AI Deception: How Tech Companies Are Fooling Us
ฝัง
- เผยแพร่เมื่อ 28 เม.ย. 2024
- Visit our sponsor aura.com/coldfusion to get started on a 2-week free trial and see where your personal information is being sold online.
AI is everywhere. Is it all it's cracked up to be or is it one big scam? Well, as it is for most things, the truth is somewhere in the middle. In this episode we see how, despite some of AI's useful contributions, the hype has also impacted the tech industry negatively.
Sources: docs.google.com/document/d/1V...
Excerpt on AI in laptops provided by Just Josh: / justjoshtech
ColdFusion Podcast:
/ @throughtheweb
ColdFusion Music:
/ @burnwatermusic7421
burnwater.bandcamp.com
Get my book:
bit.ly/NewThinkingbook
ColdFusion Socials:
/ discord
/ coldfusiontv
/ coldfusion_tv
/ coldfusiontv
Producer: Dagogo Altraide, Tawsif Akkas
Writers: Dagogo Altraide, Tawsif Akkas
Editors: Tanzim Uddin, Brayden Laffrey - วิทยาศาสตร์และเทคโนโลยี
thinking AI is doing your self checkouts and then finding out its just 1000 people in India watching the cameras 24/7 sounds like a good south park episode
🤣😅
Go read about the UHRS. Your search engine algorithm is being "trained" the same way 😂
AI : Actually Indians
Lmfao 😂 @@harrisbuild
Frr
AI: All Indians
That's brilliant 😂❤😅
Lol
Damm me human 😂
Active Indians
😂😂😂
I work in an aerospace MRO company and the marketing team is now labelling EVERYTHING a computer in our company does as “AI powered”…
Literally techs that existed decades ago (FEA simulation, self-filling forms, automated tool management system, data analytics tool, etc) is now “AI powered” just because they r jumping on the trend.
Management is now even encouraging engineers to consult chatgpt for our calculations… while i know for a fact it cant even convert some units correctly.
This is ridiculous
I've used ChatGPT for some summaries on scientific sources and the same with Bard. They literally invent sources that do not exist... No idea how someone can trust those things.
Indeed, LLMs are prone to factual hallucinations and should never be used for complex calculations
Recently I came across the term "Generative AI based design solutions" which is nothing but good old topological optimization.
Are you making parts for Boeing, by chance? 😅
@@TBolt1 no haha we deal with maintenance and spare part installations, so after-market stuffs.
AI has replaced between 60-100 people from the company I currently work for... And complaints from clients have increased..... Great informative video. Thanks 😊
lol, Go AI go Broke
Of course they have- company is dumb for thinking the transition would've been smooth
That's why it's important to enrol in Google's IT support 👏🏻
AI will improve considerably in a very short timespan. Humans won't. This is a moot point. It's akin to looking at the first cars and saying "Look, when they crash there's a far higher chance to die than if you just have an accident with a horse!".
@@LeoVital and? the goal of those companies is to make AI Autonomous,
they're not planned on making people to operate AI? because what the point of that when AI can Generate Prompt and keep churning stuff without needing input from anyone.
it's like an assembly line
you're the victim as well, and UBI won't come because it's a financial suicide.
without a job or money, are you prepared to sleep on concrete?
What amazes me the most is the fact that Amazon was trying these "walk out" stores in the US where people already perfected this type of shopping.
🤣
Hard to beat that one!
Especially in New York now lol.
The old "five finger discount". 😂
This is gold 😂😂😂
Let's not forget, Mark Cuban was 100% all-in on NFT's and defended $43 million Bored Ape prices.
This ^^^^^
Mark Cuban is an awkward clown.
😂
What's crazy is that he still defends it. I feel bad for those who blindly trust billionaires and end up betting their house on such stupid things
Mark is a fraud
When an industry focus on announcing eye-catching gimmicks rather than trying to solve the long-standing, fundamental problems(like instability and hallucination) of the technologies itself, you know this is a total scam.
This is true of all computer dev. If something is a buggy mess and they just keep announcing new features, it's because the project has no intentions of ever releasing.
