Oof, that was a HUGE gap between "free-er the market - free-er the people" and the socialist message at the end. I honestly thought you had a political schizophrenia with "we should regulate the market, but free market is actually better, as my [totally not government enforced] mantra says". Good job man! Really hq vid you got there.
_"The irony that we're automating the production of art instead of the jobs everybody hates shouldn't be lost on us."_ the fact that I and all my cynicism didn't cath this 'til you said it, leaves me feeling all different kinds of bad
To be fair I think those jobs that we dont like are the ones that most of the population are employed in so if all of a sudden they wanted automate all the unwanted jobs people would be in uproar under the threat of unemployment by automation. But I dont think people will care very much if its just a sliver of the job market thats being automated such as artist , this is also compounded by the fact that more people benifit from ai art than some higher ups that benifit from making somthing like autonomous machine that makes production more efficient.
For the average person using AI to make art I suspect it's a kind of creative expression they didn't feel able to indulge in before. Art is hard and I've known many people who've given up on it. I don't hold AI use against these folks. On the other hand, people and companies who are using AI to save a buck and cut out artists? Fuck them. It's already difficult enough to support yourself making art.
It's almost like full automation is an unrealistic sci fi meme and this video never actually talks about that aspect other than stating stuff that everyone already knows and getting angry at generated images as if the ai is taking away jobs from creators (its not). Really, so long as a robot cannot even fold a piece of fabric that is given to it, something that proves to be exceptionally hard to do, a chimp will be more useful in a factory than an AI. Looks like we are all just going to have to keep struggling in this process we call life like our ancestors before us. Maybe that's not so bad.
@@dreyri2736but It is taking jobs from creators, there already book covers who have been done by a human previously and now are AI generated, brand design, anime backgrounds, commissioned art, there already stories of people training Ai on the Art of an artists the previously commissioned and in the future it will take more jobs, specially entry level jobs, I agree there probably won't be a fully outmated future any time soon, but automation always have a downside and people often struggle to pivot and sometimes can't after their jobs are outomated out of existence and capitalists are always trying to increase their profits, so if they can fire some employees to cut cost even if AI isn't perfect, they will do it, they already done with most of phone and online costumer service and it's awful.
I'm alarmed by this use of "cath" as a verb that I haven't previously come across. "The fact that I and all my cynicism didn't"... catheter this? What?
I'm a huge AI skeptic, and as a software engineer, that's not an easy line to walk. The best sell that I've seen for AI, is that it will function as a copilot for all the shitty, high time consumption, low impact tasks we need to accomplish on our day to day. It shouldn't take away anyone's job, it should give them ability to focus on more meaningful aspects of their job. When you examine that more closely, you realize that it means that AI is going to take away the jobs of low level employees, that need to start off as basic, entry level people and work their way up to challenging, advanced problems. Where will this leave the industry in 20 years, when the junior count has declined so drastically that there will be no supply of expert level people? Is the bottom line more important than helping someone grow as a professional and enriching society as a whole? Rhetorical question, obviously the answer is yes. It sucks, on a personal level, to have to walk between raindrops to avoid unethical, shitty high tech companies. Even companies that try to make the world a better place might end up doing it in the wrong way.
@@playerslotavailable3810 I don't see any industry growing in terms of demand to such a huge scope that it'll allow the current amount of jobs to stay the same while also providing massive automation tools. It's just not the same as farming, which needed to be automated to allow for humanity to survive in scale.
@@sxnorthrop Do you really think that there will be such a massive increase in demand for production, that it'd go in line with the increased productivity that'll happen in the AI revolution? I just don't see it. I'll be happy to be proven wrong.
@@Kyfow I think I may have misunderstood your original point. I thought we were comparing technology that created a staggering transition in the economy due to it's demand. That, if looked at consistently, would show that society had a very short "grief" period for the countless lost jobs and livelihoods of illuminators and copyists (binders, bookmakers, etc.). Same will be true for ML (in my opinion). That being said there's really not enough data for anyone to make a confident assertion that ML will have the same or comparable impact. I was simply providing a hypothetical comparison to something that showed that the "bottom line [was] more important than helping someone grow as a professional and enriching society as a whole [through creative expression]", and yet today we don't think twice about how it affected multiple generations of people who experienced turmoil as a result. Every pivotal moment in history has some groundbreaking technology or discovery that was never thought possible before and it mostly ends up just making us better at what we already did before (not commenting on whether that's right or wrong ethically). TL;DR, I'm using a past example, not a current. The printing press is a perfect example of technology that destroyed lots of livelihoods (and was driven forward in-part by 'money hungry business moguls') but we don't care about it at all. It created a shift in society comparable to something like the AI revolution would create.
its already been bland tbh, seen those ugly 2010s gray on gray interior design styles or those hideous 60s concrete british council towers? but i feel like even if ai takes over artists will still prevail, its not a simple industry like many other lost positions, its a exhibition of someones creativity.
"Day by day, we live more and more in a world made *by* humans, but not *for* humans." It's a line I reiterate often, and the fact that AI is being used to replace the creative pursuits that people *want* to do instead of being used to mitigate the amount of busywork people are tasked with doing instead of being able to enjoy their lives-- it truly paints a picture more clear and vivid than anything the machine can put out about how grim the future looks if no one rises up to slay mankind's new self-proclaimed gods.
@@Pvinini Banning won't make it go away, unfortunately. It would just make it even harder to control.... As history illustrates with all attempts to ban things that there's an active demand for...
The danger is not that AI will become sentient and rebel against humans, but that it will be applied as though it were capable of human judgment when it is not. "The AI did it" will become the go-to excuse for all manner of human incompetence and deviousness, just as the magical "computer error" of fifty years ago.
@@danielblank9917 They're two different dangers, and one is vastly more immediate a concern. While it is important to learn how AI algorithms reach the decisions they do from their sampling data so we don't accidentally make a sentient AI with unintended behaviours, it's mostly a hypothetical pre-emptive problem. We have no idea if or when a truly sentient AI will be built. On the other hand, people trying to offload culpability onto non-sentient AI is a real problem we already see.
@@Jonas-SeilerWhere are you taking this confidence from? Do you think it will not happen in a near enough future to care about or do you think it can't happen in principle? Also, AI doesn't need to actually be sentient to be dangerous, being more intelligent than humans is already enough.
19:54 I'm reminded of the observation that the problem with building a bear-proof container is that the smartest bears and the dumbest humans overlap. We're getting (or already at) a point where this is true for AI.
The term "Artificial Intelligence" has been thoroughly ruined by LLMs and Image-Generators. I can't talk about the behaviour of units in a game without people thinking now that there are hooked up to ChatGPT somehow.
@@theflyingspaget No, "bots" and "NPCs" are the TYPE of entity. We're talking about the BEHAVIOR of the bots and NPCs when we use "AI". Crap like "if player is seen, chase player. If player was not detected, patrol from point A to C to D to A, repeat until player is seen."
It's all AI... if people are stupid and don't understand that it doesn't mean the term is wrong. In fact I would argue that generative AI is much MORE AI than video game AI... which is mostly simple scripts (very few games use actual machine learning for their "AI")
This video reinforced something ive been feeling for a long time. The internet doesnt seem fun anymore. Everything is funneled into narrow channels of blandness. Its driven a deeper desire for me to unplug as much as possible.
@nostalgiatrip7331 I agree but being online doesn't fix any of that, it's just an escape. I escape too so I'm not judging but it just feels less and less enjoyable
The internet is going through the same bureaucratic money-wringer that pretty much all public spaces went through before the internet existed. Cafes, parks, sports areas, town squares, etc. now require money upfront for access to their space and/or provide a bland free space and a pay-to-access fun space and/or are *technically* free but find ways to heavily discourage use without payment. It's part of the reason people fled TO the internet in the first place; you could generally do whatever you wanted, whenever you wanted and as long as you wanted, for free. Games, media, hobbies, socializing, it was all completely untethered to money and the real world. Now that enough businesses and corporations have figured out how to monetize all the "public spaces" and "free-use services" of the internet, it's just one big, digital, global mall-city, where you can either pay to do what you want (directly through giving funds or indirectly through selling personal data and watching ads) or you can sit in a corner and frown for eight hours. It was only a matter of time, but it sucks that it's already come to pass in our lifetimes.
@@nostalgiatrip7331Well, be consoled by the fact that some would actually love to be able to live and run around naked in a warm sunny place... but instead they must be dressing up and wondering how to pay the heating bills in a climate where half of the year it's wet and dark outside.
1:00:00 holy shit you have literally articulated this one thought I've had stewing in my brain for the past few years, the modern internet is becoming a conveyor belt for "content" rather than a collection of things you can look for.
i've been noticing this trend with tiktok as well, your brain is literally being conditioned to not put in the slightest bit of effort to get rewarded and in a generation or so we'll see the horrible effects of it rear its ugly head once all the iPad kiddies grow up. Even older folks aren't immune, both of my parents are nearing their 50's and they use tiktok every day. The worst part is that this is literally what some companies want, and whether or not they're cognizant of that fact this is just what the market is going to do, make you consume the most slop at the lowest cost to them.
@infiniminer7677 Good point. I think people don't understand that they think so called "free market" wants this and so it gets what it wants. What they aren't considering is that people's wants are being manipulated by the Edward Bernays market psychology even more so now with technology. What do you do when the people can be programmed into believing that their choices are there own when all of the sudden they are not?
I’ve been seeing a lot of AI on crochet subreddits. I didn’t expect to see AI leeching into those places, but it tracks. There are AI-produced patterns, which is the same kind of thing as AI-produced recipes or anything instructional. At first glance, it fits right in among other patterns, but then you get halfway through the work and Bluey has a mile-high forehead and the pattern quality just begins to degrade. Whoever is training the programs hasn’t been able to make it spit out images with accurate or natural-looking stitches yet, but it’s only a matter of time. Naturally they started with what seems like the most profitable area of crochet, amigurumi. What I’ve seen isn’t them selling a non-existent finished product, though. I haven’t looked but I’m sure there are paywalled AI patterns out there too. The grift for now seems to be getting people to the website and throwing a ton of ads on them, which is already a thing in crochet. There are tons of blogs and aggregation sites. It’s a lot faster, I assume, to cut out the part where you look for patterns to share and to just fill your site with fake ones. Then make more, which all refer to each other. It’s all the same as other content aggregation sites, but it just feels so out of place to me. I mean it all feels out of place, but finding patterns is already a time-consuming task and these AI-based sites add exponentially to the problem while masquerading as a solution in the form of an aggregator site. Idk this whole comment is just a rehashing of the same bones of an issue with crochet skin over top haha
As a software engineer, and as a comic book illustrator (with actual pen and tablet and time to hone my skills), this video hits the nail on the head on everything wrong in the industry.
Maybe we should ask for an "AI" that can do all the tasks of a CEO so we can free up billions in capital by eliminating the need for executives. I wonder how fast laws are past after that hits the market.
It's already happening, sort of. Problem is, it's China. We don't know if that one Chinese company that replaced their CEO with an AI (with obviously a human liaison) is real or just a CCP fluff.
imagine co-ops with AI CEOs. As dystopian as that sounds (to be working for a machine), the awesome part might be seeing what a companies pay structure looks like without having to pay the CEO 500 times what the rest of employees make. That seems like automation that could save some money right there.
@@jasonmast7769 if it's training data really comes from a bunch of internet comments, it would probably redistribute the wealth among workers immediately.
CEOs are the grounds keepers of the capital investment upon which companies are founded. CEOs are eliminated all the time. But the only reason CEOs exist is because the people who fund the company to begin with want them. CEOs are not required but desired. No company is forced to have a CEO if the people who fund it don't want one.
“With new technology, the line between research and commerce is razor thin, and big companies often use this fact to just ‘manifest destiny’ whenever they want and make us live with the consequences.” Jimmy, you cannot hit us with a line that hard less than 10 minutes into the video.
Would you prefer that progress stop in the meantime? What's the AI going to do that's so horrible? Jack up insurance rates when the health insurance industry should be nationalized, and we need more transit and bike infrastructure so that people don't need car insurance anyway?
@@Ilyak1986 Well you automate one thing, then the next thing, then the next and before you know it the only jobs left will be AI maintenance. Not everyone is capable of that, so then you deal with a jobless problem, next a homeless problem and next a society collapsing problem. Even if AI could eventually think for itself, we will never get there because the creator's society will collapse before we can get there. Ironically AI is not the problem, but the missuse of AI trough human hands and corporate greed... That is the problem.
@@Ilyak1986 "What's the AI going to do that's so horrible?" Just some things that AI already done: If you know how photoshop already was a problem in the past for creating fake images that were used and abused by fake news you will understand that AI image generation will be this but 1000 times worse because anyone can put the name of a real person there and generate many images that to most people will be a real photo, if AI was ethical it wouldn't generate images of humans at all because to surprise of no one is already being used for malicious end, to basic porn bots to revenge porn (and if you go to the underbelly even have the "great" children variety that already trying to push as normal and ok because no kid are being harmed). AI misinformation makes hard to find true sources and is destroying Google* and is already being deadly with the misinformation because has books about mushrooms with the wrong information so people that buy it can end up eating poisonous mushrooms and the one AI from a supermarket that made a recipe of a deadly gas as good ways to clean the house. *Look up "how many african countries start with K" on google and now remember that most people only read the first result
I genuinely never thpught there were many people THIS desperate. Fuck, I'm lonely and possibly depressed as all hell and I would never stoop to something this... sad. Dehumanised. *Artificial.* The fact the market for this exists alone is depressing...
@@backoffpeer Nay, the best thing we can do is hold ourselves to a standard of our own choosing, keep our "spine" and integrity intact. I just wish that more people chose this instead of brainless complacency.
@@Box-O-Soldier I'm the type of guy that watched Blade Runner 2049 and thought to myself "Dang! I wish I had a robot girlfriend." ngl life is bad enough I would buy that shit
Yep. I believe there's an AI service out there that charges quite a bit of money for a chance to chat up with AI, it's so scummy. At best I'd really only ever use AI like this for fun and leisure. I'm fortunate to already have people in my life to connect to, but not everyone has that luxury. . .
@@ModernCentristdon’t rely on that tho. Even if that miraculously happens, we need to do everything we can on our end. Relying on AI like it’s some savior-god isn’t enough
You feel right. They are the bastard offspring breeds of man (women) and fallen angels (demons). Satanically animated machines. And they even go as far as claiming to be God's chosen! Exponentially more amazing still is that the very people that say they love the Most High God (instead of rightly fearing Him) actually believe that to be true! Once you notice them, you will see they do business everywhere. Many - many great men and painters throughout history have; and warned about them.
The tuberculosis story reminds me of Clever Hans, the horse that could do math. He was believed to be able to solve equations and answer trivia. What was actually happening, was that the person asking the question or the audience watching would often react when the correct answer was suggested. He would then stomp his foot on seeing the reaction, signalling the correct answer. He was unable to do this when the asker/audience had no idea of the correct answer, which is how he was found out.
it was actually when hans's trainer was unaware of the answer or not present. the horse was picking up on the trainer's body language, not the audience 😊
In academic literature about explainable AI, they actually call these faulty models Clever Hans predictors haha. There is a lot of research done to try to understand, when these models are right for the wrong reason. For example search for: Unmasking Clever Hans predictors and assessing what machines really learn
It's good to realize that the "AI is going to fix everything" and the "AI is going to pull a Skynet" opinions are mostly fed from the same sources. So long as you're buying into the premise that AI hitting a singularity moment is the goalpost, you're going to miss the ways its hurting people right now.
From people like Zuckerberg, Musk or Altman who practically doubled their wealth during the pandemic but now flush jobs down the drain in the thousands but hey u can make porn with midjourney so the techbros will just shrug
Wrote a grad thesis on ML in aerospace. Nothing depresses me more than seeing how corrupt, toxic, and disenfranchising the field of AI has become recently. It’s so broken the only real solution is to let it die under the weight of its own expectations, since the math doesn’t support what it’s doing now all that well.
It's been a relief to hear experts are feeling the same way I am. I don't think we're going to get out scot-free, but OpenAI could at least go out of business once the hype really dies down. Little victories.
1) loved the video congratulations on the stellar job you did! Really like the editing style and composition! 2) without launching into a tirade and linking a bunch of papers, i optimistically give it ~10 years before it fades completely. Again, the math just doesn’t support the applications rn.
It's a shame, really. It's all very interesting stuff but silicon valley is trying to make it the next big thing, making folks think it's more than what it is.
Brhu, is just the next push on technology, digital field Its the same as when robotic arms where implemented in industry line Make job easier and more standar, better quality. Also made a lot of jobs useless. So obiously corpos are gonna capitalize on it This is not beause cof the technology, but the nature of capitalism itself Money matters, anything else is secondary with human expenses at the bottom of said list. As money is power, and power let you do watever you want.
I want to add something from my field of expertise -- computer graphics. A couple years ago there was this so-called breakthrough in this field called NeRF (Representing Scenes as Neural Radiance Fields for View Synthesis) that's basically a deep-learning-assisted rendering algorithm that produces a neural representation of a 3d scene from a bunch of images. When its first paper was introduced, it was pitched as the biggest breakthrough in the graphics and supposedly it was expected to replace all conventional rendering because of how powerful it looked. Unfortunately, at the time its inference was about 100x slower than what it needed to be in order to be usable in realtime. Fast forward 3 years later, that technology has finally advanced to the point where its performance is almost practical for realtime applications -- its performance has evolved and now it's about a 1000x faster (in the newest iteration called Gaussian Splatting) than the original version. Except one detail: there's abolutely nothing "neural" left in it. Literally nothing at all, and it's the only reason why it's becoming practical now, because the main bottleneck of that approach was the neural network inference and getting rid of it made it 3 orders of magnitude faster.
In other words: The stupid AI found a creative solution to a problem that humans didn't know before and indeed struggled to understand how it even worked before completely dissecting the AI blackbox. :D
Yup. My field is embedded hardware engineering and for a while, we too where looking at having to add AI inference in embedded systems. But much like with NeRF. Most of the planned uses just ended getting piecemeal replaced by regular algorithms and switch statements. Cause inference is expensive and often required some extra Neural processor like a Neural Compute-Stick to get anywhere fast enough. So slowly more and more tasks where being removed from the Neural Network part until eventually there where no actual models at work. Just regular algorithms that where small and efficient. The only thing we still use small neural models for, is computer vision. But that is so old and heavilly optimized via specialized libraries like OpenCV. That you can do it with any half-decent ARM system like a RPi zero.
On the one hand, this sounds hilarious because it makes it sound like progress was made, and then to justify their existence, they made a normal renderer that actually worked. On the other hand, regarding what actually happened, getting an AI to brute-force a good solution so it can later be chiseled out sounds like a very 21st century career.
@@stevencurtis7157 in this case AI didn't even bruteforce anything useful that got eventually "honed out". The actually novel and practical part of NERF(and many other DL "advances") is to formulate the task in a differentiable form so that it's _possible_ to optimize it at all with a "black box" optimizer of some sort. But as it repeatedly turns out, neural networks are just a really sub-optimal choice for such a "black box" optimizer, and there are other task-agnostic optimizer methods that perform simply better. So the big idea is to approach a task from an optimization standpoint (which was arguably never tried before NeRF in the rendering field), and that is the novel and important part, and neural networks are just one sub-optimal way of accomplishing it.
Kind of a bad take take. NERF got useful after instant ngp, and I have used it in robotics. With Gaussian Splatting, yeah, the neural network architecture got removed, but it still is machine learning because it uses gradient descent to the core.
"i don't care about mars. i want my friends and family to have the security to live good lives and a safe, quiet place to do it in." who doesn't? it's possible with our technology and resources, it's just that the execs are willing to do everything in their power to prevent it and for what, endless profits? thank you so much for making this vid, really thought provoking
I have a single course degree in machine learning from Stanford University Online, which while it isn't a credential that will get me hired, it DOES mean I've gone through the process of actually understanding what modern AI is. I spent a lot of time doing extracurricular experiments to get a good intuition and feel for HOW these statistical models we call AI actually behave, and what their limitations are. So while I may not have a PhD, I can confidently say I have more hard technical understanding and insights about AI than 95% of the population. The things you laid out in the first 12 minutes is EXACTLY what I try to tell people who say shit like "just give it more time, and AI will explode" or "I'm afraid but excited for the upcoming AI explosion". There is NO MAGIC, it's JUST A MACHINE WITH A BUNCH OF KNOBS. More data, more time, bigger models, bigger computers, quantum breakthroughs, all of that is just like polishing a turd! Yes, AI in it's current form can do a LOT of very useful things, and we absolutely should be using it to replace annoying, menial tasks like removing backgrounds or fixing lighting in photos. But I'll scream it into the clouds if I have to, NEURAL NETWORK BASED SYSTEMS WILL NOT BECOME ANYTHING REMOTELY CLOSE TO GENERAL PURPOSE. THEY WILL NOT BECOME SENTIENT. I really, really wish it were possible, I would also like to see some crazy AI future, even if that concept is scary! But if you understand HOW these systems work, then you can see how it's literally not possible for them to do these things. It's like saying with the right fuel, your car could drive over the Atlantic. Understanding how nodes work, how layers work, and how gradient descent works is fundamental to realizing this is true. While modern systems have lots of advanced features like nodes that remember things over time and complex systems made of dozens of smaller task-specific models, those things don't change the fundamental limitations of dataset availability, data quality, and the fact that the system is at it's core trying to make a number smaller. (Any time you tell someone to optimize a handful of numerical metrics, they will find "cheats" to technically fit your requirements, and AI does this too.) This video restored my faith in humanity a little. I'm sharing that segment around. You earned a like, a subscriber, and a binge watcher.
The fact that such a thoughtful and insightful critique of this video, seen by 1.2M viewers, only got 28 replies is telling. "There is no royal road to learning." Euclid and others....................
What is your point, actually? It is not clear from your message. It can't be "if somebody knows how the net's elements work, that means sentience can't exist," because we agree that sentience exists, and it is understood how neurons work in general, and there are no limits to finding out more details. Also, it can't be "math can't become sentient," because if we can simulate someone's brain with math, and the simulation is good enough, the behavior of the simulation will match the brain's behavior to the degree needed to produce the effects the brain produces. I can't think of any other points you might mean by writing your post. Can you please share?
I'm a statistical programmer - and I agree 100% with your comment! I have only a rudimentary understanding of the AI algorithms themselves (I programmed classification trees and random forests which I imagine are baby versions of them), but the obvious issues with AI content (lack of consistency, inability to 'remember' >100 word prompts, over sensitivity to training data etc) are hard baked into the black box statistical methods used and can never be improved upon. There's no model parameterisation that can control 'consistency' or 'memory', the ceiling has already been reached on how good this stuff can get. While I imagine the AI interfaces and training data sets might improve or become more efficient, these models are as dumb as a spade and always will be. I'm so sick of all the fear-mongering hype generated about this garbage! It's all to scam investors as the venture capital cash is drying up in Silicon Valley. It's not based on real science. It's based on science fiction fairytales and Terminator 2. The day we can bring the dead back to life is the day we can instill consciousness in a laptop 😂 Once the hype dies down and these money-gobbling tech companies sink into the abyss where they came from, people are going to realise that everything AI generates is rubbish. It will never be capable of replacing humans in the vast majority of jobs - that is a lure designed to scam investors for cash - and people need to stop getting their science understanding from the agenda-ridden nonsense in the mainstream media. Also, to any artists out there worried about AI stealing your work, make sure to use Glaze and Nightshade and fight back against this disgusting theft of your IP!