Exactly. Look at the world using Microsoft OS when it's more broken every update... LOL
wow, incredible take! no one ever thought of that!
because those are fundamentally unsolvable
In a US university a PhD scholar submitted her university PhD thesis , after a month she was fined and lost her degree for plagiarism, She had used Chat-GPT to write her thesis
My problem with AI is it is fundamental a statistical model based on large amounts of data. The key is the quality of the data initially used and used to update the model. It doesn't actually think in any real sense of the word. AI is very susceptible to "Garbage in, garbage out".
How are humans any different? Garbage in, garbage out applies for humans too.
Correct! AI is just a statistical predictor.
@@herp_derpingsongarbage in and garbage out humans are useless for business and productivity. Similarly, AI is useless if it generates rough output but at the bottom a disclaimer is written "Generated results may not be accurate". AI is also useless if it gives garbage out for garbage in.
True dat.
As shown by AI projects that ended up being racist. The good old Chan of 4 figured out decades ago that these things are only as good as the information they're served. Back then it was just chat bots, but the same principle applies.
They're not sentient, so all of what they are is based on what we tell it to be.
"Full Self Driving"
...actually just some dude halfway around the world remotely turning your steering wheel.
Thats exactly what Waymo is doing as we speak.
Yep, FOOL self driving is exactly like that
And that dude is used to driving on the other side of the road.
lollllll
I know this is a joke. But have you used FSD?
I worked for an accounting company from France that used hundreds of cheap laborer from Madagascar to manually input data from photos of actual bills people uploaded believing it’s OCR. Nowhere they mentioned it’s all done manually.
Wait isn't OCR old technology? You can upload your receipt to GPT-4/Claude 3 and its vision features can extract receipt data.
@@prestonr6348 I don't know what they are doing now, this was in 2022.
@@prestonr6348 its not good enough for 100% coverage
@@prestonr6348 I worked for them in 2022. I don't know what they are doing now.
damn thats so personal
Cleaning toilets for a living might be the shittiest job I've ever had, but at least AI is unlikely to replace me.
"introducing, the self cleaning toilet!"
@@MBunn-uf1we Even the robots wouldn't touch some of the things I've had to scrape off a crapper 😂
AI is like 30 years ago when Intel used the term "MIPS" to promote CPU performance. MIPS stood for "Meaningless information to promote sales"
I programmed in MIPS in one of my electrical engineering classes for microprocessors a long time ago. It was hard and tedious.
@@Marrow9000I suppose that is why x86 codebase for Windows was the only one that was stable even though Windows was originally intended for MIPS because they thought it would be the future
Wait, Amazon did something dishonest and unethical!? Gosh that's so out of character for them...
Right?!!!
Damn, this must be a one time occurrence! I doubt it would ever happen again, they are obviously the must trustworthy company in the world.
Am I the only one that refuses to support their business model? I feel like every other door has 1-2 packages from them every day.
Shhh Bezos is listening!
Pssssh... Amazon doesn't give people jobs, Amazon bad. Amazon gives people jobs, Amazon bad. Can't they do anything to please you people? Leave Britney Alone!
Just wait until they find out that every Alexa actually has a tiny person in it.
Actually yes, it was third world interpreters. Look it up
Smurfette with a megaphone
Black Mirror had an episode proposing something similar lol
😅😅😅😅😅😂😂
😂😂😂
I would go further: we don't actually have AI. We have machine learning, which has shown itself quite good at faking intelligence, but not actually intellifent. ChatGPT, under the hood, is an overpowered autopredict.
Overpowered autocorrect is still pretty nifty, but it falls way short of what all the hype would have people believe.
This! It's a mix and match data sorting machine on steroids.
But it's not sentient or sapient.
I'm scared of fungus computers though.
Those might become self-aware.
@@oompalumpus699 probably not
We don't have human intelligence yet.
The typo and nonsensical word "autopredict" in this comment is just so deliciously ironic.
Bro even edited it.
A friend of mine is an electronic engineer with +30 years experience, so he started working with paper and pencil.
He once told me :"now with a computer I can do in a few hours what once would have taken me days of work,
but I still have to stay at the office 8 hours every day".