@@silverbirch-youtubeI have some undertanding of how AI works, and it does seem like its the algorithm we use that is limiting its intelligence. it still doesnt map to a brain well enough. However, the brain is proof that some algorithm exists that can reproduce intelligence. We know our current algorithm has plenty of similarities with how humans think, like being really good at pattern recognition, and being capable of extrapolation. However, we also know there are differences. to train AI to sort apples by color and size, you show it samples of both mixed in, while humans perform better if you first teach differentiating by color and then, separately, by size. From this, we know we are missing something from AI, that the human brain does. Dismissing the concept of AI solely because of the limitations of the current version is probably wrong.
Don’t think i’ve ever commented on a video before but i just feel so strongly about this one, i genuinely think it’s the most important video I’ve seen on one of the most important topics of our future. In a time when it feels like most vocal opinions on this topic are either ignorant or completely insane, thanks for this extremely enlightening video.
Prompt: Write a comment on the topic of "desire to leave behind the current era of formulaic and stale AI generated videos", in a manner that ensures that the comment will get broad recognition and user appeal, by choosing a recently popular snowclone format textual meme and inserting the topic as its variable term.
What's funny is that since the release of this video, people generated so much AI images that the AI itself started training on its own data, making results worse again - amazing.
Somebody pointed out that AI is perfectly tuned to appeal to the "idea guy", you know, the obnoxious person who says hey, I have an idea for a book, how about you write it, and we split the money? He values the writer's work so little that he thinks you're going to give him half the money after you do absolutely all the labor to make his pedestrian idea real. AI finally gives him the slave he's always wanted, or promises to, and it's a big reason why the uptake has been so swift, despite the AI being pretty bleh at most things.
you just managed to formulate my thoughts for me (which considering the topic, is ironic) this has always been bothering me, as there's plenty of people with ideas, but very few who actually take the time to implement them, and even less people who learn to implement them better with each iteration. that same thought process also explains why its so popular.
Pretty much every AI bro I've met on the internet so far was an idea guy. And it's funny how they flip flop between "this will make artists' skills irrelevant" and "this proompt took skil, praise me"
But that's exactly how eNtRePrNeUrShIp works. U have an "idea" and a bunch of money to make ppl make sth actually useful out of it, then pocket most of the profit. The ideas bro has only one thing that separates him from the Elon Musks of the world: he's not already a millionaire.
This AI stuff just makes me sad. Apart from the stuff about killing jobs and creativity, what saddens me most is that it's killing critical thinking. I see more and more people using AI to do stuff like make drafts of their work, plan their day for them, summarize research, etc. And I take their point, it can definitely help automate some of those trivial tasks. However, it's taking away all forms of thinking from the equation. Drafts are good because it gets you to THINK about what you're going to write/create; planning your day makes you THINK about what you need to do and what to prioritize; doing research and summarizing the points of various authors makes you THINK of their views, what the key takeaways are, and allows you to assess the quality of the work. Using AI will make it easier, sure, but you've put yourself at a huge disadvantage in terms of having control over your own life and what your mind is capable of.
People who don't want to think will find a way to be stupid with or without an AI. Fuck'em. Hopefully, they're going to get exactly what they want, and be able to exist as immortal and nearly mindless hedonists in FDVR pods for eternity, and leave the rest of us alone. As for jobs, I'm actively rooting for AI to take every job ASAP. I want to live in a world where no human is able to contribute anything of any meaningful economic value to society, because I think that's the only scenario in which we pull our collective heads out of our asses and make a strong push for a more egalitarian, human-centric system, as opposed to a hierarchical, profit-based one.
Nonsense. You could make the same argument about calculators, that they hurt mathematics because we rely on them giving us the answer instead of calculating it in our heads. AI will not "kill creativity", just like calculators did not kill our mathematical abilities. People will simply use their creativity more effectively and not waste with stuff an AI can do. The critics of AI are the same as their prophets. One side claims it will be a revolution that frees us from all ills and the other that it is the devil that destroys humanity as we know it. Nonsense.
I’m a uni student who’s half way through writing my dissertation and although using ai to help hasn’t really been a temptation for me, it’s fairly often used by other students. Most of them reword and change what it spews out but they can still get caught by plagiarism scores and lectures who know their shit. My dissertation advisor even mentioned a time when a student submitted a completely AI written dissertation snd failed their entire degree because it created doubt about how much of their work is actually their own original work. For me it’s not really been a much of a thought because I don’t want to risk my degree over something so stupid, plus I enjoy knowing my work is 100% a result of my own effort and time. Even if there wasn’t the risk of disqualification I find AI is only meh at answering questions that are popular and makes very generic responses and is practically useless for anything niche.
I can't recall the last time I watched a piece of media where almost every sentence carries weight and meaning. You've put a lot of effort into producing this and it shows. Excellent work.
on the "democratizing art" point, taking it to mean that ai is creating equal grounds for all to participate in the creation of art, the real phenomenon is that people's access to the *products* of art is being made more equal. it's a very cynical view- one which erases the concepts of artistic process, of honing craft, of exploring one's own tastes and philosophies and the history of their chosen medium, of journeying to one's most authentic style of expression. it views these not as rituals of growth and development and expansion, but as mere hurdles which should be dodged as efficiently as possible to reach *the thing* that can then be shown off or sold or whatever. meanwhile, the grander issue of so many people lacking access to the pursuit of art due to whatever socioeconomic constraints may be restricting them gets left by the wayside while ai parades itself around as some heroic facsimile of a solution to the dismal lives of every overworked person who dreams to just have enough time and money to meaningfully access and practice any sort of craft
@@danilo071983 Because handing off the process and the journey to let AI do it for you is the definition of skipping the journey, not a way of having the journey.
49:22 "a good chunk of people who hire designers have no taste and think they can do the job better anyway, now they can tell the ai to make the logo pop and it won't laugh awkwardly and ignore them" this is the perfect summary of the problem. The power for the decisions is in the wrong hands and people are more and more convinced that controlling capital means been always right.
TH-cam users always write about how socialism is better than capitalism. But the irony is that when it comes to AI, youtube users judge it as if capitalism is unquestionable. So AI is attacked rather than capitalism. It is so easy that users don't even have to be paid.
When you place total power in the hands of childish gods, you had better also kiss your world goodbye at the same time. They'll destroy it in order to milk maximum personal gain out of it, you'll be powerless to do anything more than complain and watch it happen, and in the end you'll all die together. Except, they'll die with a warm binkie they refuse to share with you. Maybe best keep a shiv on you through all that so you can take it away from them as the sun goes out on a devastated planet.
As long as we keep replacing jobs with machines or programs, but leave the unemployed without a way to thrive, or even just survive, I refuse to call any of this _progress_ .
As an artist, AI generated art unironically made me apprieciate artists efforts and their thought process even more. It's just so much fascinating and interesting to me how artists manage to understand the sheer complexity that is art. Ed: that's a ton of appreciation from internet strangers, and I appreciate all of you for this. Means a lot for me. Like really, thanks!
What I really appreciate and deeply admire is that there are still people who make and share art made by hand, who still try. I pretty much quit making art these days because it seems as futile as trying to outrun a race car on foot, and I have no idea how much mental fortitude it takes to still try in the face of imminent failure.
@@nobody-nk8pd I'm truly sorry to hear this. Like really. As you said competing with AI is impractical time-wise. You could get the same level of quality tho but that requires time and dedication. If you ever wanted to tackle art again, I'd advise you not to focus on this aspect of AI. I'd tell you to try to take advantage of it as much as you can to yield the best results for yourself. I still draw and render by hand but I also want to explore the world of AI and see how I can make the process of crafting a piece (obviously not by literally generating it) even more efficient.
@@nobody-nk8pdPersonally, I love making art. Despite the fact that studying anatomy is incredibly soul shattering, the idea that one day I'll be able to make my own character from the ground up keeps me going. That's how I stay hopeful. It's less about the result, and more about the journey. I enjoy the process. Being an artist for a living is scary right now because of AI, but it has always been like that. Creative industry has always been brutal to get into, so it's nothing new. Basically, I just make art for myself, and if I get lucky to have my art be popular, I'll take it from there
As a regular art commissioner, I always appreciate the hard work put into the pieces I commission. It honestly inspires me to pickup a pencil and just draw things in my artbook, it looks crap for sure right now but I hope one day I can produce the same quality art that I regularly commission. And honestly, this has helped me understand that every stroke of a brush/pen matters in an artwork, which has made me respect hard working artists even more. Anyone who spends a long time honing their craft will always get my respect.
@@nobody-nk8pdthere's still a lot of people who appreciate art made by a human being. I don't believe it's hopeless. If people adapt to the change without giving up drawing by hand, yah know there's always a way. Blockbuster went out of business when Netflix offered something new in rental. But there's one Blockbuster left! They just found a way to adapt to the change without having to do what Netflix is doing. There's always hope.
I think one thing that will happen, is it will increase consumer fatigue. It will become harder for the consumer to be a discerning customer and they will have to 'dig' more or else more quality content will die in obscurity, or quickly be ingested and shat out by a company with more money, by just ctrl-C ctrl-V an art style or LLVMing a text dump of the original game. We are already at a point where YT creators can influence if a game/movie reaches the right audience. Otherwise there is a high chance a game will die in obscurity. A lot of this is going to hinge on having enough people 'hate' AI generic content. I don't think the majority of people will care, but hopefully there will be niche communities that emphasize 'hand-made' media as a reactionary movement.
"Hand-made" is such a weird way of putting it. It wouldn't exist if you didn't put your hand on it. What people need to do is not shun the tech, but to focus on open source, and not falling into subscribing for even more services.
@@illyaeaterI'm a purist, I guess, but I balk at the term "hand-made" when companies/people try to sell their "hand-made" furniture, that was all made with machines. Some guy recently got called out because his "hand-made" guitar was made using a CNC machine. Can you imagine the audacity? Like, if you're not gnawing at the wood like a beaver or clawing at it like a bear, is it even hand made?
We are literally just repeating what our ancestors always have done. When people were satisfied, or kind of okay with their lives, they let systems "control" them, when they had enough they changed the system (governance or nature itself, doesn't matter, just some kind of system they lived by). The only caveat I see to this is if (or when) we actually manage to "reverse engineer" our biology and mind and become able to play God with ourselves - not just our environment. We might kill our humanity in the process and a new age would begin when we no longer live by million years old rules.
"AI art is like photography, because the photographer didn't make the things he took a picture of." Okay, my dad taught photography and let me tell you, it's not that simple. Photography is all about the ability to consider pose, lighting, framing, and composition, which you have to do in the moment and with a good instinct for timing. Anyone can just snap a photo and have it contain the stuff you want it to have in it, but it takes lots of training and experience to be a photographer that can make art.
The way AI is being used is emblematic of how the expansion of human knowledge is ceaselessly commodified and weaponized against ordinary people. Makes me feel like the guy in Mad Men who lost his mind after coming into close proximity with a computer.
24:44 As someone who has used AI Dungeon before, let it be known that unless you wanna type out long walls of text directing the AI towards your desired story choice, it will just do whatever it wants.
@@Jerryfan271I think it's because it got less inconsistent. Earlier AI dungeon was far less intelligent and more creative in how it did things and I miss that
I dunno if someone mentioned this already but 12:16 reminded me of how there was a time on Google where they posed that type of image-training as a "game". Google would present random images and you (and another player) would tag them. I think the idea was that you didn't see the other player's tags until the end and you got more points if you both had similar tags (with the same words). I remember playing that a lot when I was really young. So, basically, they got 11 year olds to do their work for them, too.
Also, I am hoping that AI's algorithms become very transparent to the user, where they tell you and show you exactly what they are doing, what actual images in their database they are shifting through/melding together etc. Its very important to see the dark sides of the AI's algorithms, it will make sure that younger people make the right choices.
that's not really practically possible; one of the most fundamental limitations of any kind of computer learning is that you don't learn what the computer does; the images don't exist anymore in the dataset, they've been distilled into rules. i'm not educated enough to completely rule out the possibility of something like that existing, but it would have to be a fundamentally different technology than is currently in use@@DonnyKirkMusic
THe problem with AI is that it's neither Artificial nor Intelligent... a human somewhere has to add input into the system from then on there are a bunch of conditions (if else). Sure maybe they are using all our input all the time to train the model, but if you try to give AI problems that require some real thinking instead of repeating some existing model it fails miserably.
@@BillClinton228 thats not how AI works at all... atleast, thats not how neural network AI works, you should learn before you speak so confidently on a topic you don't actually understand
@@madcatmk2sch thats arguable, i think it was worth it to make it that long, i needed this in an internet feed full of frustrating 1minute videos, this forced me to pay attention for much longer after that intro by hooking me at the start
I'm a motion designer/2d animator, at work, I have at least 3 meeting a week discussing how I can 'integrate AI into my workflow'. I don't know how much more of this topic or those people I can take.
I'm very curious, what do they ask you and how do you respond? I have a friend who is a programmer, he was questioned about similar things by employers, but luckily for him, his workplace eventually dropped the idea of forcing programmers to use AI. It's a plague.
This was possibly one of the best video essays I have seen on youtube after a very long time! I can only imagine the amount research that went into this.
This genuinely one of the best, most well articulated video essays I’ve ever seen. This is genuinely approaching the realm of Line Goes Up by Dan Olson (although quite different). The way in which you make and word your (excellent) points in this video is masterful, and every step along the way is just as good. The conclusion was so well worded that it made me write this comment. The 2nd-to-last chapter also really stood out to me. The overstimulation section (you know the one) managed to create a truly dire and *deeply* felt portrayal of this internet nightmare that really resonated, and drove your point home in the best way it could. It might not be the most structured video, but it’s certainly incredibly compelling.
You had me at leaded gasoline. I spent the better half of my adult life trying to understand market failures and examples like DuPont's made me realized how adversarial market actors can be, hiding the externalities of what they're doing to increase profits, then using that money (which economic theory says wouldn't occur in a perfect market with perfect information) to rewrite the rules. Great job connecting the through line at the end to pricing discrimination, nudging, and commodification of time, all things that result from an information asymmetry and that stand to grant specific companies a lot of money, market power, and even political power depending on how that money is invested. The real runaway superintelligence is already here, already misaligned to human values, and will bring about a world that I think most people wouldn't want if it were presented upfront to them.
This! It's not that ai will become sentient and rebel against the humans, it's that it will supercharge the casual inhumanity of commerce and industry.
I think the current implementation of AI is the only natural conclusion of this technology under capitalism. AI is just accelerating many of the existing problems of capitalism and depersonalizing them
When an environment produces something that destabilizes its own ecosystem, is the problem the thing doing the destabilizing or the environment itself?
@@aarvlo It's the natural conclusion of technologization. I forget who said it, but "The Soviet Union and us [the US] have the same goal: technical efficiency!"
@@echoecho3155 Well, good thing that you forgot who that goofy was, because "technical efficency" was never a goal to begin with. Historically and nowadays, effeciency serve a purpose, then, an intention. Efficency for it's own it's not an idea but a closed loop that looks like what an AI would make as a gibberish to fill its emptiness.
@@Factuel-ry7zi I think you may have misunderstood my meaning. Jacques Ellul posited the idea of "technique," a mindset by which all human activities are subjugated to rational, technological methods of increasing a system's order and efficiency - reducing cost, removing roadblocks, and maximizing output. Whether it's utilizing AI to mass produce art-as-content or forcing peasant farmers off their land and into factories to maximize production, the mindset is the same. Humanity is subordinated to maximize the system's efficiency. As all economic systems seek to more efficiently utilize resources, whether socialist, capitalist, or communist, they all submit to technique. So yes, the USSR and US pursued the same goal: maximize technical efficiency at human and environmental cost. Both were at the mercy of technique as a guiding ethos. John Michael Greer's writing on "Progress" as an ersatz civil religion are also relevant, as that ensures technique proliferates.
I'm beginning to think Big Tech is actively trying to devalue art because art's what's making us think critically, too critically for their liking... A proper conspiracy theory but one impossible to dismiss, I fear.
Holy shit man. I was listening to this video absolutely enthralled with what you were saying- you have a really wonderful way of essaywriting- and I had no idea how I’d never found your channel before. But then, you brought up Rain World as “high art”, and I checked- and you’re the one who made the world’s best Rainworld deepdive a couple years ago. I’m so glad you’re doing more, super insightful stuff.
Updated as of-12/27/2023 ---------------------------------------------------- I've been fascinated by what AI dose to what we think about authorship & originality. I think that AI will make use confront the possibility that ancient hominids producing "cultural artifacts" simply may not have anything to do with being sentient life at all. So when I watched this video, and saw some of the common misunderstandings about the technology of AI in relation to American copyright law. I thought hay I'll just slam out a quick explanation. lol 1). AI training sets are full of stolen art? --In America the training sets that are essential to the functioning of art AI were originally compiled by universities and research institutions under the classroom use exception to copyright I.E. fair use. so the artistic & literary works in these training datasets arguably compiled for the purpose of education didn't need to be licenced at the time thy were made. The why these training sets were made dose not imply that companies can also use them directly to make products, this present a problem for them. That will likely need to be hashed out in court. Open Source software on the other hand. Because they are not explicitly making a for profit product. Will likely be far more successful claiming “fair use" so it is IMO vary unlikely that training sets made this way in the past. can be stopped from being shard for the foreseeable future. 2). AI can't produce original ideas or art. --The American copyright office has touched on the problem of a "non-human creator" before. ---- "The 2014 Monkey selfie copyright dispute" --Resulted in a copyright ruling that works created by a non-human "Animal" creator immediately enter the public domain upon there creation. No matter how much knowledge and labor a human contribute to making things ready for that creation. I E "sweat of the brow" ---- "The 2023 Théâtre D’opéra Spatial dispute" --Resulted in a copyright ruling stating that Art AI are authors of the original works thay produce, “when an AI technology receives solely a prompt from a human and produces complex written, visual, or musical works in response, the ‘traditional elements of authorship’ are determined and executed by the technology-not the human user.” AI Registration Guidance, 88 Fed. Reg. at 16,192. And "Because Mr. Allen has refused to limit his claim to exclude its "non-human authorship elements," the Office cannot register the Work as submitted. " ---- "The 2022 Zarya of the Dawn Dispute" --Resulted in a copyright ruling that says in-part. ""Rather than a tool that Ms. Kashtanova controlled and guided to reach her desired image, Midjourney generates images in an unpredictable way. Accordingly, Midjourney users are not the “authors” for copyright purposes of the images the technology generates. As SCOTUS has explained, the “author” of a copyrighted work is the one “who has actually formed the picture,” the one who acts as “the inventive or master mind.” Burrow-Giles, 111 U.S. at 61. A person who provides text prompts to Midjourney does not “actually form” the generated images and is not the “master mind” behind them."" ---- The processes that humans naturally follow to create original art has been enshrined in copyright law, this is why we can separate. In a court of law: A copy of a painting. From a derivative of that painting. From a painting inspired by the painting,From a painting thats in the style of that painting. Being inspired by a pice of art you’ve seen, and then painting something original in that artists style. May be in bad taste but un-controversial if the AI were a human artists and not software. In short. If AI did not produce original art based on the "inspiration" it gets from mesmerizing the abstractions of billions images. The art produced would not fall into the public domain like it dose now because of non-human authorship. Instead it would be copyright-able tho an un licensed derivatives. End. -------- Update 10/09/23 After being challenged. And re familiarizing myself w/my sources. I got a date wrong, was un-clear in places and mixed a couple arguments together. I'm making edits to correct my errors. --------- Revision 10/10/23 Revision 10/11/23 ----------- previous comment- ------------------------------------------- I thought this video was insightful aswell. but little inconsistencys with how Jimmy McGee miss describes the tech, and how it works in relation to american copyright is what changed my mind. For example: In the US the training sets were originally compiled under the fare use exception in copyright law for research and education. and therefore the works in them don't need to be licenced, And this is one of the main stumbling blocks to wide use outside of academia. Thankfully Now an entire industry has sprung up to supply different legal sets to developers. The copyright office in the states has rulings on "Art AI's" from almost twenty years ago. That apply to the art of AI's that are made to follow "the legal process" that allows humans to make new art from All the art and things thay have seen over their life or are referencing at that time. For inspiration. without violating any of those copyright holders rights. the case im referencing was a programmer who invent an AI that replicated a dying & hallucinating human mind that autonomously outputted pices of new original religious art. He sued the copyright office when thy ruled against him. They ruled that the AI was being original, and a copyright was created for each pice of art it created. The AI isn't a human being tho and so couldn't defend its own copyrights in a court of law. Nor could the owner of the AI gain use of the copyrights ether. just because of the artist that created the art at question was his legal property doesnt mein he could clame its property.
And while I do have criticisms of this video and don’t fully agree with its conclusions (as with all of his stuff) it got me thinking about the issue from a new perspective I really did find insightful- I was actually excited to read your criticism at first to maybe help me understand some of my own disagreements with the video, until I realized it was some weird legal argument written in broken English. also I wrote the original comment while high as balls
A barely 30 year old friend of mine, who doesn't understand shorts or reels and doesn't want to engage himself described it as watching yourself grow old in real time. "Growing old" in the sense of becoming incompatible with the "new world" of the next generations.
Maybe it's not him growing old, but him experiencing the older form of video "content" and now being able to compare it to the newer form of shorts forced on everybody. I don't think it's entirely an issue with age and more-so an issue with companies trying to extract the most amount of money out of people, if that makes sense.
@dormitthings4773 Obviously, there's some nuance with this, but one could few us criticising enshittification as old people being sentimental about how everything was better just a few years ago.
reason shorts dont work for him its not becouse he is old , but becouse he dont wants to consume mindless content , its not that he grew old but grew up on stuff like movies , books , and proper videos , not attention cash grab scam , i see this in food... , and just like fastfood companies are conditioning "new generation" , youtube and social media are conditioning new consumers that "this is content ofc child would want to eat just icecream , instant ramen and other garbage like that , similar is with internet content , problem is in today world you can cook for your child and feed him good stuff , but you cant really do that with online stuff , today even finding a real job requires you to engadge with a lot of that garbage compared to just 5 years ago. its not that they grew incompatible , but today world deliberately becomes more incompatible with us in order to pray on us or condition us to be exploited. ask any 15 year old how many books he read , does he read newspaper or articles? watch documentaries? good movies? , most will talk to you about roblox , minecraft , or garbage content they consume. and just like food , if you will be conditioned long enough...that this cheap low quality zero nutrition burger and fries is your tasty meal , you start to belive it , after a while you will carve it , despite it slowely kills you , similar is you youtube , social media , and attention platforms , their AI trained on model is smarter then you and me , they can indentefy weak , or individuals they can sway to their side , they got so good its not just some...but most.