I don't know what's gonna happen with the job market, but one thing's for sure. The rise in productivity is goign to go all in the pockets of the proprietors.
What did you mean? your friend says that the computer allows him to do work in a few hours, but he deliberately does not use the computer? does he use pen and paper instead? so he increases his time in the office?
just there's not gonna be a rise in productivity with this one
@@user-bm4yf6td7d I mean that legal full-time job contracts require a person to work 8 hours a day back then as much as they do today.
On the other hand productivity has increased enormously because of the advent of the computer. Just compare doing multiplications on paper to doing it with a spreadsheet.
So the point is that even though you produce much more wealth through labor nowadays thanks to computers, the daily routine of the average worker hasn't changed, 9 to 5 it was and 9 to 5 it still is.
@@user-bm4yf6td7d Three decades back, what an engineer could achieve in an 8 hour shift, takes only 8 minutes to do with today's computers, however the engineer still need to be at the work place for 8 hours. Similarly, after some years, what today's computer does in 8 minutes, could be done by AI systems in 8 seconds, still those who utilize these in their profession, would need to spend as many hours at work, as they do now. Improvements and innovations in technology is not decreasing the total efforts what humans need to put in life. It is not increasing anyone's leisure time. All gains go to performing more complex tasks in shorter time frames, and more such tasks.
I worked with a company offering AI-powered coding services. While their AI models worked well 85%-90% of the time, it still wasn't good enough on its own. The models would hallucinate small, but important details that made the code unable to compile or unsafe to run. This meant that *all* of the code had to be manually reviewed, edited, or rewritten. It wouldn't surprise me if some executive decided to replace all of their developers with AI, realized that he screwed up, and then rehired people to do most of the work they were doing before under the guise of "supervising the AI".
I think the point is that you can replace 5 juniors software engineers with 1 senior software engineer
AI is not ready yet to replace humans where it comes down to language and rational/logical reasoning. Even GPT-5 may not cut it, as it remains an input-output system.
What company? What software? Any data to back that up?
@@dk109k2dask9 I very much doubt that. LLMs struggle with basic logic. Often the error they make, render the code completely useless since its just the wrong approach.
@@dk109k2dask9 short term it works however, not having fail safes makes that 1 senior position dangerous due to... 1. being overworked, creating more opportunities to overlook mistakes/make them. 2. consolidation of all that work makes replacements harder to come by, meaning if someone quit for higher wages the replacement will most likely be worse due to starting salary being lower, and needing to learn everything. .... It honestly feels like AI in the long run if used for corporate gains only will be a mechanism of enshittification of the cogitive experince of work... and if left for long enough people will forgot what good service was... and accept.
I am a programmer and I have fully embraced using AI technology the moment it was available. Working inside of these tech companies I can comfortably say that AI is not the reason for tech layoffs. In fact the majority of the companies I work for will not even allow the use of it within the company. The bubble may be making companies more comfortable greatly cutting staff in the promise that they don't need those workers but its not reality. Most of it is being driven by market uncertaintay due to insanely high interest rates to curb inflation. Its a sledge hammer that kills the entire tech industry.
It's also the end of the pandemic, so people aren't spending as much time with tech products while not working. And the VC funding bubble popping due to higher rates.
Thank you for sharing, I think AI is the scapegoat so no one panics about a slow down coming
for me, i used to work with developers and at times interns, to develop features that company needs, we plan and get dev or intern to develop
lately, what i need, i need to plan, write a prompt, or multiple prompt, make very minor finetune, and ai (chat gpt 3) does the code.
its not perfect, hence the multiple prompts, but it gets done fast and with intern, sometimes after 2 week either they cant do it, or its done but not exactly what i want or i just hear technical BS how its not possible..
it scary, but its the reality
Interest rates are ok, they were insanely low before.
I'm also a programmer and that's exactly what I see. I'd also like to add that it doesn't replace a junior dev, I use ChatGPT-4 daily and at most it replaces google search but it's often wrong and I have to resort to google because it makes up bs.
there was a little tiny slip of the tongue moment in this video and I'm so glad you kept it in the final cut. it's such a comfort to have little human touches like that while listening to discussion about something as cold and emotionally removed as current ai
thanks for another great video!