Fu***g hell, I felt exactly the same. I'm 25 and I *felt* the curtain open on the "I'm never gonna connect to the poor younger people that are growing up currently" feeling. Which is a super sad thought that I will fight with all my strength, if only for the sake of my 10 years old nephews
The trajectory the internet and society as a whole is going towards really was nailed by MGS2 (faceless know it alls deciding what everyone should consume under the guise of providing context to content in a sea of worthless data) and MGS4 (AI and algorithms being so intertwined with reality that no one on the planet realizes that they're taking orders from machines anymore).
Sad that the people who would have the exact opposite position to yours also believe that MGS2 and MGS4 predicted what _they_ believe to be the story of today. (I.e., people who believe that people like McGee and you are economically illiterate SJWs or socialists who are either ham-fistedly trying to cage AI that could greatly improve the world, ignoring the "science fictional" type of existential threat that AI can pose to humanity, or somehow both at the same time; they also will point to MGS2 and be like "so visionary, Kojima would agree that I'm like Philanthropy and @cromtuiseagain is like the Patriot AIs".)
@@coreyander286 honestly saddened by the idea they could take in all that stuff about echo chambers, fake news and discourse in general becoming a shit show, and decide "kojima got it right, my worldview is exactly what he meant" and go back to shitting on "the SJW's/political group i don't like" in an online group designed to discuss that one and only thing without a second thought i'm lucky i'm irritated by loud noises and people, those echo chambers must be hell on the brain
MGS 2 honestly was optimistic in making its big bads the metaphysical representation of the US government, when what we got in reality is the ideological amalgamation of the most insufferable nerds on the planet
@@Graknorkecan't be any truer than that. The entire world of tech is just governed by the worst type of nerds that got soo successful that they basically became the jocks that would bully them at school instead of growing out of it.
You have outdone yourself! This is up there with Folding Ideas' videos about tech grifts in terms of research and quality! I'm glad I've been subscribed to you and to be able to see your eventual growth as a great essayist!
Its really funny to me how in about 8 months, after the hype bubble began deflating, saying something "looks like it was made by an AI" went from a celebration for technology to a straight up insult, as it should be. Another banger video.
[thing] is bad now and will never improve okay thats a real argument and definetly not cope or a placeholder for my unwarranted anger i got from reading a twitter thread
When ever was that a 'celebration,' are you just making stuff up?, from the moment that saying became, all I saw was people trying to insult the technology, especially the artist community.
Just a minor point: By "democratization of art" I don't think they mean "art will operate like a democratic government," but rather, "the ability to make and market art will be accessible to everyone." Which is a much more seductive idea, but also complete bullshit. "Hey look, anyone can make art! F the establishment!" While stealing work from those who actually put in the time and effort to create something original. I remember when they said that the Internet would lead to the "democratization" of commerce and retail, and instead we got Amazon. AI won't lead to the little guy making it big in the art world. It will just be used as a tool by giant corporations to accumulate more wealth.
Drawing is already democratized, it doesn't matter if you made a sketch of a stickman, it's still, art if you made it and has soul poured into it If you use AI, sure, it looks prettier, but there is no soul poured into it, just, Ones and Zeros,
also typing a phrase into an image generator isn't you making art. it really bothers me that people use that "anyone can make art because of ai uwu" because you're not making or doing anything except providing a prompt. like having an idea about a drawing isn't the same as trying to draw it. one is making art.
@@themischief420fr it's crazy that people genuinely compare those two things. I saw someone try to defend AI 'art' by saying that coming up with good prompts requires skill too, like, huh??? Typing some words is the same as knowing composition and anatomy? Sketching? Painting? Rendering? Color theory? Lighting and shading? Years of effort and passion? You really gotta be sniffing glue to think that they're remotely equal. Behind every artwork is a dedicated and talented or aspiring artist, and behind every AI prompt is some asshole who can type on a computer.
I am a uni student studying Data Science and AI. This video is spot on. It’s like my prof made his lectures again but 10x more enjoyable. If somebody wants to learn about the current state of “AI”. Send this.
It’s very heartening to hear, as an established and mostly isolated working adult who feels like corporations are just tearing everything to pieces, that educational spaces are bare minimum at least talking about this stuff seriously
@@alexxx4434 no, for money to live in. There is a distinction, but it's important. Money supercedes people in a post-capitalist world. Even people with money serve money
28:25 fun fact, Vizcom was used often to cheat legitimate artists out of money, as clients opted to pay for a sketch then throw it into the AI rather than PAY THE HUMAN FOR THE COMPLETE WORK
@@rene3924That’s a terrible take. That’s like asking someone to fix your plumbing, before they come in and take your money and “fix” it, and after the fact realizing that they did nothing. You lose out on time and money and they gain your money in return. The artist loses out on time while the customer gets the thing they want for free. It’s parasitism. Only one side gains while the other loses. When the customer requests a sketch they do that so they can confirm that this is the product they want. Usually it’s also a deterrent for customers to make any demands changing the artwork before it reaches its final stages, because they’ll already have the artwork they’re satisfied with.
@@rene3924 I don't think you understand. Artists are paid based on their time and effort. Getting a completed work from their sketch is cheating them out of the money they should have been paid.
@peaberry9413 A completed work, not her completed work. Any artist can take a sketch and "finish it" in many different ways. I don't see an issue with running an algorithm on someone's sketch that you purchased (so really, your sketch). I do see a moral issue with selling the new artwork, but I do not know who would truly own the piece. It is transformed, so technically new. But it was done by an algorithm which legally can't copyright artwork. 🤔 Hmm.. I think ultimately it's legal to sell, but illegal to copyright. I still stand on the grounds that it's a shitty thing to do to an artist though.
@@acebinko1 I guess I should add on that maybe some artists would also do the sketch for free BEFORE asking for payment, in which case if the client declines to go further - and thus the artist isn't paid - they could still run the sketch through the AI anyway. I can't recall the last time an artist WASN'T asking for some form of payment up-front, though, and it's probably to combat stuff like this.
It's not irony that AI is taking over leisure activities and enjoyable work like art, it's by design. It's a deliberate attack on any behaviour these ghouls cannot easily monetise for themselves.
they can though, spotify dropped the perceived value of music through the floor, and the cottage industry of playlist curation has musicians acting the part of "artificial artificial intelligence" trying to match the hyperspecific requirements of playlists, no generative ai necessary the digital painting people have it quite cushy right now even without the entire internet suddenly giving a shit about art
@flameguy3416 I understand it's a little quip but since it's a pedantic quip I feel okay with saying "well actually" because though they used words the implication is that they have not expressed how valuable this essay is. From their perspective the comment only served to express a miniscule fraction of the value and to express the true valuable would be impossible. Interestingly, if they said "I can't express in words that this essay is so valuable", then it would be ironic since then "is" serves to make the statement into a binary one of "is so valuable" or "is not so valuable" and not a matter of degree like the first case.
As someone who dabbles in several amateur but very personally cathartic creative hobbies, programming, modeling, writing, art, etc., the reduction of human choice and expression for the sake of constant, unabated monetization by people whose only goal in life is to generate profits for their shareholders hasn't been lost to me. I believe its one of the reason my concerns about AI have had a disconnect from my friends who parade and celebrate their accessibility. You can't tunnel vision on the virtuous aspects of a technology, its creation opens the door for virtuous and exploitative or malicious intentions. Throughout the 2010's we were able to ignore the growing problems because it hadn't yet disrupted our daily life, but silicon valley is unsatisfied unless they have your complete and total attention and control; that's what a vapid cult based on the principles of every piece of technology having to 'revolutionize the world' can only aspire to, ego and transforming the livelihoods of others. I always had felt alone in this concern, because most topics about AI are slap fights about one specific use case or problem, while the corporate PR jargon keeps chugging on relentlessly to capitalize on the market. It was nice at the end to realize that no matter how much it's exploited, the human condition will never change, and we will always peruse unity and contented mindfulness, and it's our responsibility to fight for the lives we want for ourselves. "We're at the point where we need to choose whether to build a world for money to live in, or a world for people to live in." is a quote that will stick with me.
Man I want to fight so bad but it's tough to find allies sometimes. I only run into anxious, frightened people who worry about skynet or people who have no spine and just surrender to what they believe to be the inevitability of it all. I've returned to university to get my masters degree in longform writing and it's fucking depressing how the professors are both uninformed and frightened too. I recommend this video to so many people - people really just need to fucking understand what's actually happening and stop worrying about fucking skynet.
Damn straight. Take a moment to realize that the most famous person that has ever existed to date spoke in parables, always speaking with metaphors and stories. One thing he said was so upfront and in your face , it was “No one can serve two masters, for either he will hate the one and love the other, or he will be devoted to the one and despise the other. You cannot serve Love and money." No metaphors, no stories to cover the true meaning. He was going to make sure there was no mistaking this.
It was good to read this from you! One of the downstream effects of the Great Tech Ubiquity--and I feel public facing AI just further exploits this--is that it has altered our ability to pay attention for decent stretches of time. Everything on demand immediately is not great, it turns out. I've been an early tester for both ChatGPT and Bard, and the degree to which they both just made shit up whole cloth was surprising and worrying. Names, dates, and narratives of whole situations that never happened. This does not make research easier.
@@JustinPogue There are 2 purposes to ChatGPT, while other companies only have one purpose. Their is no benevolent AI development that is going to take the busy work out of people's hands and give them back something better. Companies are already using this to replace people, and some industries wholesale. Anyone who believes they will use it for the greater benefit of society is just a useful idiot and should be ignored. The 2nd purpose and this is more exclusive to OpenAI and a couple of other tangential companies is the pursuit of AGI or ASI. The problem is that this pursuit again isn't for the benefit of humanity, they are looking to build a machine god, something that they believe will give them objective truths. That is worrying in itself, but even moreso when you investigate the beliefs of people like Sam Altman, Peter Thiel, Sam Harris, Elon Musk, etc. The origination for the idea of a company like OpenAI happened years ago, and the people behind the technology have displayed openly white supremacist and eugenicist ideologies. Microsoft even used a very famously racist psychological study from the 1990's as a bench marker for the "intelligence" of their machine learning program. Basically they believe that an objective machine god will reinforce the idea that white men are smarter, more capable and more deserving than anyone else. A self fulfilling prophecy given the kind of data they train on. I'm a graduate student currently working on a project about racial bias in these systems and it is beyond alarming that some of them are already getting deployed into the wild. My fellow journalists have failed massively in not exposing more of what's going on behind the scene. The panic and existential angst seem to be blinding everyone from contributing in a meaningful fashion. Nevermind that this is blatantly the biggest heist of intellectual property in history. It's astonishing that these companies have gotten away with what they have. We'll see how the lawsuits work out but permissionless innovation dictates that you simply keep breaking the rules and breaking the system until you've created a new normal, because no one is going to stop you. Educating people about chatbots and automation is important as well. The first experiments with a ChatGPT like program happened in 1967 with the ELIZA program. The computer scientist Joseph Weizenbaum developed it and was so alarmed by the results he spent the rest of his career warning against something exactly like ChatGPT. The problem is if a computer program spits out copy that seems humanlike and responsive to what you've given it, you begin to imagine a consciousness behind the text. Even in 1967 with a computer program so basic and so simple participants in the study began to believe the computer was actually sentient, even though the programmer knew it was simply spitting out text based on keywords in the entry. This opens the door for predatory behaviors we see today such as virtual girlfriends that prey on male loneliness or other friend chatbot programs. People can imagine a consciousness right up until it spits out a response that is clearly not human. Then it is revealed to be simply a computer program wearing a poorly fitted human skin suit. If you look up Emily Bender's work, she's a linguist who worked closely with google on the development of Bard and she got fired along with a number of other people on their ethics team because she demonstrated that these AI machine learning systems do not work like humans do. But people are easily conned. These companies are not your friends. They're not here to augment your life, to improve it in any way, if anything they're here to crush you into an even more authoritarian state. An example would be this, straight out of the movie minority report; www.bloomberg.com/news/articles/2022-06-30/new-algorithm-can-predict-crime-in-us-cities-a-week-before-it-happens? It should not be overstated how overhyped the claims of AI companies are as well. As you said, it's startling how often they spit out results that are simply false or nonsensical, and Bender spends a lot of time on her podcast Mystery AI Hype Theatre 3000 debunking the claims companies are making. The real threat isn't some Skynet scenario like CEO's like to keep pushing to make people frightened, it's that they slowly and systematically implement these systems into our society while no one is paying attention or doing anything about it. As Frank Herbert wrote in Dune: "Once men turned their thinking over to machines in the hope that this would set them free. But that only permitted other men with machines to enslave them." These are very dangerous times and fear and feckless politicians are what keeps us paralyzed and incapable of fighting back.
you are not alone, I want to bring value back to humanity in the eyes of the world. I want people to value each other over money or their own selfish satisfaction.
As dystopian as this whole topic is, your video gave me something important... a reinvigorated sense of meaning, and for some reason, hope. I'm working on a project, I would dare to call it a work of 'art', whatever that is... and you've added an intense fuel to my fire. 'Art' really is important, especially in these complex times. Thank you for giving me a renewed sense of worth in my own 'creativity'... Subscribed.
Me too. About halfway through this video I picked up my pencil and just started drawing while listening. Capital has taken so much from us but they'll have to pry creativity away from my cold dead hands.
I am a computer engineering student and an aspiring game developer, and I feel like I made the jack of all trades of getting fucked in the ass. I genuinely fear, not only for me but for humanity in general, that AI will take away most of what makes us human, such as the ability to make science, use our logic and brain to solve problems, and express ourselves through art, that an aspiring artist won't ever get to make a drawing or painting or whatever because "LMAO AI does it better and faster", or that there will be no one doing science because "LOL AI does that already so why bother?", and that life will be only relegated to consuming AI generated content and rotting away, that a writer will never get to voice their ideas the way they want because "LMFAO AI is more efficient". I view this as a true crime against humanity, I feel like this thing is stripping humanity away of its best traits of expression, of its freedom, and I feel rather powerless and hopeless I talked about this to one of my professors, a brilliant mind if you ask me, and they say that, no matter what, human genius and creativity will not only always have a place but it will also be needed, and that humans in all fields will have to know how to use AI as a tool. I see this prospect as rather reassuring, and even preferrable, because if we get a tool that allows us to do our work better whilst retaining the need for creativity and skills, that's honestly better, but seeing this video I can't help but ask, what if he's wrong?
I cannot find the proper use case for AI. It's obviously is "so important" that people pay for subscriptions to it. But what for? Half the stuff it says, it makes up. So the only real use case is to organize existing data, altering existing data, and generating stories. So... what am I supposed to do? The only ethical use case I can possibly think of is for generating the front end template of a story, like a mystery, and then solving the mystery yourself. How are we supposed to use it?
I like to give the benefit of the doubt to tools, but I just can’t hop on because as a hobbyist artist with 17 years worth of experience, no one uses it right and generated images aren’t creative enough to give me inspiration. Not like nature, not like other poeple’s artwork.
@@E3AloeLi Honestly the only generative image AI stuff I've seen that tickled my ADHD brain into action was when it was WORSE at images and threw more vague shapes that sparked a series of random thoughts... And I was able to get the same result by literally just shouting in the offtopic channel of any discord "GUYS QUICK FIRST WORD THAT COMES TO MIND WHEN I SAY [A]" and picking the first three non-replicated words and thinking of how to mix those together.
I've tried unsuccessfully to express many of these sentiments to friends and co-workers in the recent past. What you have created here is the definitive video to express what I could not quite articulate. Thank you.
Could've just asked an AI. I love the irony that all these dick head youtubers who want to be "hated" can't even write a more compelling summary of the threats posed by AI than an AI. And then they have the balls to blame other people for bloated nonsense in a video that is 1.5 hours. Several good points that have been touched on in many other videos but fucking hell man, what an overall waste of time.
A lot of the people are already ready to give up their own mind to become a mindless drone. They think it would take away only the bad and undesirable stuff from them, not seeing that it would take everything and leave them an empty shell, uncapable of thought, emotion, action.... life.
Fun fact: AI books exist. Now, the reason that's fun and not depressing is that they were so unsuccessful they didn't even make it into the news cycle.
There is no such thing as imagination. You can remember, and you can merge those memories together, but you can't imagine something you haven't experienced or seen before. Sorry, but there is no security in that thought, I'm afraid.
@@PCMcGee1 Your claim is objectively false. THe fact that HP Lovecraft imagined the Elder things debunks your point. HR Giger and all of his works debunks your point. The fact that our mind created the color magenda debunks your point.
The last place I worked at hired a guy to do coding. Every time I walked by the dude was using chat GPT to write simple stuff in Python instead of writing it...then he later complained that the AI compiled Python UI was hard to debug...it also suggested he download about 40 packages so the program was a mess and couldn't be easily exported to run on other machines without a boat load of libraries. Weird stuff AI is.
I strongly disagree. Weird programmer that was. Meaning, if you give an AI bad/lazy instructions with no thought put into design, it's nearly the same as doing so to a dev team. The mistakes that gentleman made were completely avoidable, but it sounds as though he also didn't want to *think* about the project.
@@nev6502 the type of people to resort to ai are the type of people that dont want to think about things too hard. Any enterprise application worth its salt will only use generated code for small functions. I personally don't use any cause for the time I have taken to massage and check the code for any errors, I may as well have taken the time to write it myself and understand it from the top down.
@@stang9806 I do agree with part of your comment as a software developer myself. I don't agree with the opener of "the type of people to resort to...". Leads were saying that about autocomplete / the beginnings of intellisense in the 90s and now I know very few professionals that don't use some form of it. Personally I will resort to anything that allows me to work faster. The problem / inefficiency arises when you are able to read code much faster than you can write it (even if you can write it very fast). I've been considering this more as of late and wonder about the potential of a hybrid approach (where most of the generation is done with AI, and a human audits the result). With plenty of guard rails code generation can be tailored to various styles. Food for thought; probably not quite practical just yet but if not, it's getting very close in my eyes. One thing we can agree on; it currently requires a human.
@@GavenJr when the JXL standard is finalised webp will be obliterated. Imagine 30% better compression, thumbnails are redundant due to arbitrary progressive load through file truncation (no excuses, just cut off the data stream at data/quality limit), and perfect jpeg back compat transcoding.
I mean webp has both the qualities of jpegs and pngs with better compression than either, great for web development due to it having a low tax load on a server's bandwidth, it's just a shame that so much software has seemingly out right refused to support it for some reason.
PSA: WEBP can be changed into PNG seamlessly simply by renaming the file extension. I'm unsure if this is a feature of Win10 doing something behind the scenes, or if the data itself can be reinterpreted so easily.
Speaking of which I just got recommended this video without even noticing its length till midway. This is an important (and extremely entertaining, funny) video as it actually broadens the AI conversion to tangible real world implications that aren't really talked about much. Genuinely speechless from the amount covered, it's a definite rewatch to process what I want to say.
Anyone else notice that the AI generated captions on any shortform video like TH-cam Shorts or Instagram Reels is almost always wrong? It's infuriating.
it's almost deliberately done to make people comment, imo... People will jump in and correct what was wrong but in turn drive up "engagement" and then the algorithm will push the video even more.
Drives me crazy too. And it's not content creators comment-baiting with intentional errors, it's the subtitle generators simply being awful and nobody caring. In a world where writing "should of" is almost becoming the norm, this stuff worries me. The internet and technology should be making us smarter but I feel it's just making the average person dumber.
@@AdaptorLive the internet has unfortunately given the illiterate enough tools to skate by without learning how to communicate properly. It's almost created a simplified/informal digital English compared to actual written English. AI will be bad at implementing audio recognition while also learning from increasingly poor training data in a race to the bottom. Yay :)
"It's democratizing art and writing! Now everyone can do it!" Everyone already could, buddy, get a pencil and a piece of paper and start experimenting! You probably won't be good right away, but nobody is, it takes practice.
I studied machine learning in university but I've spent the last 7 or so years working as an engineer, programming business software. In the last few months I've seen a gold rush to build contrived implementations leaning into AI. When I say contrived I mean... often completely stupid. Most of the people singing the praises of AI don't know how it works and chalk it up to a black box, filled with magic. It's... It's gonna lead to a lot of lawlsuits. I've heard asks range from piping private user information into language models, to using AI to build a drop-down menu. The most common ask is to, "train" a model on an extremely small data set. It's dumb. What I see happening is big companies like Microsoft selling pick axes, and a bunch of ding-dongs who failed their way to the top of middling tech companies shouldering their way into the gold rush, begging to be the first to give MS money, if only to say, "we haz the AI."
Your way of writing made me read about the usage of "ask" in the English language. So you basically use the word "ask" in place of the word "question" or "request".
Ah the things you can do with LGBM or a random forest, and the people will come sprinting just to put a neural network there instead. It's all shapey statistics peeps, jeez! Get a glass of water and think about it for 2 minutes!
Yeah. Machine Learning long has had a place. Said place being mostly either for Computer Vision where a margin of failure is acceptable OR as a proof-of-concept to slowly get chipped away in regards of what is being doing by a model, till it is just a general Algoritm. Cause you rarely want something that says "Maybe it is that" to be in charge of anything critical and you often want it to be able to run on a Real-Time embedded system without affecting other tasks. But we had a breakthrough in a single sliver of AI as a whole (LLM) and suddenly Machine-Learning had to be crammed into everything for the sole purpose of tricking dumb gullible Investors into handing over big bags of cash cause AI bubble...
“in a free market, when something bad happens to you , you must be punished” almost brought tears to my eyes man . great work as usual, you’re one of my favourite writers on this website
@@hagoryopi2101 because when two people want the same thing, the person with more power has many ways and connections to make it happen, and the free market eventually just has boards of companies who only see the stock price, so the thing that they want most is the expansion and growth, this happens at every large company. And compared to normal people who still has some options, people down on their luck usually have few alternatives, as literal life saving services are also privatized, thus when something bad happens to you, you become valuable asset to companies, and not a fellow human being deserving of compassion. Just like gravity forms planets out of cosmic dust, free market eventually becomes a system well adjusted for making money at people's expense(for example planned obsolescence)
@@hagoryopi2101 When something bad happens to you, you are in a vulnerable position. When you are in a vulerable position, then a "rational value maximizer" *has* to exploit you (if they dont then they are either not being rational or not maximizing value).
@@justincenter4061 the most profitable way to "exploit" vulnerable people is to offer them some security. Instant loyal customer, most profit in the long run. Anything less than that is irrational violent "logic," the fallacious belief that keeping others down benefits oneself more than mutual growth.
effectively what I currently see being the most likely scenario is that eventually the internet and digital media in general will become so utterly worthless as a source of information and entertainment that most people with self-respect will just drop out and move on from it
Oh god no, it will create the next gen of "Conspiracy theorists" and "anti-vaxxers". It's definitely inevitable and probably already happening, but no I don't think that is ever a good thing.
Ive been considering that. Internets only been around for 30 or 40 years, really still in its infancy. As incredible as it is, it could be replaced over the next century.