What was the slip?
Yes, wonderful slip, the funniest and most thought-provoking slip I've heard in 20 years. Please don't tell anyone what it was, let people discover for themselves (if they can).
😂😂
@@swagatpatra2139😅😅😅
@@binkwillans5138 now I can't sleep at night until I find it
The older one becomes, the more it seems that most big business involves lying, manipulation.. even fraud.
Yup
Congratulations.. you're starting to catch on!
That's the game if you want success. Otherwise its best to remain silent
yep, capitalism is mostly about exporting the work to 3rd world countries that pay less and then using psychology to trick people into buying an overpriced product that they very likely dont need in the 1st world.
you don’t need to get old to realize this
The biggest sign we are in a bubble is experts saying we are not in the bubble.
I don't know if that's how it actually works but you got my upvote
Yes, same thing happened during Dot Com rise also.
Too much leverage from cheap money.
@@Tubeytime
So, building on your argument Sturmeko, what is the scenario that we are not in a bubble?
@@KevanMajere33when the experts say we are in a bubble, then we’re in a cube.
I feel like a huge part of AI development that is missed in this video is how the models are largely trained on stolen work. I don't mean in the way that workers are paid to do work for companies, I mean usage of copywrited content without the consent of the copywrite holder. You can get chat GPT to spit out lyrics of various songs using key words, and phrases from various websites. The only reason it is viable for these companies is due to all the work of others they are taking, and hoping they can displace them fast enough before regulations and fines catch up to what has happened.
If they actually paid for the work they are taking, it wouldn't be even remotely profitable to run these systems (not that they currently are profitable in most cases).
In my mind, that's what makes this a bubble. It isn't just the hype, it is that they are flagrantly breaking existing law to the point that they should be fined into non-existance in most cases.
Yes, precisely this. It's profitable because law tends to lag behind tech. It's plain stealing.
I lost faith in the AI revolution when the NFT/Crypto bros started moving in.
You went to school? You‘ve learnt from books?
That's not a case they'll win. Any person can use the Internet to view publicly available web pages and learn from them. Surely an algorithm / AI has the same freedom.
well here's to hoping for the nytimes case
Fascinating look at the struggles and deceptive practices around AI implementation. It is disappointing to see how some companies opt for shortcuts and hype to misinform public and investors. We need more transparency and ethical practices in the tech industry.
You are an AI.
@@rajeshranjan3199 wake up, Neo...
The Matrix has you...
Bruh. Did u use ChatGPT to make this comment?? 😂
Bruh. U used chat gpt to make this comment. 🤣. It’s a nothing burger.
they must rely on shortcuts because the technology simply can't support their use cases on it's own
Honestly AI is the biggest "Fake until you make it" i've seen in a good while, it's sad but it's the current gold rush.
No! It depends on the industry and use case. Sadly, there aren't very many practical A.I use cases. It's mostly akin to a very sophisticated toy. However, a toy that can be very lucrative with integrity, sustainability and judgment. It's really mostly lip service with already establish corporations that gives it worth rather than the A.I alone.
@coreym162 'AI AI AI AI AI .... ' : the sound of wall Street
@@daporter8790 After i saw an "AI Rice cooker" i gave up.
Wow. That was the best and most concise description of this snake oil garbage I have seen.
@@coreym162 there are actually many good use cases. Summarizing, search, coding, material science, protein folding, developement of medicine and so on.
I think it's important the customer service chat bots swear when appropriate.
i would love it if an AI helper heard my complaint and said ‘godDAMN that’s terrible lol’
"Anyway, I'm kinda busy rn so take your complaint somewhere else alright?"@@WeeWeeJumbo
That's a joke, right?
In my country, there are literally freelancer jobs where you identify text in an image and put it in a document.
I looked into the company paying for these freelancers and they market their service as 'AI-powered' image to text conversion.
Thats an image annotator job, It'S the same thing the "Amazon check out ai is just 1000 indians" is based on, It IS Training an AI / ML Model, You are Annotating Images for their MAchine learning algorithm so it can Eventually do it solely by itself, reliably.