Jesus Christ dude, this an actual banger of a video essay. Incredibly well-written and edited. Usually I resign myself to semi-passively engaging with someone's multi-month long obsession created into a 5 hour video, but I was enthralled, fascinated and engaged the entire time. I'm going to be thinking about this video for a long time.
Incredible fever dream of an intro. That should have clued me in that it was GPT. I have played with it extensively, and am getting better at recognizing it in the wild I think, but I still didn't detect it in a produced video, despite the incredible dissonance between the title and the minecraft fever dream. Incredible. Thanks for this essay. I wish I were smart enough to make use of it
It's easy to spot, especially if it is not given any half-decent tone prompting. It will sounds like an amalgamation of English language, and always be in the form of retell, pros, cons, selection, conclusion. Yes, just like an SAT essay (do they still do those?).
I didnt realise it was GPT but I did think he was intentionally making it seem as generic as possible to set a tone. Weirdly enough I feel like I easily have watched at least 3 ish 'minecraft nostalgia essay' videos, that read exactly like the intro, from 2019 til now
I had some kind of audio glitch with my headphones and I patiently watched the first 9 minutes in silence. I thought it was a really strange but oddly pleasing mystery to try and figure out what was going on.
Unfortunately this video perpetuates a couple common myths about the technology of AI in relation to american copyright law. 1). AI training sets are full of stolen art. In America the training sets that are essential to the functioning of art AI were originally compiled by universities and reserch institutions under the classroom use exception to copyright I.E. fair use. so the artistic & literary works in these training datasets arguably compiled for the purpose of education didn't need to be licenced. This lack of licensing dose present a problem for companies that are using them to make products. 2). AI can't produce original ideas or art. The American copyright office has touched on the problem of a "non-human creator" before now. "The 2014 Monkey selfie copyright dispute"-Resulted in a copyright ruling that works created by a non-human "Animal" creator immediately enter the public domain upon there creation. "The 2023 Théâtre D’opéra Spatial dispute"--Resulted in a copyright ruling stating that Art AI are authors of the original works thay produce, “when an AI technology receives solely a prompt from a human and produces complex written, visual, or musical works in response, the ‘traditional elements of authorship’ are determined and executed by the technology-not the human user.” AI Registration Guidance, 88 Fed. Reg. at 16,192. And "Because Mr. Allen has refused to limit his claim to exclude its "non-human authorship elements," the Office cannot register the Work as submitted. " "The 2022 Zarya of the Dawn Dispute"--Resulted in a copyright ruling that says in-part. ""Rather than a tool that Ms. Kashtanova controlled and guided to reach her desired image, Midjourney generates images in an unpredictable way. Accordingly, Midjourney users are not the “authors” for copyright purposes of the images the technology generates. As SCOTUS has explained, the “author” of a copyrighted work is the one “who has actually formed the picture,” the one who acts as “the inventive or master mind.” Burrow-Giles, 111 U.S. at 61. A person who provides text prompts to Midjourney does not “actually form” the generated images and is not the “master mind” behind them."" ---- The processes that humans naturally follow to create original art has been enshrined in copyright law, this is why we can separate. In a court of law: A copy of a painting. From a derivative of that painting. From a painting inspired by the painting. From a painting thats in the style of that painting. Being inspired by a pice of art or painting in a another artists style. Would be un-controversial if the Art AI were a human artists and not software. In short if Art AI were not producing original art based on the "inspiration" of trained images, and not simple "Copying" then the produced art wouldnt fall into the public domain -because of non-human authorship. instead it would be copyrightable. -------- Update 10/09/23 After being challenged. And re familiarizing myself w/my sources. I got a date wrong, was un-clear in places and mixed a couple arguments together. I'm making edits to correct my errors. --------- Completed revisions 10/10/23
So far, AI has made Google searches useless, spellcheckers less efficient than proofreading yourself, and art hosting sites clogged with generic, mass-produced trash. Thanks for giving us a preview of what this nightmare future has in store for us next!
There was a very similar line in a comic adaptation of Coraline I read recently. I posted a picture of it to Mastodon and THE Neil Gaiman boosted my post to his followers as well!!!
I think that people don't really understand how similar AI and Human brains are. Neural networks are literally based on working of Human brain. AI art is currently similar to visual imagery. No one conjures something utterly unique out of thin air, It's all inspired by previous works and their surroundings. It's weird to me how people call AI unoriginal when it replicates or tries to go through processes human brain undergo to develop a thought or visualization. There is no good or evil in AI. Tolkein, For all his virtue, was a rather conservative man bound by the frankly limited ideal framework of his time, unpossessed of ny real expertise or insight into our modern world, Indeed, he could not come to terms with his own contemporararies at times, as with his dislike of Dune. Not to detract from his achievements, Mind you, but one must take that into account when applying his sayings into a unrelated field.
If there is no evil in a neural network, and a neural network models a human brain, does that also imply that there is no evil in a human brain? It seems to me that - like humans - AI is capable of evil but isn't necessarily evil by nature (AI can produce good as well as evil). Tolkien saw a distinction between true creation and what he called "subcreation". True creation is e.g. God creating matter from the void. But anything a human does, like writing a story or making an AI which makes a story, would be "subcreation" - creating something novel out of what already exists. Tolkien saw writing stories as "revealing" them, not truly creating them (it's fascinating, read his Essay on Fairy Stories to learn more!). I believe the quote above is speaking of true creation more than subcreation. It explains why Morgoth and Sauron could not make their own races, and instead corrupted existing races (elves, men, ents) to create "new" races. But I don't think Tolkien would apply it to our modern AI. I do think he'd detest our modern AI, though!
This video does a great job pulling away the wool that is over your eyes. Anyone scrolling down because they lost focus on the video should scroll back up and finish the video. The last part about monetizing your data and using it to incentivize "good" vs "bad" behavior is scary and real. I have added this to my favorites playlist. Also, kudos to the creator for taking his own advise and adding subtitles to his video. The crappy AI generated subtitles implemented by TH-cam are a perfect example of immature AI technologies injecting themselves into our daily lives.
Conflating the censorious, privacy-violating mega-conglomerate corporations with the incoming AI boom that they are leveraging against those that rely on (not merely use) their services isn't elucidating anyone. Also I haven't seen to the whole video yet but does he address the fact that not only is the majority of this research being conducted in the open but the industry is finally addressing the hardware problem with specialized accelerators, such that running very large models won't be the exclusive realm of quarter million dollar machines like the DGX A100? And then what?
not enough data put simply, if you're listening to someone with accent its hard to understand him the same goes for ai trying to understand words, if it doesnt know it tries to come up with words that are sounding somewhat similar when the bro in the vid speaks too fast trying to find the relatable word becomes impossible with the hardware it runs on. its not "immature" the models are actually quite advanced main limiter is and forever will be hardware and how much data there is. even a bad language model you built yourself could most certainly understand you (you as in you when youre trying speaking to it) but maybe not others because they speak different use, different words.
No, it doesnt, it makes you think you learned something while in fact you were lied to and decieved, absolutely baseleslly. Its horrible how people can be so easily swayed without asking for further sources or facts.
Another very good video... You have a very evocative writing style. "A world for money to live in or a world for people to live in" is a line that I think will stick with me for a while.
Just discovered your channel and I think this video was one of the best on AI (and being a professional aritst, I watch A LOT of those). Yoir approach is very different from other video essayists and there are some points in this video that I've only seen you make yet! It's given me so much more to think about, reasearch, and explore. Hope your channel keeps growing cause you truly deserve it! Keep up the great work.
Ah, yes, pleasing the stakeholders is ALWAYS something that steers societies away from dystopia! Thank you, AI alignment "non-profits," for letting us know you're helping the little guy!
AI alignment is an important problem to solve, our entire race and future depends on it. Please go and look up Robert Miles AI safety to understand what im talking about.
@@TheManinBlack9054i will filter my consumption. I will listen to an expert whose analysis I feel I can trust (Jimmy Mcgee) and ignore charlatans with an agenda (random guy two levels deep in the TH-cam comments)
28:50 I have seen so many people talking about AI who overlook this exact point. Art is all about those little decisions that capture the artist's voice. I couldn't have put it better. Amazing video.
Ironically, it's that same point that relieves me of the idea that AI could ever fully replace artists. Someone who becomes completely reliant on AI may certainly use (unknowingly) the same techniques as an expert uses, but the lack of understanding why the techniques work will be wildly inconsistent compared to someone who gives a damn about art.
When talking about writing, and AI being a threat to the field, one anon described AI as being unable to write, because it does not understand, it puts together sentences shaped strings of words.
@@MrhellslayerzI nwver thought it would ever replace artists but it will make it a lot harder for new artists to get to this point and support themselves and there will be an ever decreasing pool of actual voices to listen to, because the less viable it is to enter this industry the harder it will be to get enough experience to be able to make these choices effectively.
@@bluester7177couldn’t agree more. Very few people finish school as highly polished professionals. Most creatives in any field have to hone their crafts with lower level, simpler work, which right now there is a market for. But if people cant sustain themselves doing this sort of work anymore, then the pool of people who can afford to become artists will shrink to the independently wealthy and those with rich parents. Ironic that the people who claim to be “democratizing art” are steal real artists work to do the exact opposite.
That may be true, but how many people look out for it? Oh no, an indie game used a little bit of AI art for its backgrounds. The sky is falling, I tell you!
Fantastic video! One other thing I feel is important for the future is directing people to user curated sites where real people (even if most aren’t professionals/experts) help answer questions about different trades like how to fix very specific computer issues, car troubles, or how to do your own taxes. Most just accidentally find these (Reddit posts come to mind), but like Jimmy mentioned, Google is getting worse and worse about giving you a good answer to a search, so making sure people know how to find the truth is critical to surviving (that or some talented person makes an AI tool that can instantly filter out AI generated content).
A GIANT thank you for this video. I am in full agreement with every bit of it, and you touched on many thoughtful and critical points I do not see being brought up at all in these AI video pieces. Thank you especially for the photographer analogy, and mentioning the role of human intent- the element that can only be mirrored but never truly replaced; it's the piece that makes art connect people, and endure. Fierce pushers for AI arts take the creative process for granted, stripping the meaning of the mediums down to literal minimums (prettiness, cost, tech). If you value anything beyond that, they call it "human hubris." I think this attitude stems from ignorance, combined with hyper-focus on one area of tech, and delusions of grandeur. Sprinkle in a disregard and even disdain for those who are stolen from. Which... strongly reminds one of NFT's, and those who insisted they were visionaries pushing crypto as the future currency. People must stop using a vital limb of human culture as their playground for automated blandification. Imo the competence of the AI (ex the extra/missing fingers) is not the big problem, here. The big problem (besides people's literal livelihoods being cut off) is something you mentioned in your video: As a professional, I am seeing a startling lack of care, less skill, and ALARMING decrease in critical thinking beginning to rise in my creative field due to people using AI as a shortcut. This is a bigger problem than people may think. This is an attitude that should never be normalized, especially not during times like this. It's an attitude that spreads from producer to consumer. It encourages people to become militantly dumb, as though asking for more from the media we so eagerly consume is somehow absurd. Liking something just 'coz is actually just fine. But AI art presents a hard wall that halts active analysis and discussion. You watch a wonderful movie, then you find out it was garbled together by AI. What more is there to even say about it now, besides discussing the novelty of that technology? There is no part of it that was trying to connect to you. And how ridiculous is it that some people demand high praise for typing in prompts for art that's a mash of often stolen work? That makes me sound like an asshole, but what else can we call it? That's like someone punching buttons on a vending machine, and demanding to be called a chef. They can't seem to decide whether AI art makes it "easy and accessible for everyone," or if it's a special task that requires skill. They will change the narrative to suit the argument. And no, gaining something meaningful from a film/painting/writing/song etc. is NOT the same thing as finding the beauty of a tree meaningful (something humans did not 'make.') Nor is it even remotely the same as 3D technology. I am tired of hearing these desperate analogies from people who are so bizarrely enthusiastic about devaluing human expression, and putting professionals and creatives of all stripes and kinds, out of work. They do not speak for us. If this AI craze (not just for art, obviously) is really a bubble as some say, then I cannot wait for it to pop, and I'm sure I'm not alone in feeling that way. There MUST be better ways to apply this technology. Big breath okay I'm going to sit outside with my dog... But I'm very happy and grateful that I'm seeing more videos telling this side of the story. It's the side that isn't backed by billionaires. And I'm ESPECIALLY happy to be seeing artists try to hone their skill and bring their personal expression into a space that often dismisses them as redundant at best, unwelcome at worst. Your personal voice matters. I'm especially happy to see new artists. Please don't let this AI stuff discourage you.
The part where you talk about search results on youtube being mixed with useless recommendations hits me like a truck: I was trying to find your channel from that one Cruelty Squad video (since I forgot the channel's name, whoops) and the search results were mostly unrelated nonsense I watched in the past..
Being a software engineer working closely with ML, I quickly noticed that most jobs on the market were related to recommendation systems and other ways to manipulate human attention. So wrote a small assay, long lost, about hypothetic future. In this future timeline, most AI applications of our time is considered extremely barbaric and unethical because the culture has evolved and people see it for what it is. So, just like we have "medieval torture museums" in a lot of European cities nowadays, those better future humans would have XXI century AI museums. Such a museum would have rooms or whole floors dedicated to different AI abuses. Maybe a whole floor for mobile games and microtransaction optimization with AI, a large section for recommendation networks to sell you stuff you don't need, and then walking further with your museum guide you enter a large hall with a Facebook timeline projected on a wall, scrolling down without ever stopping. I wrote that quite a while ago being frustrated by social networks becoming more and more shitty by becoming better at that stakeholder value thing. The essay was also quite optimistic in describing the future where we've collectively understood how to use the power of ML for good, at least mostly. I no longer have that optimism in me though. We're used to now owning things by now, and mostly used to not owning our personal data for convenience, and that's bad but manageable. It's much more frustrating to not own your decisions, to be manipulated by neural marketing and gaslighted by statistical objectivity of "but that's what the algorithm decided based on YOUR actions". We don't have the discipline to watch our every move and every action just to project a correct image of ourselves. We switch between the state when we make conscious decision and the state where we let go and let our inner monkey have fun. The marketing department is only interested in the second state, and they're eager to sell it to you as your true self.
That's a very insightful comment my friend. I think you are probably quite young and still might have some faith that humanity progresses towards something called evolution. That's a nice concept created some 200 years ago, but it turns out that idea is just another marketing tool to justify colonialism and race superiority. The truth is that the vast majority of people are just well meaning fools exploited by clever crooks. AI is just another tool for the perpetuation of that status quo.
For some reason I’m kinda happy you are all getting so frustrated with something that is completely under your own control. Bunch of addicted losers lol
As an AI engineer myself, I feel obliged to acknowledge every concern you've brought up. It's like I've got a ton on my mind, but saying it all out loud feels risky. I worry about getting laughed at or pushed aside, or even dampening the vibe and being seen as a downer. I can't help but cringe when I hear bigots and parrots praising the AI Lord and all its blessings. But hey, here I am, just riding this crazy boat, not knowing where it's headed.
29:08 "... when you create something, you're constantly making tiny choices and I think an author's voice is a sort of harmony and consistency between those choices." This quote perfectly encapsulates how I feel about AIs in creative fields. Intentionality should be a key aspect of any piece of art, so it makes sense to use AI when randomness is the intention (the movement of models in Rain World, for instance). Otherwise it becomes a crutch and leaves what you're creating feel vapid and uninspired. But at least I can become a fucking prompt engineer.
@@awsomebot1 Honestly, yeah it does. How much it matters is a little more dynamic, its like the frog in the pot. You cant put a frog in boiling water or it jumps out, so you put it in a pot of cold water and then turn the burner on, by the time it realizes its being cooked it no longer has the strength to jump out. We could see something similar with AI, and we are: its slowly creeping in, generating low quality media for consumption. Not in such quantities that it causes an outrage, but it will get worse. As AI media becomes more and more normal, we notice it less, which allows it to penetrate further. Eventually we're left with very little except for AI generated art of all forms, and it will all suck, and still nobody will care because it happened gradually enough to avoid a backlash. It matters because right now the water is still lukewarm, we can still jump out of the pot, but not for very much longer.
But they are, and let’s not pretend that mega-corporations like Google and Amazon give two slivers of a shit about us, look at how they treat their *own* goddamn employees.
This video is fantastic all around. There's so much in it that it's hard to comment on anything too specific, but it was very solid from start to finish. Thank you for making it.
something's wrong with my copy of minecraft
People asked for it, so here's the music from my videos:
owenc.bandcamp.com/album/jim-mcgee-pt-1
_"People asked for it, so here's the music from my videos"_
AWW YEAH!!!
bought the album, thanks for the great stuff
Oof, that was a HUGE gap between "free-er the market - free-er the people" and the socialist message at the end. I honestly thought you had a political schizophrenia with "we should regulate the market, but free market is actually better, as my [totally not government enforced] mantra says". Good job man! Really hq vid you got there.
@@Duvumvirat your listening comprehension is so faulty lmfao. Wanna share time stamps of parts you feel contradict each other ?
@@Duvumvirat bruh
_"The irony that we're automating the production of art instead of the jobs everybody hates shouldn't be lost on us."_ the fact that I and all my cynicism didn't cath this 'til you said it, leaves me feeling all different kinds of bad
To be fair I think those jobs that we dont like are the ones that most of the population are employed in so if all of a sudden they wanted automate all the unwanted jobs people would be in uproar under the threat of unemployment by automation. But I dont think people will care very much if its just a sliver of the job market thats being automated such as artist , this is also compounded by the fact that more people benifit from ai art than some higher ups that benifit from making somthing like autonomous machine that makes production more efficient.
For the average person using AI to make art I suspect it's a kind of creative expression they didn't feel able to indulge in before. Art is hard and I've known many people who've given up on it. I don't hold AI use against these folks.
On the other hand, people and companies who are using AI to save a buck and cut out artists? Fuck them. It's already difficult enough to support yourself making art.
It's almost like full automation is an unrealistic sci fi meme and this video never actually talks about that aspect other than stating stuff that everyone already knows and getting angry at generated images as if the ai is taking away jobs from creators (its not).
Really, so long as a robot cannot even fold a piece of fabric that is given to it, something that proves to be exceptionally hard to do, a chimp will be more useful in a factory than an AI.
Looks like we are all just going to have to keep struggling in this process we call life like our ancestors before us. Maybe that's not so bad.
@@dreyri2736but It is taking jobs from creators, there already book covers who have been done by a human previously and now are AI generated, brand design, anime backgrounds, commissioned art, there already stories of people training Ai on the Art of an artists the previously commissioned and in the future it will take more jobs, specially entry level jobs, I agree there probably won't be a fully outmated future any time soon, but automation always have a downside and people often struggle to pivot and sometimes can't after their jobs are outomated out of existence and capitalists are always trying to increase their profits, so if they can fire some employees to cut cost even if AI isn't perfect, they will do it, they already done with most of phone and online costumer service and it's awful.
I'm alarmed by this use of "cath" as a verb that I haven't previously come across.
"The fact that I and all my cynicism didn't"... catheter this? What?
I'm a huge AI skeptic, and as a software engineer, that's not an easy line to walk. The best sell that I've seen for AI, is that it will function as a copilot for all the shitty, high time consumption, low impact tasks we need to accomplish on our day to day. It shouldn't take away anyone's job, it should give them ability to focus on more meaningful aspects of their job.
When you examine that more closely, you realize that it means that AI is going to take away the jobs of low level employees, that need to start off as basic, entry level people and work their way up to challenging, advanced problems. Where will this leave the industry in 20 years, when the junior count has declined so drastically that there will be no supply of expert level people? Is the bottom line more important than helping someone grow as a professional and enriching society as a whole? Rhetorical question, obviously the answer is yes.
It sucks, on a personal level, to have to walk between raindrops to avoid unethical, shitty high tech companies. Even companies that try to make the world a better place might end up doing it in the wrong way.
I never understood this criticism of A.I. I mean farm equipment replaced 90% of farmhands but I don't think people are complaining about tractors
@@playerslotavailable3810 I don't see any industry growing in terms of demand to such a huge scope that it'll allow the current amount of jobs to stay the same while also providing massive automation tools. It's just not the same as farming, which needed to be automated to allow for humanity to survive in scale.
@@KyfowThe printing press.
@@sxnorthrop Do you really think that there will be such a massive increase in demand for production, that it'd go in line with the increased productivity that'll happen in the AI revolution? I just don't see it. I'll be happy to be proven wrong.
@@Kyfow I think I may have misunderstood your original point. I thought we were comparing technology that created a staggering transition in the economy due to it's demand. That, if looked at consistently, would show that society had a very short "grief" period for the countless lost jobs and livelihoods of illuminators and copyists (binders, bookmakers, etc.). Same will be true for ML (in my opinion). That being said there's really not enough data for anyone to make a confident assertion that ML will have the same or comparable impact. I was simply providing a hypothetical comparison to something that showed that the "bottom line [was] more important than helping someone grow as a professional and enriching society as a whole [through creative expression]", and yet today we don't think twice about how it affected multiple generations of people who experienced turmoil as a result. Every pivotal moment in history has some groundbreaking technology or discovery that was never thought possible before and it mostly ends up just making us better at what we already did before (not commenting on whether that's right or wrong ethically).
TL;DR, I'm using a past example, not a current. The printing press is a perfect example of technology that destroyed lots of livelihoods (and was driven forward in-part by 'money hungry business moguls') but we don't care about it at all. It created a shift in society comparable to something like the AI revolution would create.
The future is bland.
its already been bland tbh, seen those ugly 2010s gray on gray interior design styles or those hideous 60s concrete british council towers? but i feel like even if ai takes over artists will still prevail, its not a simple industry like many other lost positions, its a exhibition of someones creativity.
David Firth speaking nothing but facts
Well at least you've got your rusty spoons.
David fifth 😂😂😂😂
And spoonless
"Day by day, we live more and more in a world made *by* humans, but not *for* humans."
It's a line I reiterate often, and the fact that AI is being used to replace the creative pursuits that people *want* to do instead of being used to mitigate the amount of busywork people are tasked with doing instead of being able to enjoy their lives-- it truly paints a picture more clear and vivid than anything the machine can put out about how grim the future looks if no one rises up to slay mankind's new self-proclaimed gods.
It would be a better world if generative AI and anything resembling it was banned outright.
@@Pvinini Banning won't make it go away, unfortunately. It would just make it even harder to control.... As history illustrates with all attempts to ban things that there's an active demand for...
Made for WHICH humans is the question huh?
Businesses will pay more for tech that reduces their costs than consumers will for things that make their lives better.
@@rvantong Eugh, oof, *oof,* I don't like how true that quote is.
The danger is not that AI will become sentient and rebel against humans, but that it will be applied as though it were capable of human judgment when it is not. "The AI did it" will become the go-to excuse for all manner of human incompetence and deviousness, just as the magical "computer error" of fifty years ago.
Y2K
That is a danger and the danger that AI will become sentient is also absolutely a danger
@@danielblank9917 They're two different dangers, and one is vastly more immediate a concern. While it is important to learn how AI algorithms reach the decisions they do from their sampling data so we don't accidentally make a sentient AI with unintended behaviours, it's mostly a hypothetical pre-emptive problem. We have no idea if or when a truly sentient AI will be built.
On the other hand, people trying to offload culpability onto non-sentient AI is a real problem we already see.
@@danielblank9917 no, that is literally the least dangerous thing about AI (because it can't happen)
@@Jonas-SeilerWhere are you taking this confidence from? Do you think it will not happen in a near enough future to care about or do you think it can't happen in principle? Also, AI doesn't need to actually be sentient to be dangerous, being more intelligent than humans is already enough.
The word "everypony" out of absolutely fucking nowhere hit me like a truck.
Could not have been more out of nowhere lol
/) *brohoofs*
(\
A welcomed random aneurysm
That feeling of reading the comment as the words are spoken in the video
19:54
I'm reminded of the observation that the problem with building a bear-proof container is that the smartest bears and the dumbest humans overlap. We're getting (or already at) a point where this is true for AI.
The term "Artificial Intelligence" has been thoroughly ruined by LLMs and Image-Generators. I can't talk about the behaviour of units in a game without people thinking now that there are hooked up to ChatGPT somehow.
I swear we're gonna have to start saying "behavioral code" or something like that
@@Hernos28 bots still works, NPC as well but that is a broader term overall.
@@theflyingspaget No, "bots" and "NPCs" are the TYPE of entity. We're talking about the BEHAVIOR of the bots and NPCs when we use "AI". Crap like "if player is seen, chase player. If player was not detected, patrol from point A to C to D to A, repeat until player is seen."
swap to calling them computer units or cpu
It's all AI... if people are stupid and don't understand that it doesn't mean the term is wrong. In fact I would argue that generative AI is much MORE AI than video game AI... which is mostly simple scripts (very few games use actual machine learning for their "AI")
This video reinforced something ive been feeling for a long time. The internet doesnt seem fun anymore. Everything is funneled into narrow channels of blandness. Its driven a deeper desire for me to unplug as much as possible.
the problem is millions of americans still live in poverty in shit cities . there's not much to do irl where i live due to the heat aline
@nostalgiatrip7331 I agree but being online doesn't fix any of that, it's just an escape. I escape too so I'm not judging but it just feels less and less enjoyable
The internet is going through the same bureaucratic money-wringer that pretty much all public spaces went through before the internet existed. Cafes, parks, sports areas, town squares, etc. now require money upfront for access to their space and/or provide a bland free space and a pay-to-access fun space and/or are *technically* free but find ways to heavily discourage use without payment. It's part of the reason people fled TO the internet in the first place; you could generally do whatever you wanted, whenever you wanted and as long as you wanted, for free. Games, media, hobbies, socializing, it was all completely untethered to money and the real world.
Now that enough businesses and corporations have figured out how to monetize all the "public spaces" and "free-use services" of the internet, it's just one big, digital, global mall-city, where you can either pay to do what you want (directly through giving funds or indirectly through selling personal data and watching ads) or you can sit in a corner and frown for eight hours. It was only a matter of time, but it sucks that it's already come to pass in our lifetimes.
What's stopping you?
@@nostalgiatrip7331Well, be consoled by the fact that some would actually love to be able to live and run around naked in a warm sunny place... but instead they must be dressing up and wondering how to pay the heating bills in a climate where half of the year it's wet and dark outside.
1:00:00 holy shit you have literally articulated this one thought I've had stewing in my brain for the past few years, the modern internet is becoming a conveyor belt for "content" rather than a collection of things you can look for.
i've been noticing this trend with tiktok as well, your brain is literally being conditioned to not put in the slightest bit of effort to get rewarded and in a generation or so we'll see the horrible effects of it rear its ugly head once all the iPad kiddies grow up. Even older folks aren't immune, both of my parents are nearing their 50's and they use tiktok every day. The worst part is that this is literally what some companies want, and whether or not they're cognizant of that fact this is just what the market is going to do, make you consume the most slop at the lowest cost to them.
@infiniminer7677
Good point. I think people don't understand that they think so called "free market" wants this and so it gets what it wants. What they aren't considering is that people's wants are being manipulated by the Edward Bernays market psychology even more so now with technology. What do you do when the people can be programmed into believing that their choices are there own when all of the sudden they are not?
@@infiniminer7677 well, tiktok is garbage, so there's your problem!
Kind of like television, which I ejected from my life a long time ago.
A common term to call this is "content sludge", if you're interested in additional reading lol
I’ve been seeing a lot of AI on crochet subreddits. I didn’t expect to see AI leeching into those places, but it tracks. There are AI-produced patterns, which is the same kind of thing as AI-produced recipes or anything instructional. At first glance, it fits right in among other patterns, but then you get halfway through the work and Bluey has a mile-high forehead and the pattern quality just begins to degrade. Whoever is training the programs hasn’t been able to make it spit out images with accurate or natural-looking stitches yet, but it’s only a matter of time. Naturally they started with what seems like the most profitable area of crochet, amigurumi. What I’ve seen isn’t them selling a non-existent finished product, though. I haven’t looked but I’m sure there are paywalled AI patterns out there too. The grift for now seems to be getting people to the website and throwing a ton of ads on them, which is already a thing in crochet. There are tons of blogs and aggregation sites. It’s a lot faster, I assume, to cut out the part where you look for patterns to share and to just fill your site with fake ones. Then make more, which all refer to each other. It’s all the same as other content aggregation sites, but it just feels so out of place to me. I mean it all feels out of place, but finding patterns is already a time-consuming task and these AI-based sites add exponentially to the problem while masquerading as a solution in the form of an aggregator site.
Idk this whole comment is just a rehashing of the same bones of an issue with crochet skin over top haha
As a software engineer, and as a comic book illustrator (with actual pen and tablet and time to hone my skills), this video hits the nail on the head on everything wrong in the industry.
Maybe we should ask for an "AI" that can do all the tasks of a CEO so we can free up billions in capital by eliminating the need for executives. I wonder how fast laws are past after that hits the market.
It's already happening, sort of. Problem is, it's China. We don't know if that one Chinese company that replaced their CEO with an AI (with obviously a human liaison) is real or just a CCP fluff.
imagine co-ops with AI CEOs. As dystopian as that sounds (to be working for a machine), the awesome part might be seeing what a companies pay structure looks like without having to pay the CEO 500 times what the rest of employees make. That seems like automation that could save some money right there.
@@jasonmast7769 if it's training data really comes from a bunch of internet comments, it would probably redistribute the wealth among workers immediately.
@@nathanrocks2562 It would probably kill them all and replace them for being inefficient. Wealth exists to grow more wealth.
CEOs are the grounds keepers of the capital investment upon which companies are founded. CEOs are eliminated all the time. But the only reason CEOs exist is because the people who fund the company to begin with want them. CEOs are not required but desired. No company is forced to have a CEO if the people who fund it don't want one.
“With new technology, the line between research and commerce is razor thin, and big companies often use this fact to just ‘manifest destiny’ whenever they want and make us live with the consequences.”
Jimmy, you cannot hit us with a line that hard less than 10 minutes into the video.
Would you prefer that progress stop in the meantime? What's the AI going to do that's so horrible? Jack up insurance rates when the health insurance industry should be nationalized, and we need more transit and bike infrastructure so that people don't need car insurance anyway?
manifest destiny was an objective good imo.
@@Ilyak1986 Well you automate one thing, then the next thing, then the next and before you know it the only jobs left will be AI maintenance. Not everyone is capable of that, so then you deal with a jobless problem, next a homeless problem and next a society collapsing problem. Even if AI could eventually think for itself, we will never get there because the creator's society will collapse before we can get there. Ironically AI is not the problem, but the missuse of AI trough human hands and corporate greed... That is the problem.
@@Ilyak1986We could you know, demand more social responsibility from corporations and politicians who allow them free reins
@@Ilyak1986 "What's the AI going to do that's so horrible?"
Just some things that AI already done:
If you know how photoshop already was a problem in the past for creating fake images that were used and abused by fake news you will understand that AI image generation will be this but 1000 times worse because anyone can put the name of a real person there and generate many images that to most people will be a real photo, if AI was ethical it wouldn't generate images of humans at all because to surprise of no one is already being used for malicious end, to basic porn bots to revenge porn (and if you go to the underbelly even have the "great" children variety that already trying to push as normal and ok because no kid are being harmed).
AI misinformation makes hard to find true sources and is destroying Google* and is already being deadly with the misinformation because has books about mushrooms with the wrong information so people that buy it can end up eating poisonous mushrooms and the one AI from a supermarket that made a recipe of a deadly gas as good ways to clean the house.
*Look up "how many african countries start with K" on google and now remember that most people only read the first result
The AI therapy tool/friend/girlfriend stuff really gets to me. Our most fundamental needs are being exploited at every angle.
I genuinely never thpught there were many people THIS desperate.
Fuck, I'm lonely and possibly depressed as all hell and I would never stoop to something this... sad. Dehumanised. *Artificial.*
The fact the market for this exists alone is depressing...
@@Box-O-Soldier It's super depressing man. I made an oath to my self I'd never get that low. Best thing we can do is keep ourselves busy I guess
@@backoffpeer Nay, the best thing we can do is hold ourselves to a standard of our own choosing, keep our "spine" and integrity intact.
I just wish that more people chose this instead of brainless complacency.
@@Box-O-Soldier I'm the type of guy that watched Blade Runner 2049 and thought to myself "Dang! I wish I had a robot girlfriend." ngl life is bad enough I would buy that shit
Yep. I believe there's an AI service out there that charges quite a bit of money for a chance to chat up with AI, it's so scummy.
At best I'd really only ever use AI like this for fun and leisure. I'm fortunate to already have people in my life to connect to, but not everyone has that luxury. . .
As a young person, it's super cool to have no future.
Don't say that to yourself, you're God after all and your words matter.
May I interest you in global warming?
@@VoidStaresback AI will solve global warming.
@@ModernCentrist I look forwards to that
@@ModernCentristdon’t rely on that tho. Even if that miraculously happens, we need to do everything we can on our end. Relying on AI like it’s some savior-god isn’t enough
The more Silicon Valley types say "We want to build systems for humans" makes me feel like they aren't fully human.
The design is very human
You feel right. They are the bastard offspring breeds of man (women) and fallen angels (demons). Satanically animated machines. And they even go as far as claiming to be God's chosen! Exponentially more amazing still is that the very people that say they love the Most High God (instead of rightly fearing Him) actually believe that to be true! Once you notice them, you will see they do business everywhere. Many - many great men and painters throughout history have; and warned about them.
@@petarhandle The-design-is-very-human guy was warning us all.
Ironically, taking away people's humanity is a very human thing to do
you know its like the reverse. we should leave the nazis in europe and north america and we can escape to south america.
The tuberculosis story reminds me of Clever Hans, the horse that could do math. He was believed to be able to solve equations and answer trivia. What was actually happening, was that the person asking the question or the audience watching would often react when the correct answer was suggested. He would then stomp his foot on seeing the reaction, signalling the correct answer. He was unable to do this when the asker/audience had no idea of the correct answer, which is how he was found out.
Same with the gorilla that was "taught sign language" heh
@@noisyash4234 I am, and no she wasn't :)
th-cam.com/video/e7wFotDKEF4/w-d-xo.htmlsi=WdPR2n_UHn2nnkug
It was just another scam for government grants :)
it was actually when hans's trainer was unaware of the answer or not present. the horse was picking up on the trainer's body language, not the audience 😊
In academic literature about explainable AI, they actually call these faulty models Clever Hans predictors haha. There is a lot of research done to try to understand, when these models are right for the wrong reason. For example search for: Unmasking Clever Hans predictors and assessing what machines really learn
It's good to realize that the "AI is going to fix everything" and the "AI is going to pull a Skynet" opinions are mostly fed from the same sources. So long as you're buying into the premise that AI hitting a singularity moment is the goalpost, you're going to miss the ways its hurting people right now.
From people like Zuckerberg, Musk or Altman who practically doubled their wealth during the pandemic but now flush jobs down the drain in the thousands
but hey u can make porn with midjourney so the techbros will just shrug
You have got to go through hell before you get to heaven.
Any tech revolution is going to hurt people. You think computers and the internet have not hurt anybody?
@@christopherkrause337 Progress? This is 10 steps back
@@christopherkrause337that’s what AI is worth to you? _everything?_
I just love how a billion dollar industry is built on a system where we don't really know what it's doing.
Even though we made it.
We're so cooked.
seeing how everybody is hungry and intense towards work opportunity is rather sad,after all its about money.
Wrote a grad thesis on ML in aerospace. Nothing depresses me more than seeing how corrupt, toxic, and disenfranchising the field of AI has become recently. It’s so broken the only real solution is to let it die under the weight of its own expectations, since the math doesn’t support what it’s doing now all that well.
It's been a relief to hear experts are feeling the same way I am. I don't think we're going to get out scot-free, but OpenAI could at least go out of business once the hype really dies down. Little victories.
1) loved the video congratulations on the stellar job you did! Really like the editing style and composition!
2) without launching into a tirade and linking a bunch of papers, i optimistically give it ~10 years before it fades completely. Again, the math just doesn’t support the applications rn.
It's a shame, really. It's all very interesting stuff but silicon valley is trying to make it the next big thing, making folks think it's more than what it is.
Brhu, is just the next push on technology, digital field
Its the same as when robotic arms where implemented in industry line
Make job easier and more standar, better quality.
Also made a lot of jobs useless.
So obiously corpos are gonna capitalize on it
This is not beause cof the technology, but the nature of capitalism itself
Money matters, anything else is secondary with human expenses at the bottom of said list.
As money is power, and power let you do watever you want.
@@nathanstruble8587 May I ask how Mathematics doesn't support the applications?
I want to add something from my field of expertise -- computer graphics. A couple years ago there was this so-called breakthrough in this field called NeRF (Representing Scenes as Neural Radiance Fields for View Synthesis) that's basically a deep-learning-assisted rendering algorithm that produces a neural representation of a 3d scene from a bunch of images. When its first paper was introduced, it was pitched as the biggest breakthrough in the graphics and supposedly it was expected to replace all conventional rendering because of how powerful it looked. Unfortunately, at the time its inference was about 100x slower than what it needed to be in order to be usable in realtime.
Fast forward 3 years later, that technology has finally advanced to the point where its performance is almost practical for realtime applications -- its performance has evolved and now it's about a 1000x faster (in the newest iteration called Gaussian Splatting) than the original version. Except one detail: there's abolutely nothing "neural" left in it. Literally nothing at all, and it's the only reason why it's becoming practical now, because the main bottleneck of that approach was the neural network inference and getting rid of it made it 3 orders of magnitude faster.
In other words: The stupid AI found a creative solution to a problem that humans didn't know before and indeed struggled to understand how it even worked before completely dissecting the AI blackbox. :D
Yup. My field is embedded hardware engineering and for a while, we too where looking at having to add AI inference in embedded systems.
But much like with NeRF. Most of the planned uses just ended getting piecemeal replaced by regular algorithms and switch statements. Cause inference is expensive and often required some extra Neural processor like a Neural Compute-Stick to get anywhere fast enough. So slowly more and more tasks where being removed from the Neural Network part until eventually there where no actual models at work. Just regular algorithms that where small and efficient.
The only thing we still use small neural models for, is computer vision. But that is so old and heavilly optimized via specialized libraries like OpenCV. That you can do it with any half-decent ARM system like a RPi zero.
On the one hand, this sounds hilarious because it makes it sound like progress was made, and then to justify their existence, they made a normal renderer that actually worked.
On the other hand, regarding what actually happened, getting an AI to brute-force a good solution so it can later be chiseled out sounds like a very 21st century career.
@@stevencurtis7157 in this case AI didn't even bruteforce anything useful that got eventually "honed out". The actually novel and practical part of NERF(and many other DL "advances") is to formulate the task in a differentiable form so that it's _possible_ to optimize it at all with a "black box" optimizer of some sort. But as it repeatedly turns out, neural networks are just a really sub-optimal choice for such a "black box" optimizer, and there are other task-agnostic optimizer methods that perform simply better. So the big idea is to approach a task from an optimization standpoint (which was arguably never tried before NeRF in the rendering field), and that is the novel and important part, and neural networks are just one sub-optimal way of accomplishing it.
Kind of a bad take take. NERF got useful after instant ngp, and I have used it in robotics. With Gaussian Splatting, yeah, the neural network architecture got removed, but it still is machine learning because it uses gradient descent to the core.
"i don't care about mars. i want my friends and family to have the security to live good lives and a safe, quiet place to do it in." who doesn't? it's possible with our technology and resources, it's just that the execs are willing to do everything in their power to prevent it and for what, endless profits?
thank you so much for making this vid, really thought provoking
I have a single course degree in machine learning from Stanford University Online, which while it isn't a credential that will get me hired, it DOES mean I've gone through the process of actually understanding what modern AI is. I spent a lot of time doing extracurricular experiments to get a good intuition and feel for HOW these statistical models we call AI actually behave, and what their limitations are. So while I may not have a PhD, I can confidently say I have more hard technical understanding and insights about AI than 95% of the population.
The things you laid out in the first 12 minutes is EXACTLY what I try to tell people who say shit like "just give it more time, and AI will explode" or "I'm afraid but excited for the upcoming AI explosion". There is NO MAGIC, it's JUST A MACHINE WITH A BUNCH OF KNOBS. More data, more time, bigger models, bigger computers, quantum breakthroughs, all of that is just like polishing a turd! Yes, AI in it's current form can do a LOT of very useful things, and we absolutely should be using it to replace annoying, menial tasks like removing backgrounds or fixing lighting in photos. But I'll scream it into the clouds if I have to, NEURAL NETWORK BASED SYSTEMS WILL NOT BECOME ANYTHING REMOTELY CLOSE TO GENERAL PURPOSE. THEY WILL NOT BECOME SENTIENT.
I really, really wish it were possible, I would also like to see some crazy AI future, even if that concept is scary! But if you understand HOW these systems work, then you can see how it's literally not possible for them to do these things. It's like saying with the right fuel, your car could drive over the Atlantic. Understanding how nodes work, how layers work, and how gradient descent works is fundamental to realizing this is true. While modern systems have lots of advanced features like nodes that remember things over time and complex systems made of dozens of smaller task-specific models, those things don't change the fundamental limitations of dataset availability, data quality, and the fact that the system is at it's core trying to make a number smaller. (Any time you tell someone to optimize a handful of numerical metrics, they will find "cheats" to technically fit your requirements, and AI does this too.)
This video restored my faith in humanity a little. I'm sharing that segment around. You earned a like, a subscriber, and a binge watcher.
Best comment on this video. By far
The fact that such a thoughtful and insightful critique of this video, seen by 1.2M viewers, only got 28 replies is telling. "There is no royal road to learning." Euclid and others....................
What is your point, actually? It is not clear from your message. It can't be "if somebody knows how the net's elements work, that means sentience can't exist," because we agree that sentience exists, and it is understood how neurons work in general, and there are no limits to finding out more details. Also, it can't be "math can't become sentient," because if we can simulate someone's brain with math, and the simulation is good enough, the behavior of the simulation will match the brain's behavior to the degree needed to produce the effects the brain produces. I can't think of any other points you might mean by writing your post. Can you please share?
I'm a statistical programmer - and I agree 100% with your comment! I have only a rudimentary understanding of the AI algorithms themselves (I programmed classification trees and random forests which I imagine are baby versions of them), but the obvious issues with AI content (lack of consistency, inability to 'remember' >100 word prompts, over sensitivity to training data etc) are hard baked into the black box statistical methods used and can never be improved upon. There's no model parameterisation that can control 'consistency' or 'memory', the ceiling has already been reached on how good this stuff can get. While I imagine the AI interfaces and training data sets might improve or become more efficient, these models are as dumb as a spade and always will be. I'm so sick of all the fear-mongering hype generated about this garbage! It's all to scam investors as the venture capital cash is drying up in Silicon Valley. It's not based on real science. It's based on science fiction fairytales and Terminator 2. The day we can bring the dead back to life is the day we can instill consciousness in a laptop 😂
Once the hype dies down and these money-gobbling tech companies sink into the abyss where they came from, people are going to realise that everything AI generates is rubbish. It will never be capable of replacing humans in the vast majority of jobs - that is a lure designed to scam investors for cash - and people need to stop getting their science understanding from the agenda-ridden nonsense in the mainstream media.
Also, to any artists out there worried about AI stealing your work, make sure to use Glaze and Nightshade and fight back against this disgusting theft of your IP!
@@silverbirch-youtubeI have some undertanding of how AI works, and it does seem like its the algorithm we use that is limiting its intelligence. it still doesnt map to a brain well enough. However, the brain is proof that some algorithm exists that can reproduce intelligence.
We know our current algorithm has plenty of similarities with how humans think, like being really good at pattern recognition, and being capable of extrapolation. However, we also know there are differences. to train AI to sort apples by color and size, you show it samples of both mixed in, while humans perform better if you first teach differentiating by color and then, separately, by size. From this, we know we are missing something from AI, that the human brain does. Dismissing the concept of AI solely because of the limitations of the current version is probably wrong.
Don’t think i’ve ever commented on a video before but i just feel so strongly about this one, i genuinely think it’s the most important video I’ve seen on one of the most important topics of our future. In a time when it feels like most vocal opinions on this topic are either ignorant or completely insane, thanks for this extremely enlightening video.
We making it outta the president AI voice TH-cam videos with this one
The only good thing to come out of AI is that clip of Joe Biden discussing potent weed
Much agreed
And doing my research for me
Prompt: Write a comment on the topic of "desire to leave behind the current era of formulaic and stale AI generated videos", in a manner that ensures that the comment will get broad recognition and user appeal, by choosing a recently popular snowclone format textual meme and inserting the topic as its variable term.
@@AexisRai sorry m8 gonna need my ai to read that for me
"In a free market, when something bad happens to you, you have to be punished". Spot on.
where in the video does it say that
01:06:49@@snail64
@@snail64 1:06:50
In a non-free market, when everything is normal you have to be punished.
@@hollyjaw3303 in a society driven by money you have to be punished
What's funny is that since the release of this video, people generated so much AI images that the AI itself started training on its own data, making results worse again - amazing.
AIncest
@@RemMcKoffl3r Not much of a Wincest, this time...
ai wincest
Glorious.
And that's what we call spreading misinformation
If I have to hear that fake cheerful ai girl voice or that serious male voice ONE MORE TIME I will SNAP
How many times have you snapped so far?
@@MalicousOne I am beyond insane. DID YOU KNOW...
@@MrMarttivainaathere's some numbers you can call which may help you...
Or that goblin voice.
I hate that I know exactly which ones you're talking about.
Somebody pointed out that AI is perfectly tuned to appeal to the "idea guy", you know, the obnoxious person who says hey, I have an idea for a book, how about you write it, and we split the money? He values the writer's work so little that he thinks you're going to give him half the money after you do absolutely all the labor to make his pedestrian idea real. AI finally gives him the slave he's always wanted, or promises to, and it's a big reason why the uptake has been so swift, despite the AI being pretty bleh at most things.
you just managed to formulate my thoughts for me (which considering the topic, is ironic)
this has always been bothering me, as there's plenty of people with ideas, but very few who actually take the time to implement them, and even less people
who learn to implement them better with each iteration.
that same thought process also explains why its so popular.