GPT 4 by openai was annotated by thousands of kenyans but the company they used now shut down because OpenAI can now use GPT 4 to annotate things for GPT 5 after reaching the performance threshold of a "low skilled worker"
Wow. Hmm privacy issues?
In a US university a PhD scholar submitted her university PhD thesis , after a month she was fined and lost her degree for plagiarism, She had used Chat-GPT to write her thesis
isn't that data classification for training AI models? AI needs to be trained off of verified data. People need to say "this is a cat" thousands of times for an AI to recognise what a cat is. Although identifying text is a bit of a rudimentary task at this time
❤Over the years, I think I'v fallen in love with that soothing warm voiceover at the beginning saying: *You are watching ColdFusion TV* ... then the same voice, saying: *Cold Fusion, its new thinking* at the end! ❤
Incidentally, Amazon operates a crowdsourcing platform called Amazon Mechanical Turk.
The jokes write themselves...
LOLZ
I thought the same thing about Amazon, kinda sketchy
I thought this was a joke but it's actually real.
@@roe_ lmao same
So Amazon just took local jobs and outsourced them under the guise of "AI/Progress."
Shocker.
Exactly
The only thing real about AI is how it's been used to fuck over the working class.
Never heard of 😉
I say an advert of either Twitter or Reddit yesterday by Samsung stating that their new vacuum cleaner, not an automated one like a Roomba but a handheld one, used AI. It's a handheld vacuum cleaner.
As a developer I can assure you AI has 0 to do with industry layoffs. Most people don't use it, it hasn't changed the way we work dramatically, it's literally being used as an improved version of an IDEs autocomplete function. Neat, but not revolutionary. The reason behind the layoffs is much simpler and less interesting: high interest rates.
Economic growth in the past 15 years has been driven solely by low interest rates. Now that rates are up, dramatically so in most countries, companies are forced to cut spending to service their debts and prepare for further economic turmoil.
Precisely. And most of those hires weren't necessary in the first place, literally fodder for DEI.
This 👍👍
This… you win for the best comment. So true
Finally someone who sees the light. The video made some good points, but with such a cop-out stupid conclusion. AI is absolutely not the reason for all these recent layoffs.
I think AI layoffs have more to do with Musk laying off 75% of the Twitter staff and other companies realizing they have thousands of unproductive workers in their ranks.
Such a great video, AI is such a marketing buzzword these days for anything where the computer seems to "think" for itself
"AI" is a marketing buzzword indicating that the company is paying to use OpenAI's technology in some capacity.
I hate it so much
I think AI will end up the same like Blockchain, Web3, NFT, and all other dead project.
here the thing, AI is Unpredictable and Uncertain,
because they often give you a false information, you'll be forced to Triple check each Results.
it end up becoming way slower than if you just google it yourself.
things like booking Appointment, buying ticket, and Checking schedule,
AI often gives you the wrong information.
also if you generate 10 times, 7 out of 10 they'll give you a different answer,
and i bet you'll stop trusting AI, and it's only as useful as a Party trick like Siri
and this will build a sense of Distrust amongst users,
where instead of taking Risk of being wrong or end up in embarrassing situation,
they would rather do it themself and not using any AI App.
like do you feel comfortable to book Hotel, rent a car, or buy a new Iphone using AI alone?
Then business will also stop using AI before they do a Major mistake, Marketing Blunder, Lawsuit caused by AI.
even right now, there are already ton of Self driving car accident caused by AI,
AI is more dangerous than a Traditional programmed Algorithm
I think AI will end up the same like Blockchain, Web3, NFT, and all other dead project.
here the thing, AI is Unpredictable and Uncertain,
because they often give you a false information, you'll be forced to Triple check each Results.
it end up becoming way slower than if you just google it yourself.
things like booking Appointment, buying ticket, and Checking schedule,
AI often gives you the wrong information.
also if you generate 10 times, 7 out of 10 they'll give you a different answer,
and i bet you'll stop trusting AI, and it's only as useful as a Party trick like Siri
and this will build a sense of Distrust amongst users
AI in itself isn't even that big of a deal. It's been around since the 60s. AI is just a slightly fancier linear algebra algorithm, at the end of the day.