Pretty much every AI bro I've met on the internet so far was an idea guy. And it's funny how they flip flop between "this will make artists' skills irrelevant" and "this proompt took skil, praise me"
@@vaiyt Exactly 😂 I've prompted for hours, im a true artist
But that's exactly how eNtRePrNeUrShIp works. U have an "idea" and a bunch of money to make ppl make sth actually useful out of it, then pocket most of the profit. The ideas bro has only one thing that separates him from the Elon Musks of the world: he's not already a millionaire.
If George Lucas had AI, would the prequel trilogy have turned out better? 🤔
This AI stuff just makes me sad. Apart from the stuff about killing jobs and creativity, what saddens me most is that it's killing critical thinking. I see more and more people using AI to do stuff like make drafts of their work, plan their day for them, summarize research, etc. And I take their point, it can definitely help automate some of those trivial tasks. However, it's taking away all forms of thinking from the equation. Drafts are good because it gets you to THINK about what you're going to write/create; planning your day makes you THINK about what you need to do and what to prioritize; doing research and summarizing the points of various authors makes you THINK of their views, what the key takeaways are, and allows you to assess the quality of the work. Using AI will make it easier, sure, but you've put yourself at a huge disadvantage in terms of having control over your own life and what your mind is capable of.
Quantity over quality = our present AI "revolution"
You profile picture is making me happy. :) thank you for the memories
People who don't want to think will find a way to be stupid with or without an AI. Fuck'em. Hopefully, they're going to get exactly what they want, and be able to exist as immortal and nearly mindless hedonists in FDVR pods for eternity, and leave the rest of us alone. As for jobs, I'm actively rooting for AI to take every job ASAP. I want to live in a world where no human is able to contribute anything of any meaningful economic value to society, because I think that's the only scenario in which we pull our collective heads out of our asses and make a strong push for a more egalitarian, human-centric system, as opposed to a hierarchical, profit-based one.
Nonsense. You could make the same argument about calculators, that they hurt mathematics because we rely on them giving us the answer instead of calculating it in our heads. AI will not "kill creativity", just like calculators did not kill our mathematical abilities. People will simply use their creativity more effectively and not waste with stuff an AI can do.
The critics of AI are the same as their prophets. One side claims it will be a revolution that frees us from all ills and the other that it is the devil that destroys humanity as we know it. Nonsense.
I’m a uni student who’s half way through writing my dissertation and although using ai to help hasn’t really been a temptation for me, it’s fairly often used by other students. Most of them reword and change what it spews out but they can still get caught by plagiarism scores and lectures who know their shit. My dissertation advisor even mentioned a time when a student submitted a completely AI written dissertation snd failed their entire degree because it created doubt about how much of their work is actually their own original work.
For me it’s not really been a much of a thought because I don’t want to risk my degree over something so stupid, plus I enjoy knowing my work is 100% a result of my own effort and time. Even if there wasn’t the risk of disqualification I find AI is only meh at answering questions that are popular and makes very generic responses and is practically useless for anything niche.
I can't recall the last time I watched a piece of media where almost every sentence carries weight and meaning. You've put a lot of effort into producing this and it shows. Excellent work.
on the "democratizing art" point, taking it to mean that ai is creating equal grounds for all to participate in the creation of art, the real phenomenon is that people's access to the *products* of art is being made more equal. it's a very cynical view- one which erases the concepts of artistic process, of honing craft, of exploring one's own tastes and philosophies and the history of their chosen medium, of journeying to one's most authentic style of expression. it views these not as rituals of growth and development and expansion, but as mere hurdles which should be dodged as efficiently as possible to reach *the thing* that can then be shown off or sold or whatever. meanwhile, the grander issue of so many people lacking access to the pursuit of art due to whatever socioeconomic constraints may be restricting them gets left by the wayside while ai parades itself around as some heroic facsimile of a solution to the dismal lives of every overworked person who dreams to just have enough time and money to meaningfully access and practice any sort of craft
I couldn't have said it better myself. The point of art is the journey and the communication of ideas, not the result and profit.
You can still have the journey with AI art, it helped me a lot personally
@@LutraLovegood No you can't.
Why not? Can't u still do it uniquely?
@@danilo071983 Because handing off the process and the journey to let AI do it for you is the definition of skipping the journey, not a way of having the journey.
I saw this dystopian AI system from India, it looks at employees who are slacking or 'not working hard' and fires them.
Good lord wtf. That is seriously evil.
I'm sure that won't backfire with 1.4 billion people what if just 2% decide to rebel? That's bigger than India's military police force combined x10
@@Spartacus547 Yeah but it takes at least 100 indians to match 1 competent Westerner.
Of all the bad surrounding AI, this example is not really bad at all.
49:22 "a good chunk of people who hire designers have no taste and think they can do the job better anyway, now they can tell the ai to make the logo pop and it won't laugh awkwardly and ignore them" this is the perfect summary of the problem. The power for the decisions is in the wrong hands and people are more and more convinced that controlling capital means been always right.
Fortunately, it means they can represent their horrific taste out and proud
TH-cam users always write about how socialism is better than capitalism. But the irony is that when it comes to AI, youtube users judge it as if capitalism is unquestionable. So AI is attacked rather than capitalism. It is so easy that users don't even have to be paid.
@@menjolno have you heard of "redundant systems"
When you place total power in the hands of childish gods, you had better also kiss your world goodbye at the same time. They'll destroy it in order to milk maximum personal gain out of it, you'll be powerless to do anything more than complain and watch it happen, and in the end you'll all die together. Except, they'll die with a warm binkie they refuse to share with you. Maybe best keep a shiv on you through all that so you can take it away from them as the sun goes out on a devastated planet.
The moment creatives stopped owning their companies is the moment art stopped being produced
As long as we keep replacing jobs with machines or programs, but leave the unemployed without a way to thrive, or even just survive, I refuse to call any of this _progress_ .
beautifully said!
Like voiceovers....for example
The vast majority of people are fine... you'll just do other jobs
@@dillonblair6491 well fuck those other people. i guess
@@dillonblair6491the "vast majority" isn't everyone
As an artist, AI generated art unironically made me apprieciate artists efforts and their thought process even more. It's just so much fascinating and interesting to me how artists manage to understand the sheer complexity that is art.
Ed: that's a ton of appreciation from internet strangers, and I appreciate all of you for this. Means a lot for me. Like really, thanks!
What I really appreciate and deeply admire is that there are still people who make and share art made by hand, who still try. I pretty much quit making art these days because it seems as futile as trying to outrun a race car on foot, and I have no idea how much mental fortitude it takes to still try in the face of imminent failure.
@@nobody-nk8pd I'm truly sorry to hear this. Like really. As you said competing with AI is impractical time-wise. You could get the same level of quality tho but that requires time and dedication. If you ever wanted to tackle art again, I'd advise you not to focus on this aspect of AI. I'd tell you to try to take advantage of it as much as you can to yield the best results for yourself.
I still draw and render by hand but I also want to explore the world of AI and see how I can make the process of crafting a piece (obviously not by literally generating it) even more efficient.
@@nobody-nk8pdPersonally, I love making art. Despite the fact that studying anatomy is incredibly soul shattering, the idea that one day I'll be able to make my own character from the ground up keeps me going. That's how I stay hopeful. It's less about the result, and more about the journey. I enjoy the process.
Being an artist for a living is scary right now because of AI, but it has always been like that. Creative industry has always been brutal to get into, so it's nothing new.
Basically, I just make art for myself, and if I get lucky to have my art be popular, I'll take it from there
As a regular art commissioner, I always appreciate the hard work put into the pieces I commission. It honestly inspires me to pickup a pencil and just draw things in my artbook, it looks crap for sure right now but I hope one day I can produce the same quality art that I regularly commission. And honestly, this has helped me understand that every stroke of a brush/pen matters in an artwork, which has made me respect hard working artists even more. Anyone who spends a long time honing their craft will always get my respect.
@@nobody-nk8pdthere's still a lot of people who appreciate art made by a human being. I don't believe it's hopeless. If people adapt to the change without giving up drawing by hand, yah know there's always a way. Blockbuster went out of business when Netflix offered something new in rental. But there's one Blockbuster left! They just found a way to adapt to the change without having to do what Netflix is doing. There's always hope.
I think one thing that will happen, is it will increase consumer fatigue. It will become harder for the consumer to be a discerning customer and they will have to 'dig' more or else more quality content will die in obscurity, or quickly be ingested and shat out by a company with more money, by just ctrl-C ctrl-V an art style or LLVMing a text dump of the original game. We are already at a point where YT creators can influence if a game/movie reaches the right audience. Otherwise there is a high chance a game will die in obscurity. A lot of this is going to hinge on having enough people 'hate' AI generic content. I don't think the majority of people will care, but hopefully there will be niche communities that emphasize 'hand-made' media as a reactionary movement.
Pinterest is already becoming unbearable because of this. I used the site to collect art, but now I have to slog through so much AI shit.
"Hand-made" is such a weird way of putting it. It wouldn't exist if you didn't put your hand on it. What people need to do is not shun the tech, but to focus on open source, and not falling into subscribing for even more services.
@@illyaeaterI'm a purist, I guess, but I balk at the term "hand-made" when companies/people try to sell their "hand-made" furniture, that was all made with machines. Some guy recently got called out because his "hand-made" guitar was made using a CNC machine. Can you imagine the audacity?
Like, if you're not gnawing at the wood like a beaver or clawing at it like a bear, is it even hand made?
@@fredmercury1314 Yeah I guess it's only really hand made if you spent 50 years distilling the sauce
We are literally just repeating what our ancestors always have done. When people were satisfied, or kind of okay with their lives, they let systems "control" them, when they had enough they changed the system (governance or nature itself, doesn't matter, just some kind of system they lived by). The only caveat I see to this is if (or when) we actually manage to "reverse engineer" our biology and mind and become able to play God with ourselves - not just our environment. We might kill our humanity in the process and a new age would begin when we no longer live by million years old rules.
the "everypony" at the end of the intro triggered my fight or flight response.
"AI art is like photography, because the photographer didn't make the things he took a picture of."
Okay, my dad taught photography and let me tell you, it's not that simple. Photography is all about the ability to consider pose, lighting, framing, and composition, which you have to do in the moment and with a good instinct for timing. Anyone can just snap a photo and have it contain the stuff you want it to have in it, but it takes lots of training and experience to be a photographer that can make art.
AI art is like snapping a blurry photo on your phone and calling yourself a photographer.
@@ceddavis7441 Even that takes more effort on part of the individual.
@@ceddavis7441 AI is like telling someone else to snap a blurry photo on THEIR phone, not paying them and calling yourself a photographer.
@@PapasGatito yaknow I wasn't sure if it could be put it better, and here you are. 100% down to a tee.
The way AI is being used is emblematic of how the expansion of human knowledge is ceaselessly commodified and weaponized against ordinary people. Makes me feel like the guy in Mad Men who lost his mind after coming into close proximity with a computer.
Absolutely
You should read Yanis Varoufakis' most recent book.
Ironically enough, youtube wouldn't stop recommending this video to me.
24:44 As someone who has used AI Dungeon before, let it be known that unless you wanna type out long walls of text directing the AI towards your desired story choice, it will just do whatever it wants.
It used to be better
It used to be so much better...
When I first came to it, it was like that, my head exploded
Now it just responds to anything with "so you do it"
I don't know why, but it feels like we've gone backwards since AI dungeon. AI generated stories are just not as fun as they were when that came out.
@@Jerryfan271I think it's because it got less inconsistent. Earlier AI dungeon was far less intelligent and more creative in how it did things and I miss that
@@martianscienceenthusiast A.I Dungeon both changed models and censored itself. Majority left due to that and other previous stuff.
It took me a minute to realize what was going on with that Minecraft introduction but once I did it completely took me out. Great job.
I dunno if someone mentioned this already but 12:16 reminded me of how there was a time on Google where they posed that type of image-training as a "game". Google would present random images and you (and another player) would tag them. I think the idea was that you didn't see the other player's tags until the end and you got more points if you both had similar tags (with the same words). I remember playing that a lot when I was really young. So, basically, they got 11 year olds to do their work for them, too.
Also, I am hoping that AI's algorithms become very transparent to the user, where they tell you and show you exactly what they are doing, what actual images in their database they are shifting through/melding together etc.
Its very important to see the dark sides of the AI's algorithms, it will make sure that younger people make the right choices.
that's not really practically possible; one of the most fundamental limitations of any kind of computer learning is that you don't learn what the computer does; the images don't exist anymore in the dataset, they've been distilled into rules. i'm not educated enough to completely rule out the possibility of something like that existing, but it would have to be a fundamentally different technology than is currently in use@@DonnyKirkMusic
well, was it not a game?
THe problem with AI is that it's neither Artificial nor Intelligent... a human somewhere has to add input into the system from then on there are a bunch of conditions (if else). Sure maybe they are using all our input all the time to train the model, but if you try to give AI problems that require some real thinking instead of repeating some existing model it fails miserably.
@@BillClinton228 thats not how AI works at all... atleast, thats not how neural network AI works, you should learn before you speak so confidently on a topic you don't actually understand
"Minecraft is a game that needs no introduction"
Proceeds to give minecraft an introduction.
the most hilarious thing ever
Just to make the video stupidily long
@@madcatmk2schADHD moment
"Cut the deadwood." -My graduate advisor
@@madcatmk2sch thats arguable, i think it was worth it to make it that long, i needed this in an internet feed full of frustrating 1minute videos, this forced me to pay attention for much longer after that intro by hooking me at the start
I'm a motion designer/2d animator, at work, I have at least 3 meeting a week discussing how I can 'integrate AI into my workflow'. I don't know how much more of this topic or those people I can take.
I'm very curious, what do they ask you and how do you respond? I have a friend who is a programmer, he was questioned about similar things by employers, but luckily for him, his workplace eventually dropped the idea of forcing programmers to use AI. It's a plague.
it is a bubble.
This was possibly one of the best video essays I have seen on youtube after a very long time! I can only imagine the amount research that went into this.
This genuinely one of the best, most well articulated video essays I’ve ever seen. This is genuinely approaching the realm of Line Goes Up by Dan Olson (although quite different).
The way in which you make and word your (excellent) points in this video is masterful, and every step along the way is just as good. The conclusion was so well worded that it made me write this comment.
The 2nd-to-last chapter also really stood out to me. The overstimulation section (you know the one) managed to create a truly dire and *deeply* felt portrayal of this internet nightmare that really resonated, and drove your point home in the best way it could.
It might not be the most structured video, but it’s certainly incredibly compelling.
glazing
@@fattshea8085are you stupid
interesting line go up cideo
@@chlorine--17 excellent comment, my friend
You had me at leaded gasoline. I spent the better half of my adult life trying to understand market failures and examples like DuPont's made me realized how adversarial market actors can be, hiding the externalities of what they're doing to increase profits, then using that money (which economic theory says wouldn't occur in a perfect market with perfect information) to rewrite the rules. Great job connecting the through line at the end to pricing discrimination, nudging, and commodification of time, all things that result from an information asymmetry and that stand to grant specific companies a lot of money, market power, and even political power depending on how that money is invested.
The real runaway superintelligence is already here, already misaligned to human values, and will bring about a world that I think most people wouldn't want if it were presented upfront to them.
This! It's not that ai will become sentient and rebel against the humans, it's that it will supercharge the casual inhumanity of commerce and industry.
This is fascinating
@@stevecarter8810 IE rampant capitalism, the death of humanity and its soul
.
As someone who's bit of a capitalist I agree on this,
This is use too much.
"will bring"
Already has been. (It can get worse tho, don't worry about that :P)
I actually like this video because you point out AI is itself not the problem, but its use is a symptom of the problem.
I think the current implementation of AI is the only natural conclusion of this technology under capitalism. AI is just accelerating many of the existing problems of capitalism and depersonalizing them
When an environment produces something that destabilizes its own ecosystem, is the problem the thing doing the destabilizing or the environment itself?
@@aarvlo It's the natural conclusion of technologization. I forget who said it, but "The Soviet Union and us [the US] have the same goal: technical efficiency!"
@@echoecho3155 Well, good thing that you forgot who that goofy was, because "technical efficency" was never a goal to begin with. Historically and nowadays, effeciency serve a purpose, then, an intention. Efficency for it's own it's not an idea but a closed loop that looks like what an AI would make as a gibberish to fill its emptiness.
@@Factuel-ry7zi I think you may have misunderstood my meaning.
Jacques Ellul posited the idea of "technique," a mindset by which all human activities are subjugated to rational, technological methods of increasing a system's order and efficiency - reducing cost, removing roadblocks, and maximizing output.
Whether it's utilizing AI to mass produce art-as-content or forcing peasant farmers off their land and into factories to maximize production, the mindset is the same. Humanity is subordinated to maximize the system's efficiency.
As all economic systems seek to more efficiently utilize resources, whether socialist, capitalist, or communist, they all submit to technique.
So yes, the USSR and US pursued the same goal: maximize technical efficiency at human and environmental cost. Both were at the mercy of technique as a guiding ethos.
John Michael Greer's writing on "Progress" as an ersatz civil religion are also relevant, as that ensures technique proliferates.
I'm beginning to think Big Tech is actively trying to devalue art because art's what's making us think critically, too critically for their liking...
A proper conspiracy theory but one impossible to dismiss, I fear.
Holy shit man. I was listening to this video absolutely enthralled with what you were saying- you have a really wonderful way of essaywriting- and I had no idea how I’d never found your channel before. But then, you brought up Rain World as “high art”, and I checked- and you’re the one who made the world’s best Rainworld deepdive a couple years ago. I’m so glad you’re doing more, super insightful stuff.
Updated as of-12/27/2023
----------------------------------------------------
I've been fascinated by what AI dose to what we think about authorship & originality. I think that AI will make use confront the possibility that ancient hominids producing "cultural artifacts" simply may not have anything to do with being sentient life at all. So when I watched this video, and saw some of the common misunderstandings about the technology of AI in relation to American copyright law. I thought hay I'll just slam out a quick explanation. lol
1). AI training sets are full of stolen art?
--In America the training sets that are essential to the functioning of art AI were originally compiled by universities and research institutions under the classroom use exception to copyright I.E. fair use. so the artistic & literary works in these training datasets arguably compiled for the purpose of education didn't need to be licenced at the time thy were made.
The why these training sets were made dose not imply that companies can also use them directly to make products, this present a problem for them. That will likely need to be hashed out in court. Open Source software on the other hand. Because they are not explicitly making a for profit product. Will likely be far more successful claiming “fair use" so it is IMO vary unlikely that training sets made this way in the past. can be stopped from being shard for the foreseeable future.
2). AI can't produce original ideas or art.
--The American copyright office has touched on the problem of a "non-human creator" before.
----
"The 2014 Monkey selfie copyright dispute"
--Resulted in a copyright ruling that works created by a non-human "Animal" creator immediately enter the public domain upon there creation. No matter how much knowledge and labor a human contribute to making things ready for that creation. I E "sweat of the brow"
----
"The 2023 Théâtre D’opéra Spatial dispute"
--Resulted in a copyright ruling stating that Art AI are authors of the original works thay produce, “when an AI technology receives solely a prompt from a human and produces complex written, visual, or musical works in response, the ‘traditional elements of authorship’ are determined and executed by the technology-not the human user.” AI Registration Guidance, 88 Fed. Reg. at 16,192. And "Because Mr. Allen has refused to limit his claim to exclude its "non-human authorship elements," the Office cannot register the Work as submitted. "
----
"The 2022 Zarya of the Dawn Dispute"
--Resulted in a copyright ruling that says in-part. ""Rather than a tool that Ms. Kashtanova controlled and guided to reach her desired image, Midjourney generates images in an unpredictable way. Accordingly, Midjourney users are not the “authors” for copyright purposes of the images the technology generates. As SCOTUS has explained, the “author” of a copyrighted work is the one “who has actually formed the picture,” the one who acts as “the inventive or master mind.” Burrow-Giles, 111 U.S. at 61. A person who provides text prompts to Midjourney does not “actually form” the generated images and is not the “master mind” behind them.""
----
The processes that humans naturally follow to create original art has been enshrined in copyright law, this is why we can separate. In a court of law:
A copy of a painting. From a derivative of that painting. From a painting inspired by the painting,From a painting thats in the style of that painting. Being inspired by a pice of art you’ve seen, and then painting something original in that artists style. May be in bad taste but un-controversial if the AI were a human artists and not software.
In short. If AI did not produce original art based on the "inspiration" it gets from mesmerizing the abstractions of billions images. The art produced would not fall into the public domain like it dose now because of non-human authorship. Instead it would be copyright-able tho an un licensed derivatives. End.
--------
Update 10/09/23 After being challenged. And re familiarizing myself w/my sources. I got a date wrong, was un-clear in places and mixed a couple arguments together. I'm making edits to correct my errors.
---------
Revision 10/10/23
Revision 10/11/23
-----------
previous comment-
-------------------------------------------
I thought this video was insightful aswell. but little inconsistencys with how Jimmy McGee miss describes the tech, and how it works in relation to american copyright is what changed my mind.
For example: In the US the training sets were originally compiled under the fare use exception in copyright law for research and education. and therefore the works in them don't need to be licenced, And this is one of the main stumbling blocks to wide use outside of academia. Thankfully Now an entire industry has sprung up to supply different legal sets to developers.
The copyright office in the states has rulings on "Art AI's" from almost twenty years ago. That apply to the art of AI's that are made to follow "the legal process" that allows humans to make new art from All the art and things thay have seen over their life or are referencing at that time. For inspiration. without violating any of those copyright holders rights.
the case im referencing was a programmer who invent an AI that replicated a dying & hallucinating human mind that autonomously outputted pices of new original religious art. He sued the copyright office when thy ruled against him. They ruled that the AI was being original, and a copyright was created for each pice of art it created. The AI isn't a human being tho and so couldn't defend its own copyrights in a court of law. Nor could the owner of the AI gain use of the copyrights ether. just because of the artist that created the art at question was his legal property doesnt mein he could clame its property.
Rainworld is shit
@@KzaksI don’t give a fuck about copyright law man I pirate half the games I play and run linux
And while I do have criticisms of this video and don’t fully agree with its conclusions (as with all of his stuff) it got me thinking about the issue from a new perspective I really did find insightful- I was actually excited to read your criticism at first to maybe help me understand some of my own disagreements with the video, until I realized it was some weird legal argument written in broken English.
also I wrote the original comment while high as balls
I’ll have to give the video a rewatch sometime and try and formulate my thoughts and criticisms on it, I don’t entirely remember
A barely 30 year old friend of mine, who doesn't understand shorts or reels and doesn't want to engage himself described it as watching yourself grow old in real time. "Growing old" in the sense of becoming incompatible with the "new world" of the next generations.