Mark Cuban saying we aren't in a bubble is the biggest indicator that we Are in a bubble 😂
Yeah lol. There would be AI IPOs left and right if interest rates were low, that's why they aren't. Existing, large companies are definitely hyping it up
So really he could say anything and it would indicate that we're in a bubble?
I didn't really get it. What actually happens if this bubble pops? Can someone please explain
@@venkateshtelu7815 A bubble is when the value of something is speculated to be much higher than it should be, and this happens because people are valuing a thing for how much more they think other people are willing to pay for it in the future.
People don't understand that the value of their stocks are not simply StockQuantity X MarketValue, and they don't realize that behind the market price, there are many orders placed for stocks to buy below market value and to sell above market value. People mass selling their shares at the market price causes the highest buy orders to be filled automatically, so when too many shares are traded too quickly, the market price will drop. If it drops too quickly, people will notice and either panic sell or "buy the dip", but if overall sentiment is that they would prefer to get out now to either pay off loans or play it safe, the price plummets and the people that "bought the dip" end up "holding the bag" i.e. they overpaid for something people don't value anymore.
tl;dr The market price is an equilibrium determined by buyers and sellers, and when too many people speculate on the value to be very high, this inevitably leads to a mass selloff and price drop because you will always have wealthy investors that understand that the bubble can't go on forever, and some big investors took out large loans to invest that they need to pay back.
lol
Before 2022? I feel old... I was using LSTM neural networks back in 2016 since that is what was THE THING to use and what replaced my general reccurent network models. I read the "Attention is all you need" paper when it published but never really used it activly. Before that I made my own frameworks for ANNs in C++ with genetic algorithms back in 2010.
What got me started was a neural net image classifier winning a competition in 2009-2010 where they classified handwritten digits from the MNIST dataset. The graphics that was made in a video about it here on youtube regarding how the net "sees" these numbers was amazing and changed my life. I knew I had to know more about and have loved ANNs and AI since.
Deep learning is overrated to be honest. It's just a pile of obscure models that can hardly be deciphered.
Thank you for this! I have been watching Coldfusion TV for a while now. The amount of research you put into your pieces is really astounding. Obviously the mechanical Turk was an embarrassment to some really powerful people, yet there you go showing picture after picture of it be used against them! Also loved how well you explained the term “AI washing”. Great job.
I plan on "Block Chaining" my "A.I." in order to "Disrupt" the industry with "Agile" "Dev Ops". This will create "Synergy" with my "Cloud" services until I can "Pivot" my "Digital Transformation" to an "Internet of Things" model so that I have the "Bandwidth" for the "Bleeding Edge" of "Disruptive Technologies".
But can it "bridge" to "smart contracts?"
Dude you'll be a trillionaire in no time! Can I buy some of your stock?
Damn bro, you could put that old WeWork CEO out of a job! All you need is a turtleneck and cool hair and you got it 👍🏾
Don't forget Software as a service.
Bullshit Bingo Bonanza!
I think the largest hurdle for AI is going to be reliability, especially if it is going to be integrated into engineering / healthcare. I don't want to constantly have to baby-sit, fact-check, and correct output from these models. If I'm going to be paying a subscription fee that's more expensive than Netflix/Disney+/etc (for my work, not leisure), then I'm going to want these models to not just write correct and efficient code, but secure code as well. Without more reliability, A.I. will always be a tool for people who don't know what they're doing, which is fine if you're a hobbyist, but not if you're a professional.
Don't worry. Some overworked folks in a Data Center in Mumbai will review the output beforehand
Exactly the point I'm bringing up as I have the impression that AI is good when it doesn't need to be precise. I tested it with a simple question (how many times further away is the sun compared to the moon from earth) and it came up with the wrong result. But it tends to work when there is some leeway like language, pictures etc. Bottom line: It's painful to trust a system when you know that from time to time is gives erroneous answers.
Well spoken. AI has been praised for 30 plus years now despite it still being bad when industries are trying too hard to make it replace humans in every activities
I would argue the opposite point. Due to hallucination, Large Language Models (not all of AI) will continue to require human supervision. But you need experts to discern hallucination, not "people who don't know what they're doing". So for the time being, LLMs are most useful for professionals.
The question is if double checking it is faster than doing it yourself. Those are the use cases, few as they are.
As someone who works customer service for a tech company, I'm planning on going back to school in 2026 to get a new degree. I've already survived one round of layoffs, but I don't anticipate surviving many more. Our leadership has promised profusely that AI won't replace us, but they're already trialing it on some of our tickets. One of my close friends is on the team that's testing it, and he's horrified because he knows he's helping train the machine that will eventually replace him. But that's the project he was handed so he doesn't have a choice.
All of my siblings are very into tech and the future of tech. They remain purposefully oblivious to why I'm so scared. There's the chance that the bubble pops and it's all hype and dies away. But I doubt it. Companies smell a savings. And they don't really care who or what they destroy in the process.
True. Corporate leadership only looks out for the balance sheet. Never trust them to look out for your well being. People need to look out for themselves.
@@Iron-BridgeThe shareholders also need to see that growth no matter the cost
Smart to anticipate. Jobs being replaced by "machines" isn't new. Not that long ago a farm had tens or hundreds of employees. Now they can run them alone or just a few people.
Go get a degree in managing AI dude. :) boom. Always gonna be needed. 😊
That's why the handyman coming to repair a leak makes as much as a software engineer lately 😂
I'm old enough to remember when Transputers, Robots, and Expert Systems were the tech darlings pre-public-internet.
I'm 43 years old. I lived in a world with and without the internet. I'm not going to lie, I worry a LOT about how things will be when I'm actually really old. Scary times.
Yeah, i also livet before internet, and remember the start of internet, it was wild and fun .
@@Brato1986 remember? We had so much hope for the future! We thought people would be more tolerant because they would be able to look for information, education would be great, etc... We were promised flying cars and we've got this mess...
@@princonsuella_ I'm still waiting on my jet pack.
What do you mean? Everything is objectively better, people are living longer, we have better tech, and we'll be able to live in space!!!!! Seek and you shall find, seek garbage find garbage, seek excellence find excellence.
@@coenraadloubser5768 Sure, but that doesn't reach the MAJORITY of people on Earth. We have many enhancements and still terrible salaries, education, health and others. All this advancement and most people are still struggling. In the grand scheme of the world, few people are benefiting from all that.Lots of people are surrounded by garbage, with no perspective at all. You're talking as someone who has options. Many don't.
I worked for a company that claimed they leveraged AI powered system. It was just a room of people there was legit no AI. I went on to work for the same guys at a new tech company and they were doing the same thing except this time they laid off my dept and claimed the AI was doing it now lol
Your job was offshored to India, Philippines or some other region where labor and benefits are much lower
@timcasady4750 unfortunately I know. It was Phillipines they started working before we were laid off. It was a crappy situation and I saw it coming but rode it out.
Profit speaks.
My favorite is when the marketing people get involved and decide that anything that has been programmed or run by an algorithm is now AI since it follows computer logic, which is somehow more real than outsourcing to a hidden group from a 3rd world country
I wonder if they can use it for processing Guar Gum. That is used in a lot of foods as a thickening agent. It is used in ice cream and other foods that you consume. It is also used in fracking to get the sand and other proppants down the well bore into where it needs to be. Interesting stuff!
You forgot to mention what Google did with the Bard Gemini AI chat , people thought that the AI was talking in real time because it sounded so casual and natural , but it was a lie, they manually prompted the responses and added them , their excuse was " We wanted to show how the tech will look like in the futue " 😂😂😂
WHAT!!!!..Who the hell did i chat all my feelings with.
@@MrRahulKumarKandula You're chatting with the a.i. It would cost a nonsensical amount of money for people to do it
news be like "AI is the future!
AI in reality "sorry I can't do that for you I'm only a chat bot"
Remember that next time the "news" tries to tell you anything at all
_Sorry, I can't answer that question, I'm not a biologist._
This was an AWESOME conversation. I appreciate you and your team for educating me about AI. A lot of what you said provided food for thought.