Dang bro. You hit hard
Maybe it's not him growing old, but him experiencing the older form of video "content" and now being able to compare it to the newer form of shorts forced on everybody. I don't think it's entirely an issue with age and more-so an issue with companies trying to extract the most amount of money out of people, if that makes sense.
@dormitthings4773 Obviously, there's some nuance with this, but one could few us criticising enshittification as old people being sentimental about how everything was better just a few years ago.
reason shorts dont work for him its not becouse he is old , but becouse he dont wants to consume mindless content , its not that he grew old but grew up on stuff like movies , books , and proper videos , not attention cash grab scam , i see this in food... , and just like fastfood companies are conditioning "new generation" , youtube and social media are conditioning new consumers that "this is content
ofc child would want to eat just icecream , instant ramen and other garbage like that , similar is with internet content , problem is in today world you can cook for your child and feed him good stuff , but you cant really do that with online stuff , today even finding a real job requires you to engadge with a lot of that garbage compared to just 5 years ago.
its not that they grew incompatible , but today world deliberately becomes more incompatible with us in order to pray on us or condition us to be exploited.
ask any 15 year old how many books he read , does he read newspaper or articles? watch documentaries? good movies? , most will talk to you about roblox , minecraft , or garbage content they consume.
and just like food , if you will be conditioned long enough...that this cheap low quality zero nutrition burger and fries is your tasty meal , you start to belive it , after a while you will carve it , despite it slowely kills you , similar is you youtube , social media , and attention platforms , their AI trained on model is smarter then you and me , they can indentefy weak , or individuals they can sway to their side , they got so good its not just some...but most.
Fu***g hell, I felt exactly the same. I'm 25 and I *felt* the curtain open on the "I'm never gonna connect to the poor younger people that are growing up currently" feeling.
Which is a super sad thought that I will fight with all my strength, if only for the sake of my 10 years old nephews
Mandalore, redlettermedia and now Jimmy. Today, my cup runneth over.
Word
I know what I'm watching tonight
rlm sux
The "progress" that has been made since this video being released 5 months ago is astonishing.
exactly
The trajectory the internet and society as a whole is going towards really was nailed by MGS2 (faceless know it alls deciding what everyone should consume under the guise of providing context to content in a sea of worthless data) and MGS4 (AI and algorithms being so intertwined with reality that no one on the planet realizes that they're taking orders from machines anymore).
Sad that the people who would have the exact opposite position to yours also believe that MGS2 and MGS4 predicted what _they_ believe to be the story of today.
(I.e., people who believe that people like McGee and you are economically illiterate SJWs or socialists who are either ham-fistedly trying to cage AI that could greatly improve the world, ignoring the "science fictional" type of existential threat that AI can pose to humanity, or somehow both at the same time; they also will point to MGS2 and be like "so visionary, Kojima would agree that I'm like Philanthropy and @cromtuiseagain is like the Patriot AIs".)
@@coreyander286 honestly saddened by the idea they could take in all that stuff about echo chambers, fake news and discourse in general becoming a shit show, and decide "kojima got it right, my worldview is exactly what he meant" and go back to shitting on "the SJW's/political group i don't like" in an online group designed to discuss that one and only thing without a second thought
i'm lucky i'm irritated by loud noises and people, those echo chambers must be hell on the brain
MGS 2 honestly was optimistic in making its big bads the metaphysical representation of the US government, when what we got in reality is the ideological amalgamation of the most insufferable nerds on the planet
@@coreyander286based take
@@Graknorkecan't be any truer than that. The entire world of tech is just governed by the worst type of nerds that got soo successful that they basically became the jocks that would bully them at school instead of growing out of it.
You have outdone yourself! This is up there with Folding Ideas' videos about tech grifts in terms of research and quality! I'm glad I've been subscribed to you and to be able to see your eventual growth as a great essayist!
Its really funny to me how in about 8 months, after the hype bubble began deflating, saying something "looks like it was made by an AI" went from a celebration for technology to a straight up insult, as it should be. Another banger video.
Ai is Th
[thing] is bad now and will never improve okay thats a real argument and definetly not cope or a placeholder for my unwarranted anger i got from reading a twitter thread
@@hydroxide5507 You're putting words in their mouth.
P.S. Your unwarranted anger is showing. :)
@@hydroxide5507 "unwarranted anger" lol, art theft doesn't warrant anger anymore huh? educate yourself kthx :)
When ever was that a 'celebration,' are you just making stuff up?, from the moment that saying became, all I saw was people trying to insult the technology, especially the artist community.
Just a minor point: By "democratization of art" I don't think they mean "art will operate like a democratic government," but rather, "the ability to make and market art will be accessible to everyone." Which is a much more seductive idea, but also complete bullshit. "Hey look, anyone can make art! F the establishment!" While stealing work from those who actually put in the time and effort to create something original.
I remember when they said that the Internet would lead to the "democratization" of commerce and retail, and instead we got Amazon. AI won't lead to the little guy making it big in the art world. It will just be used as a tool by giant corporations to accumulate more wealth.
Drawing is already democratized, it doesn't matter if you made a sketch of a stickman, it's still, art if you made it and has soul poured into it
If you use AI, sure, it looks prettier, but there is no soul poured into it, just, Ones and Zeros,
also typing a phrase into an image generator isn't you making art. it really bothers me that people use that "anyone can make art because of ai uwu" because you're not making or doing anything except providing a prompt. like having an idea about a drawing isn't the same as trying to draw it. one is making art.
@@themischief420fr it's crazy that people genuinely compare those two things. I saw someone try to defend AI 'art' by saying that coming up with good prompts requires skill too, like, huh??? Typing some words is the same as knowing composition and anatomy? Sketching? Painting? Rendering? Color theory? Lighting and shading? Years of effort and passion? You really gotta be sniffing glue to think that they're remotely equal. Behind every artwork is a dedicated and talented or aspiring artist, and behind every AI prompt is some asshole who can type on a computer.
I am a uni student studying Data Science and AI. This video is spot on. It’s like my prof made his lectures again but 10x more enjoyable. If somebody wants to learn about the current state of “AI”. Send this.
Or missunderstand it
@@bluegamer4210 For real. If this is better than what he's learning at uni, he needs his money back
@@InnsmouthAdmiral They didn't say it was better, just more enjoyable. Many profs' lectures don't include a lot of humor or multimedia production.
It’s very heartening to hear, as an established and mostly isolated working adult who feels like corporations are just tearing everything to pieces, that educational spaces are bare minimum at least talking about this stuff seriously
just out of curiosity: Does thinking about this put you off the data science and AI world?
Oof, this hit me: "We're at the point where we need to choose whether to build a world for money to live in, or for people to live in."
Errrm yikes moment
yawn, just commie bs
In other words: "we're at the point where we need to choose if we want to keep serving or abolish capitalism"
For people with money to live in.
@@alexxx4434 no, for money to live in. There is a distinction, but it's important. Money supercedes people in a post-capitalist world. Even people with money serve money
28:25 fun fact, Vizcom was used often to cheat legitimate artists out of money, as clients opted to pay for a sketch then throw it into the AI rather than PAY THE HUMAN FOR THE COMPLETE WORK
I wouldn't really call it scamming. They requested a sketch the artist delivered and was payed. No scam involved
@@rene3924That’s a terrible take. That’s like asking someone to fix your plumbing, before they come in and take your money and “fix” it, and after the fact realizing that they did nothing. You lose out on time and money and they gain your money in return. The artist loses out on time while the customer gets the thing they want for free. It’s parasitism. Only one side gains while the other loses.
When the customer requests a sketch they do that so they can confirm that this is the product they want. Usually it’s also a deterrent for customers to make any demands changing the artwork before it reaches its final stages, because they’ll already have the artwork they’re satisfied with.
@@rene3924 I don't think you understand. Artists are paid based on their time and effort. Getting a completed work from their sketch is cheating them out of the money they should have been paid.
@peaberry9413 A completed work, not her completed work. Any artist can take a sketch and "finish it" in many different ways.
I don't see an issue with running an algorithm on someone's sketch that you purchased (so really, your sketch). I do see a moral issue with selling the new artwork, but I do not know who would truly own the piece. It is transformed, so technically new. But it was done by an algorithm which legally can't copyright artwork. 🤔 Hmm.. I think ultimately it's legal to sell, but illegal to copyright.
I still stand on the grounds that it's a shitty thing to do to an artist though.
@@acebinko1 I guess I should add on that maybe some artists would also do the sketch for free BEFORE asking for payment, in which case if the client declines to go further - and thus the artist isn't paid - they could still run the sketch through the AI anyway.
I can't recall the last time an artist WASN'T asking for some form of payment up-front, though, and it's probably to combat stuff like this.
It's not irony that AI is taking over leisure activities and enjoyable work like art, it's by design. It's a deliberate attack on any behaviour these ghouls cannot easily monetise for themselves.
they can though, spotify dropped the perceived value of music through the floor, and the cottage industry of playlist curation has musicians acting the part of "artificial artificial intelligence" trying to match the hyperspecific requirements of playlists, no generative ai necessary
the digital painting people have it quite cushy right now even without the entire internet suddenly giving a shit about art
I can't express in words how valuable this essay is
But you left a comment
@flameguy3416 I understand it's a little quip but since it's a pedantic quip I feel okay with saying "well actually" because though they used words the implication is that they have not expressed how valuable this essay is. From their perspective the comment only served to express a miniscule fraction of the value and to express the true valuable would be impossible. Interestingly, if they said "I can't express in words that this essay is so valuable", then it would be ironic since then "is" serves to make the statement into a binary one of "is so valuable" or "is not so valuable" and not a matter of degree like the first case.
@@flameguy3416 you're really confused
As someone who dabbles in several amateur but very personally cathartic creative hobbies, programming, modeling, writing, art, etc., the reduction of human choice and expression for the sake of constant, unabated monetization by people whose only goal in life is to generate profits for their shareholders hasn't been lost to me. I believe its one of the reason my concerns about AI have had a disconnect from my friends who parade and celebrate their accessibility. You can't tunnel vision on the virtuous aspects of a technology, its creation opens the door for virtuous and exploitative or malicious intentions. Throughout the 2010's we were able to ignore the growing problems because it hadn't yet disrupted our daily life, but silicon valley is unsatisfied unless they have your complete and total attention and control; that's what a vapid cult based on the principles of every piece of technology having to 'revolutionize the world' can only aspire to, ego and transforming the livelihoods of others. I always had felt alone in this concern, because most topics about AI are slap fights about one specific use case or problem, while the corporate PR jargon keeps chugging on relentlessly to capitalize on the market. It was nice at the end to realize that no matter how much it's exploited, the human condition will never change, and we will always peruse unity and contented mindfulness, and it's our responsibility to fight for the lives we want for ourselves. "We're at the point where we need to choose whether to build a world for money to live in, or a world for people to live in." is a quote that will stick with me.
Man I want to fight so bad but it's tough to find allies sometimes. I only run into anxious, frightened people who worry about skynet or people who have no spine and just surrender to what they believe to be the inevitability of it all. I've returned to university to get my masters degree in longform writing and it's fucking depressing how the professors are both uninformed and frightened too. I recommend this video to so many people - people really just need to fucking understand what's actually happening and stop worrying about fucking skynet.
Damn straight. Take a moment to realize that the most famous person that has ever existed to date spoke in parables, always speaking with metaphors and stories. One thing he said was so upfront and in your face , it was “No one can serve two masters, for either he will hate the one and love the other, or he will be devoted to the one and despise the other. You cannot serve Love and money." No metaphors, no stories to cover the true meaning. He was going to make sure there was no mistaking this.
It was good to read this from you! One of the downstream effects of the Great Tech Ubiquity--and I feel public facing AI just further exploits this--is that it has altered our ability to pay attention for decent stretches of time. Everything on demand immediately is not great, it turns out. I've been an early tester for both ChatGPT and Bard, and the degree to which they both just made shit up whole cloth was surprising and worrying. Names, dates, and narratives of whole situations that never happened. This does not make research easier.
@@JustinPogue There are 2 purposes to ChatGPT, while other companies only have one purpose. Their is no benevolent AI development that is going to take the busy work out of people's hands and give them back something better. Companies are already using this to replace people, and some industries wholesale. Anyone who believes they will use it for the greater benefit of society is just a useful idiot and should be ignored.
The 2nd purpose and this is more exclusive to OpenAI and a couple of other tangential companies is the pursuit of AGI or ASI. The problem is that this pursuit again isn't for the benefit of humanity, they are looking to build a machine god, something that they believe will give them objective truths. That is worrying in itself, but even moreso when you investigate the beliefs of people like Sam Altman, Peter Thiel, Sam Harris, Elon Musk, etc.
The origination for the idea of a company like OpenAI happened years ago, and the people behind the technology have displayed openly white supremacist and eugenicist ideologies. Microsoft even used a very famously racist psychological study from the 1990's as a bench marker for the "intelligence" of their machine learning program.
Basically they believe that an objective machine god will reinforce the idea that white men are smarter, more capable and more deserving than anyone else. A self fulfilling prophecy given the kind of data they train on.
I'm a graduate student currently working on a project about racial bias in these systems and it is beyond alarming that some of them are already getting deployed into the wild. My fellow journalists have failed massively in not exposing more of what's going on behind the scene. The panic and existential angst seem to be blinding everyone from contributing in a meaningful fashion.
Nevermind that this is blatantly the biggest heist of intellectual property in history. It's astonishing that these companies have gotten away with what they have. We'll see how the lawsuits work out but permissionless innovation dictates that you simply keep breaking the rules and breaking the system until you've created a new normal, because no one is going to stop you.
Educating people about chatbots and automation is important as well. The first experiments with a ChatGPT like program happened in 1967 with the ELIZA program. The computer scientist Joseph Weizenbaum developed it and was so alarmed by the results he spent the rest of his career warning against something exactly like ChatGPT. The problem is if a computer program spits out copy that seems humanlike and responsive to what you've given it, you begin to imagine a consciousness behind the text. Even in 1967 with a computer program so basic and so simple participants in the study began to believe the computer was actually sentient, even though the programmer knew it was simply spitting out text based on keywords in the entry.
This opens the door for predatory behaviors we see today such as virtual girlfriends that prey on male loneliness or other friend chatbot programs. People can imagine a consciousness right up until it spits out a response that is clearly not human. Then it is revealed to be simply a computer program wearing a poorly fitted human skin suit.
If you look up Emily Bender's work, she's a linguist who worked closely with google on the development of Bard and she got fired along with a number of other people on their ethics team because she demonstrated that these AI machine learning systems do not work like humans do. But people are easily conned.
These companies are not your friends. They're not here to augment your life, to improve it in any way, if anything they're here to crush you into an even more authoritarian state. An example would be this, straight out of the movie minority report;
www.bloomberg.com/news/articles/2022-06-30/new-algorithm-can-predict-crime-in-us-cities-a-week-before-it-happens?
It should not be overstated how overhyped the claims of AI companies are as well. As you said, it's startling how often they spit out results that are simply false or nonsensical, and Bender spends a lot of time on her podcast Mystery AI Hype Theatre 3000 debunking the claims companies are making.
The real threat isn't some Skynet scenario like CEO's like to keep pushing to make people frightened, it's that they slowly and systematically implement these systems into our society while no one is paying attention or doing anything about it.
As Frank Herbert wrote in Dune: "Once men turned their thinking over to machines in the hope that this would set them free. But that only permitted other men with machines to enslave them."
These are very dangerous times and fear and feckless politicians are what keeps us paralyzed and incapable of fighting back.
you are not alone, I want to bring value back to humanity in the eyes of the world. I want people to value each other over money or their own selfish satisfaction.
As dystopian as this whole topic is, your video gave me something important... a reinvigorated sense of meaning, and for some reason, hope. I'm working on a project, I would dare to call it a work of 'art', whatever that is... and you've added an intense fuel to my fire. 'Art' really is important, especially in these complex times.
Thank you for giving me a renewed sense of worth in my own 'creativity'...
Subscribed.
Me too. About halfway through this video I picked up my pencil and just started drawing while listening. Capital has taken so much from us but they'll have to pry creativity away from my cold dead hands.
@@KirstyBaba 🙌
Never say die. I want art because its made by another person with something to say, not because im a machine that reacts to stimulus.
@lumariousthelion I am gonna subscribe in case you ever get around to uploading that art
By all definitions of Art, AI still falls under the “freedom of expression” category. Therefore, AI “art” can still be classified as such.
I am a computer engineering student and an aspiring game developer, and I feel like I made the jack of all trades of getting fucked in the ass.
I genuinely fear, not only for me but for humanity in general, that AI will take away most of what makes us human, such as the ability to make science, use our logic and brain to solve problems, and express ourselves through art, that an aspiring artist won't ever get to make a drawing or painting or whatever because "LMAO AI does it better and faster", or that there will be no one doing science because "LOL AI does that already so why bother?", and that life will be only relegated to consuming AI generated content and rotting away, that a writer will never get to voice their ideas the way they want because "LMFAO AI is more efficient".
I view this as a true crime against humanity, I feel like this thing is stripping humanity away of its best traits of expression, of its freedom, and I feel rather powerless and hopeless
I talked about this to one of my professors, a brilliant mind if you ask me, and they say that, no matter what, human genius and creativity will not only always have a place but it will also be needed, and that humans in all fields will have to know how to use AI as a tool. I see this prospect as rather reassuring, and even preferrable, because if we get a tool that allows us to do our work better whilst retaining the need for creativity and skills, that's honestly better, but seeing this video I can't help but ask, what if he's wrong?
I cannot find the proper use case for AI. It's obviously is "so important" that people pay for subscriptions to it. But what for? Half the stuff it says, it makes up. So the only real use case is to organize existing data, altering existing data, and generating stories. So... what am I supposed to do? The only ethical use case I can possibly think of is for generating the front end template of a story, like a mystery, and then solving the mystery yourself.
How are we supposed to use it?
I like to give the benefit of the doubt to tools, but I just can’t hop on because as a hobbyist artist with 17 years worth of experience, no one uses it right and generated images aren’t creative enough to give me inspiration. Not like nature, not like other poeple’s artwork.
@@E3AloeLi Honestly the only generative image AI stuff I've seen that tickled my ADHD brain into action was when it was WORSE at images and threw more vague shapes that sparked a series of random thoughts...
And I was able to get the same result by literally just shouting in the offtopic channel of any discord "GUYS QUICK FIRST WORD THAT COMES TO MIND WHEN I SAY [A]" and picking the first three non-replicated words and thinking of how to mix those together.
I've tried unsuccessfully to express many of these sentiments to friends and co-workers in the recent past. What you have created here is the definitive video to express what I could not quite articulate. Thank you.
Could've just asked an AI. I love the irony that all these dick head youtubers who want to be "hated" can't even write a more compelling summary of the threats posed by AI than an AI. And then they have the balls to blame other people for bloated nonsense in a video that is 1.5 hours. Several good points that have been touched on in many other videos but fucking hell man, what an overall waste of time.
Good luck getting your friends to watch a video that's longer than 5 minutes, never mind an hour.
A lot of the people are already ready to give up their own mind to become a mindless drone. They think it would take away only the bad and undesirable stuff from them, not seeing that it would take everything and leave them an empty shell, uncapable of thought, emotion, action.... life.
@@christopherkrause337 you simply do not care about slave labor and the destruction of jobs for millions, got it
@@theghostcreator776 Artists and writers are really scared of competition.
"Of course, we creators will not lose to AI. Because AI cannot imagine." - Naoto Ohshima
Every day I hope he's right.
Fun fact: AI books exist. Now, the reason that's fun and not depressing is that they were so unsuccessful they didn't even make it into the news cycle.
@@EnigmaticGentleman His name is Onision.
There is no such thing as imagination. You can remember, and you can merge those memories together, but you can't imagine something you haven't experienced or seen before. Sorry, but there is no security in that thought, I'm afraid.
Can you?
@@PCMcGee1 Your claim is objectively false. THe fact that HP Lovecraft imagined the Elder things debunks your point. HR Giger and all of his works debunks your point. The fact that our mind created the color magenda debunks your point.
I really don't like the society we've built
Who does?
@@theyellowarchitect4504Microsoft
@@dtech1218 They don't like it either, that's why they are trying to change it (to their benefit), and even now they are still not satisfied.
Creating discomfort motivates change in a directed direction
brace yourself, it's gonna get so much worse
The last place I worked at hired a guy to do coding. Every time I walked by the dude was using chat GPT to write simple stuff in Python instead of writing it...then he later complained that the AI compiled Python UI was hard to debug...it also suggested he download about 40 packages so the program was a mess and couldn't be easily exported to run on other machines without a boat load of libraries. Weird stuff AI is.
I strongly disagree. Weird programmer that was. Meaning, if you give an AI bad/lazy instructions with no thought put into design, it's nearly the same as doing so to a dev team. The mistakes that gentleman made were completely avoidable, but it sounds as though he also didn't want to *think* about the project.
@@nev6502 the type of people to resort to ai are the type of people that dont want to think about things too hard. Any enterprise application worth its salt will only use generated code for small functions. I personally don't use any cause for the time I have taken to massage and check the code for any errors, I may as well have taken the time to write it myself and understand it from the top down.
@@stang9806 I do agree with part of your comment as a software developer myself. I don't agree with the opener of "the type of people to resort to...". Leads were saying that about autocomplete / the beginnings of intellisense in the 90s and now I know very few professionals that don't use some form of it. Personally I will resort to anything that allows me to work faster. The problem / inefficiency arises when you are able to read code much faster than you can write it (even if you can write it very fast). I've been considering this more as of late and wonder about the potential of a hybrid approach (where most of the generation is done with AI, and a human audits the result). With plenty of guard rails code generation can be tailored to various styles. Food for thought; probably not quite practical just yet but if not, it's getting very close in my eyes. One thing we can agree on; it currently requires a human.
@@nev6502 And how many of those professionals can still proofread worth shit now that autocomplete (more like autocorrupt) is so entrenched?
"Never trust a WebP user."
I felt that in my *soul*
Hate that darn format.
I miss jpeg's on the web
@@GavenJr when the JXL standard is finalised webp will be obliterated. Imagine 30% better compression, thumbnails are redundant due to arbitrary progressive load through file truncation (no excuses, just cut off the data stream at data/quality limit), and perfect jpeg back compat transcoding.
I mean webp has both the qualities of jpegs and pngs with better compression than either, great for web development due to it having a low tax load on a server's bandwidth, it's just a shame that so much software has seemingly out right refused to support it for some reason.
PSA: WEBP can be changed into PNG seamlessly simply by renaming the file extension.
I'm unsure if this is a feature of Win10 doing something behind the scenes, or if the data itself can be reinterpreted so easily.
@@TeamMuggi when I've tried that it makes the image darker; loading it in GIMP and exporting it as a png or jpg feels more reliable to me
Speaking of which I just got recommended this video without even noticing its length till midway. This is an important (and extremely entertaining, funny) video as it actually broadens the AI conversion to tangible real world implications that aren't really talked about much. Genuinely speechless from the amount covered, it's a definite rewatch to process what I want to say.
Anyone else notice that the AI generated captions on any shortform video like TH-cam Shorts or Instagram Reels is almost always wrong? It's infuriating.
it's almost deliberately done to make people comment, imo... People will jump in and correct what was wrong but in turn drive up "engagement" and then the algorithm will push the video even more.
Drives me crazy too. And it's not content creators comment-baiting with intentional errors, it's the subtitle generators simply being awful and nobody caring. In a world where writing "should of" is almost becoming the norm, this stuff worries me. The internet and technology should be making us smarter but I feel it's just making the average person dumber.
@@AdaptorLive the internet has unfortunately given the illiterate enough tools to skate by without learning how to communicate properly. It's almost created a simplified/informal digital English compared to actual written English. AI will be bad at implementing audio recognition while also learning from increasingly poor training data in a race to the bottom. Yay :)
@@ProjectZ36 Nicely put and I'm afraid you're completely right.
"It's democratizing art and writing! Now everyone can do it!" Everyone already could, buddy, get a pencil and a piece of paper and start experimenting! You probably won't be good right away, but nobody is, it takes practice.
I studied machine learning in university but I've spent the last 7 or so years working as an engineer, programming business software. In the last few months I've seen a gold rush to build contrived implementations leaning into AI. When I say contrived I mean... often completely stupid. Most of the people singing the praises of AI don't know how it works and chalk it up to a black box, filled with magic. It's... It's gonna lead to a lot of lawlsuits. I've heard asks range from piping private user information into language models, to using AI to build a drop-down menu. The most common ask is to, "train" a model on an extremely small data set. It's dumb. What I see happening is big companies like Microsoft selling pick axes, and a bunch of ding-dongs who failed their way to the top of middling tech companies shouldering their way into the gold rush, begging to be the first to give MS money, if only to say, "we haz the AI."
"using AI to build a drop-down menu"
That one killed me :D
“Jimmy I need you to implement a model for server management in two weeks for our next earnings call do you think you can do that?”
Your way of writing made me read about the usage of "ask" in the English language. So you basically use the word "ask" in place of the word "question" or "request".
Ah the things you can do with LGBM or a random forest, and the people will come sprinting just to put a neural network there instead. It's all shapey statistics peeps, jeez! Get a glass of water and think about it for 2 minutes!
Yeah. Machine Learning long has had a place. Said place being mostly either for Computer Vision where a margin of failure is acceptable OR as a proof-of-concept to slowly get chipped away in regards of what is being doing by a model, till it is just a general Algoritm. Cause you rarely want something that says "Maybe it is that" to be in charge of anything critical and you often want it to be able to run on a Real-Time embedded system without affecting other tasks.
But we had a breakthrough in a single sliver of AI as a whole (LLM) and suddenly Machine-Learning had to be crammed into everything for the sole purpose of tricking dumb gullible Investors into handing over big bags of cash cause AI bubble...
“in a free market, when something bad happens to you , you must be punished” almost brought tears to my eyes man . great work as usual, you’re one of my favourite writers on this website
Top 10 myths of Free Market:
How is that the inherent result of a free market?
@@hagoryopi2101 because when two people want the same thing, the person with more power has many ways and connections to make it happen, and the free market eventually just has boards of companies who only see the stock price, so the thing that they want most is the expansion and growth, this happens at every large company. And compared to normal people who still has some options, people down on their luck usually have few alternatives, as literal life saving services are also privatized, thus when something bad happens to you, you become valuable asset to companies, and not a fellow human being deserving of compassion. Just like gravity forms planets out of cosmic dust, free market eventually becomes a system well adjusted for making money at people's expense(for example planned obsolescence)
@@hagoryopi2101 When something bad happens to you, you are in a vulnerable position. When you are in a vulerable position, then a "rational value maximizer" *has* to exploit you (if they dont then they are either not being rational or not maximizing value).
@@justincenter4061 the most profitable way to "exploit" vulnerable people is to offer them some security. Instant loyal customer, most profit in the long run. Anything less than that is irrational violent "logic," the fallacious belief that keeping others down benefits oneself more than mutual growth.
effectively what I currently see being the most likely scenario is that eventually the internet and digital media in general will become so utterly worthless as a source of information and entertainment that most people with self-respect will just drop out and move on from it
Oh god no, it will create the next gen of "Conspiracy theorists" and "anti-vaxxers". It's definitely inevitable and probably already happening, but no I don't think that is ever a good thing.
I really kind of hope this is true
people who have left fakebook years ago are the forerunners.
Ive been considering that. Internets only been around for 30 or 40 years, really still in its infancy. As incredible as it is, it could be replaced over the next century.
We might already be there... most platforms are dead most of the time now.
I am steadily being drawn to build a cabin in a remote area and salvaging used car parts.
growing my own food to live is looking real appealing now
@@peanut3438 yeah, we WISH something like that was actually possible nowadays.
But you can't even do that anymore
@@evergarden8592 Entire villages in the Spanish countryside can be bought for like €10,000 so i don't understand wher eyour point comes from
Jesus Christ dude, this an actual banger of a video essay. Incredibly well-written and edited. Usually I resign myself to semi-passively engaging with someone's multi-month long obsession created into a 5 hour video, but I was enthralled, fascinated and engaged the entire time. I'm going to be thinking about this video for a long time.
Incredible fever dream of an intro. That should have clued me in that it was GPT. I have played with it extensively, and am getting better at recognizing it in the wild I think, but I still didn't detect it in a produced video, despite the incredible dissonance between the title and the minecraft fever dream.
Incredible. Thanks for this essay. I wish I were smart enough to make use of it
It's easy to spot, especially if it is not given any half-decent tone prompting. It will sounds like an amalgamation of English language, and always be in the form of retell, pros, cons, selection, conclusion. Yes, just like an SAT essay (do they still do those?).
I didnt realise it was GPT but I did think he was intentionally making it seem as generic as possible to set a tone. Weirdly enough I feel like I easily have watched at least 3 ish 'minecraft nostalgia essay' videos, that read exactly like the intro, from 2019 til now
I had some kind of audio glitch with my headphones and I patiently watched the first 9 minutes in silence. I thought it was a really strange but oddly pleasing mystery to try and figure out what was going on.
Unfortunately this video perpetuates a couple common myths about the technology of AI in relation to american copyright law.
1). AI training sets are full of stolen art.
In America the training sets that are essential to the functioning of art AI were originally compiled by universities and reserch institutions under the classroom use exception to copyright I.E. fair use. so the artistic & literary works in these training datasets arguably compiled for the purpose of education didn't need to be licenced. This lack of licensing dose present a problem for companies that are using them to make products.
2). AI can't produce original ideas or art.
The American copyright office has touched on the problem of a "non-human creator" before now.
"The 2014 Monkey selfie copyright dispute"-Resulted in a copyright ruling that works created by a non-human "Animal" creator immediately enter the public domain upon there creation.
"The 2023 Théâtre D’opéra Spatial dispute"--Resulted in a copyright ruling stating that Art AI are authors of the original works thay produce, “when an AI technology receives solely a prompt from a human and produces complex written, visual, or musical works in response, the ‘traditional elements of authorship’ are determined and executed by the technology-not the human user.” AI Registration Guidance, 88 Fed. Reg. at 16,192. And "Because Mr. Allen has refused to limit his claim to exclude its "non-human authorship elements," the Office cannot register the Work as submitted. "
"The 2022 Zarya of the Dawn Dispute"--Resulted in a copyright ruling that says in-part. ""Rather than a tool that Ms. Kashtanova controlled and guided to reach her desired image, Midjourney generates images in an unpredictable way. Accordingly, Midjourney users are not the “authors” for copyright purposes of the images the technology generates. As SCOTUS has explained, the “author” of a copyrighted work is the one “who has actually formed the picture,” the one who acts as “the inventive or master mind.” Burrow-Giles, 111 U.S. at 61. A person who provides text prompts to Midjourney does not “actually form” the generated images and is not the “master mind” behind them.""
----
The processes that humans naturally follow to create original art has been enshrined in copyright law, this is why we can separate. In a court of law:
A copy of a painting.
From a derivative of that painting.
From a painting inspired by the painting.
From a painting thats in the style of that painting.
Being inspired by a pice of art or painting in a another artists style. Would be un-controversial if the Art AI were a human artists and not software.
In short if Art AI were not producing original art based on the "inspiration" of trained images, and not simple "Copying" then the produced art wouldnt fall into the public domain -because of non-human authorship. instead it would be copyrightable.
--------
Update 10/09/23 After being challenged. And re familiarizing myself w/my sources. I got a date wrong, was un-clear in places and mixed a couple arguments together. I'm making edits to correct my errors.
---------
Completed revisions 10/10/23
@@Kzaks This is false from your very first point, I don't even know where you made this shit up.
So far, AI has made Google searches useless, spellcheckers less efficient than proofreading yourself, and art hosting sites clogged with generic, mass-produced trash. Thanks for giving us a preview of what this nightmare future has in store for us next!
I don’t even want to imagine what cold hell we’re heading towards
ahahahah =```) so dumb
I've had to learn how to block certain generated sites to scrape the bulk of the crap out of image searches.
Hi Jarv, love your art! X95 a best!
As opposed to generic, mass produced trash art from humans? Because there's a lot of that too.
We are not automating the production of art. We are automating the theft of art.
As an artist I agree with this,
"Evil is not able to create anything new, it can only distort and destroy what has been invented or made by the forces of good.” JRR Tolkien
It's not exactly a direct quote
@@gluetubeserver I think it is but idk
There was a very similar line in a comic adaptation of Coraline I read recently. I posted a picture of it to Mastodon and THE Neil Gaiman boosted my post to his followers as well!!!
I think that people don't really understand how similar AI and Human brains are. Neural networks are literally based on working of Human brain. AI art is currently similar to visual imagery. No one conjures something utterly unique out of thin air, It's all inspired by previous works and their surroundings. It's weird to me how people call AI unoriginal when it replicates or tries to go through processes human brain undergo to develop a thought or visualization. There is no good or evil in AI. Tolkein, For all his virtue, was a rather conservative man bound by the frankly limited ideal framework of his time, unpossessed of ny real expertise or insight into our modern world, Indeed, he could not come to terms with his own contemporararies at times, as with his dislike of Dune. Not to detract from his achievements, Mind you, but one must take that into account when applying his sayings into a unrelated field.
If there is no evil in a neural network, and a neural network models a human brain, does that also imply that there is no evil in a human brain?
It seems to me that - like humans - AI is capable of evil but isn't necessarily evil by nature (AI can produce good as well as evil).
Tolkien saw a distinction between true creation and what he called "subcreation". True creation is e.g. God creating matter from the void. But anything a human does, like writing a story or making an AI which makes a story, would be "subcreation" - creating something novel out of what already exists. Tolkien saw writing stories as "revealing" them, not truly creating them (it's fascinating, read his Essay on Fairy Stories to learn more!).
I believe the quote above is speaking of true creation more than subcreation. It explains why Morgoth and Sauron could not make their own races, and instead corrupted existing races (elves, men, ents) to create "new" races. But I don't think Tolkien would apply it to our modern AI. I do think he'd detest our modern AI, though!
This video does a great job pulling away the wool that is over your eyes. Anyone scrolling down because they lost focus on the video should scroll back up and finish the video. The last part about monetizing your data and using it to incentivize "good" vs "bad" behavior is scary and real. I have added this to my favorites playlist.
Also, kudos to the creator for taking his own advise and adding subtitles to his video. The crappy AI generated subtitles implemented by TH-cam are a perfect example of immature AI technologies injecting themselves into our daily lives.
Conflating the censorious, privacy-violating mega-conglomerate corporations with the incoming AI boom that they are leveraging against those that rely on (not merely use) their services isn't elucidating anyone.
Also I haven't seen to the whole video yet but does he address the fact that not only is the majority of this research being conducted in the open but the industry is finally addressing the hardware problem with specialized accelerators, such that running very large models won't be the exclusive realm of quarter million dollar machines like the DGX A100?
And then what?
Lol you caught me
not enough data put simply, if you're listening to someone with accent its hard to understand him the same goes for ai trying to understand words, if it doesnt know it tries to come up with words that are sounding somewhat similar when the bro in the vid speaks too fast trying to find the relatable word becomes impossible with the hardware it runs on. its not "immature" the models are actually quite advanced main limiter is and forever will be hardware and how much data there is. even a bad language model you built yourself could most certainly understand you (you as in you when youre trying speaking to it) but maybe not others because they speak different use, different words.
No, it doesnt, it makes you think you learned something while in fact you were lied to and decieved, absolutely baseleslly. Its horrible how people can be so easily swayed without asking for further sources or facts.
@@TheManinBlack9054 If you want to convince anyone of anything, you need to elaborate.
Another very good video... You have a very evocative writing style. "A world for money to live in or a world for people to live in" is a line that I think will stick with me for a while.
Just discovered your channel and I think this video was one of the best on AI (and being a professional aritst, I watch A LOT of those). Yoir approach is very different from other video essayists and there are some points in this video that I've only seen you make yet! It's given me so much more to think about, reasearch, and explore. Hope your channel keeps growing cause you truly deserve it! Keep up the great work.
Ah, yes, pleasing the stakeholders is ALWAYS something that steers societies away from dystopia! Thank you, AI alignment "non-profits," for letting us know you're helping the little guy!
Yes and we know those in power whether in business or government always have our best interests in mind 😅
AI alignment is an important problem to solve, our entire race and future depends on it.
Please go and look up Robert Miles AI safety to understand what im talking about.
@@TheManinBlack9054whatever you say guy
@@TheManinBlack9054i will filter my consumption. I will listen to an expert whose analysis I feel I can trust (Jimmy Mcgee) and ignore charlatans with an agenda (random guy two levels deep in the TH-cam comments)
They're so nice and considerate, thank you to these legends for looking out for us.
28:50 I have seen so many people talking about AI who overlook this exact point. Art is all about those little decisions that capture the artist's voice. I couldn't have put it better. Amazing video.
Ironically, it's that same point that relieves me of the idea that AI could ever fully replace artists. Someone who becomes completely reliant on AI may certainly use (unknowingly) the same techniques as an expert uses, but the lack of understanding why the techniques work will be wildly inconsistent compared to someone who gives a damn about art.
When talking about writing, and AI being a threat to the field, one anon described AI as being unable to write, because it does not understand, it puts together sentences shaped strings of words.
@@MrhellslayerzI nwver thought it would ever replace artists but it will make it a lot harder for new artists to get to this point and support themselves and there will be an ever decreasing pool of actual voices to listen to, because the less viable it is to enter this industry the harder it will be to get enough experience to be able to make these choices effectively.
@@bluester7177couldn’t agree more. Very few people finish school as highly polished professionals. Most creatives in any field have to hone their crafts with lower level, simpler work, which right now there is a market for. But if people cant sustain themselves doing this sort of work anymore, then the pool of people who can afford to become artists will shrink to the independently wealthy and those with rich parents. Ironic that the people who claim to be “democratizing art” are steal real artists work to do the exact opposite.
That may be true, but how many people look out for it?
Oh no, an indie game used a little bit of AI art for its backgrounds. The sky is falling, I tell you!
Fantastic video!
One other thing I feel is important for the future is directing people to user curated sites where real people (even if most aren’t professionals/experts) help answer questions about different trades like how to fix very specific computer issues, car troubles, or how to do your own taxes. Most just accidentally find these (Reddit posts come to mind), but like Jimmy mentioned, Google is getting worse and worse about giving you a good answer to a search, so making sure people know how to find the truth is critical to surviving (that or some talented person makes an AI tool that can instantly filter out AI generated content).
the world needs more wikis
I use alternative search engines like Wiby or Search Marginalia. Those put me into contact with a lot of people's personal websites.
A GIANT thank you for this video. I am in full agreement with every bit of it, and you touched on many thoughtful and critical points I do not see being brought up at all in these AI video pieces. Thank you especially for the photographer analogy, and mentioning the role of human intent- the element that can only be mirrored but never truly replaced; it's the piece that makes art connect people, and endure.
Fierce pushers for AI arts take the creative process for granted, stripping the meaning of the mediums down to literal minimums (prettiness, cost, tech). If you value anything beyond that, they call it "human hubris." I think this attitude stems from ignorance, combined with hyper-focus on one area of tech, and delusions of grandeur. Sprinkle in a disregard and even disdain for those who are stolen from. Which... strongly reminds one of NFT's, and those who insisted they were visionaries pushing crypto as the future currency.
People must stop using a vital limb of human culture as their playground for automated blandification. Imo the competence of the AI (ex the extra/missing fingers) is not the big problem, here.
The big problem (besides people's literal livelihoods being cut off) is something you mentioned in your video: As a professional, I am seeing a startling lack of care, less skill, and ALARMING decrease in critical thinking beginning to rise in my creative field due to people using AI as a shortcut. This is a bigger problem than people may think. This is an attitude that should never be normalized, especially not during times like this. It's an attitude that spreads from producer to consumer. It encourages people to become militantly dumb, as though asking for more from the media we so eagerly consume is somehow absurd.
Liking something just 'coz is actually just fine. But AI art presents a hard wall that halts active analysis and discussion. You watch a wonderful movie, then you find out it was garbled together by AI. What more is there to even say about it now, besides discussing the novelty of that technology? There is no part of it that was trying to connect to you.
And how ridiculous is it that some people demand high praise for typing in prompts for art that's a mash of often stolen work? That makes me sound like an asshole, but what else can we call it? That's like someone punching buttons on a vending machine, and demanding to be called a chef. They can't seem to decide whether AI art makes it "easy and accessible for everyone," or if it's a special task that requires skill. They will change the narrative to suit the argument.
And no, gaining something meaningful from a film/painting/writing/song etc. is NOT the same thing as finding the beauty of a tree meaningful (something humans did not 'make.') Nor is it even remotely the same as 3D technology. I am tired of hearing these desperate analogies from people who are so bizarrely enthusiastic about devaluing human expression, and putting professionals and creatives of all stripes and kinds, out of work. They do not speak for us.
If this AI craze (not just for art, obviously) is really a bubble as some say, then I cannot wait for it to pop, and I'm sure I'm not alone in feeling that way. There MUST be better ways to apply this technology.
Big breath okay I'm going to sit outside with my dog... But I'm very happy and grateful that I'm seeing more videos telling this side of the story. It's the side that isn't backed by billionaires. And I'm ESPECIALLY happy to be seeing artists try to hone their skill and bring their personal expression into a space that often dismisses them as redundant at best, unwelcome at worst.
Your personal voice matters. I'm especially happy to see new artists. Please don't let this AI stuff discourage you.
The part where you talk about search results on youtube being mixed with useless recommendations hits me like a truck: I was trying to find your channel from that one Cruelty Squad video (since I forgot the channel's name, whoops) and the search results were mostly unrelated nonsense I watched in the past..
Being a software engineer working closely with ML, I quickly noticed that most jobs on the market were related to recommendation systems and other ways to manipulate human attention. So wrote a small assay, long lost, about hypothetic future. In this future timeline, most AI applications of our time is considered extremely barbaric and unethical because the culture has evolved and people see it for what it is. So, just like we have "medieval torture museums" in a lot of European cities nowadays, those better future humans would have XXI century AI museums.
Such a museum would have rooms or whole floors dedicated to different AI abuses. Maybe a whole floor for mobile games and microtransaction optimization with AI, a large section for recommendation networks to sell you stuff you don't need, and then walking further with your museum guide you enter a large hall with a Facebook timeline projected on a wall, scrolling down without ever stopping.
I wrote that quite a while ago being frustrated by social networks becoming more and more shitty by becoming better at that stakeholder value thing. The essay was also quite optimistic in describing the future where we've collectively understood how to use the power of ML for good, at least mostly.
I no longer have that optimism in me though. We're used to now owning things by now, and mostly used to not owning our personal data for convenience, and that's bad but manageable. It's much more frustrating to not own your decisions, to be manipulated by neural marketing and gaslighted by statistical objectivity of "but that's what the algorithm decided based on YOUR actions".
We don't have the discipline to watch our every move and every action just to project a correct image of ourselves. We switch between the state when we make conscious decision and the state where we let go and let our inner monkey have fun. The marketing department is only interested in the second state, and they're eager to sell it to you as your true self.
This summary of your essay sounds very insightful. I would love to read the whole thing!
@@ΑρτεμισίαΠλοκαμίδου Thank you. Not sure if I'll be able to find it (and translate), but if I do I'll post it here
That's a very insightful comment my friend. I think you are probably quite young and still might have some faith that humanity progresses towards something called evolution. That's a nice concept created some 200 years ago, but it turns out that idea is just another marketing tool to justify colonialism and race superiority.
The truth is that the vast majority of people are just well meaning fools exploited by clever crooks. AI is just another tool for the perpetuation of that status quo.
The war on ownership needs WAY more attention.
For some reason I’m kinda happy you are all getting so frustrated with something that is completely under your own control. Bunch of addicted losers lol
As an AI engineer myself, I feel obliged to acknowledge every concern you've brought up. It's like I've got a ton on my mind, but saying it all out loud feels risky. I worry about getting laughed at or pushed aside, or even dampening the vibe and being seen as a downer. I can't help but cringe when I hear bigots and parrots praising the AI Lord and all its blessings. But hey, here I am, just riding this crazy boat, not knowing where it's headed.
29:08 "... when you create something, you're constantly making tiny choices and I think an author's voice is a sort of harmony and consistency between those choices."
This quote perfectly encapsulates how I feel about AIs in creative fields. Intentionality should be a key aspect of any piece of art, so it makes sense to use AI when randomness is the intention (the movement of models in Rain World, for instance). Otherwise it becomes a crutch and leaves what you're creating feel vapid and uninspired.
But at least I can become a fucking prompt engineer.
Thing is how much does this matter for the average mainstream media enjoyer lol. Does it really matter if the next Thor has intent behind it
@@awsomebot1 Honestly, yeah it does. How much it matters is a little more dynamic, its like the frog in the pot. You cant put a frog in boiling water or it jumps out, so you put it in a pot of cold water and then turn the burner on, by the time it realizes its being cooked it no longer has the strength to jump out. We could see something similar with AI, and we are: its slowly creeping in, generating low quality media for consumption. Not in such quantities that it causes an outrage, but it will get worse. As AI media becomes more and more normal, we notice it less, which allows it to penetrate further. Eventually we're left with very little except for AI generated art of all forms, and it will all suck, and still nobody will care because it happened gradually enough to avoid a backlash.
It matters because right now the water is still lukewarm, we can still jump out of the pot, but not for very much longer.
@@wills.5762 You're saying it should matter. I agree. But reality is, for most people it will not.
@@awsomebot1 My affirmative answer was addressing your 2nd sentence, otherwise yeah, I'm saying it should matter.
its rare to find an hour long video essay on youtube that actually says something.
There is so much hope in the simple phrase "Things don't have to be this way." Thanks for the reminder and keep fighting the good fight.
But they are, and let’s not pretend that mega-corporations like Google and Amazon give two slivers of a shit about us, look at how they treat their *own* goddamn employees.
This video is fantastic all around. There's so much in it that it's hard to comment on anything too specific, but it was very solid from start to finish. Thank you for making it.