This is a conversation that all of us should be having. Let’s be honest, corporations/company’s (different industries and in tech) don’t have a great track record for ONLY creating positivity for consumers or their work force (e.g. Enron, DuPont, Goldman Sachs, Boeing, etc.). So hearing CEOs were not being fully transparent was NO SURPRISE (NDAs are a thing, right? 😂😂).
Continuing to educate ourselves about the subject and being real about what humans are capable of (Good, Bad, and Evil) is one way to not be taken by surprise. Keep your ears to the ground in your company and have your own side hustle, because no one is immune to a pink slip 📝 ❤
Same things. I see many companies labelling something ai when its just some basic python function.
Lots of them are also just wrappers for chatgpt, which I find hilarious.
Yeah an algorithm that changes from feedback is now AI
I think the real takeaway is how awesome & productive 1,000 Indians can be!
In all fairness, Amazon should argue that “AI” actual stood for “All Indians” 😂
Kek
😂😂
That's what they'll say if taken to court😂
I think the 'all indians' where checking that the AI works. It seems to me that one store would have spent less money on few cashiers.
More like Anonymous Indians 😂
You have the best intro to your videos out of anyone on TH-cam.
The song at the end had me thinking "eesh, not sure why Cold Fusion would stick this on the end" then I read the it was made by AI and everything makes sense. Those music making AI's that are being used recently feel like they will be even more impactful in general life/pop culture than anything visual AI created. They hit differently, interesting to see what the next CF video has to say about those.
kinda doubt they'll even be able to outperform stock music
*FANTASTIC TO FINALLY SEE PEOPLE CALLING THIS HYPE OUT.*
I'm an aerospace engineer (by degree) who works in industrial control systems and the staggering amount of nonsense I hear on AI is ridiculous. In the early 1990s I had neural nets and how they actually functioned explained by an electrical engineer who'd worked on them while at University. The origins of neural nets goes back to the 1950s BEFORE the emergence of digital computers.
The first generation neural nets were fundamentally analog computing devices set up for complex calculations. The YT channel Veratasium did a good video on what analog computers are, why they were used in the past and why they have potential applications again today. AI (as we know it) is basically digital versions of analog computers and being digital they can be scaled up to immense complexity in a reliable and repeatable way.
The most important thing that I see people misunderstanding is that these AI systems DO NOT THINK, but instead they CALCULATE ANSWERS to very complex statistical math. We see this most often when the search box on YT or Google starts giving possible questions as we start typing. Its *CALCULATING* the most likely search to what you are typing.
I BLAME the promoters and media pundits for this misconception.
Agreed... but firing all these people probably has even less to do with AI... I'm no economist but all these companies sales and revenues have slowed. THAT is the likely culprit, these massive companies are thinking scale and employee pay is a large cost for running a business.
As for AI, I don't think it will replace any jobs (even programing jobs), these jobs will always be TOO specific for them to be replaced by a computer.
@@qazmko22Also what the real economy is doing and what the bs money print part os doing are 2 different things
@@qazmko22proof your claim. All quarterly reports state otherwise. Shrinking sales and revenues? Where? Maybe at your grocery store but for sure not on the stock market (Q4/23 and Q1/24). That’s just a false claim.
They do think. Unless you are in the philosophy school of dualism, then the human thinking is nothing more than calculating asnwers to very complex statistical math as defined by the complex structure of our brain. There is a reason they are called "Neural" networks.
You really like CAPS LOCK
It’s almost like how 4-5 years ago every single company was investigating how to use blockchain. It’s just the latest buzzword that everyone wants to be a part of even if they don’t really have a way of actually using it.
omg yes I remember a clip from some reality TV show where they follow some rich young socialites and they were talking about setting up blockchain for their businesses over dinner and look where we are now...
Block chain didn't cut 30% of tech jobs did it?
@@ziggs123Neither has AI yet. That's just the excuse large companies are using to justify layoffs they're really making because of high interest rates and a coming recession showing how most of them don't have viable business models.
@@PXAbstraction I mean, they aren't, they're just saying its a shit market to retain excess talent and we need a cull in the ranks.
A blockchain is a highly specific tool that resolves a very particular problem. AI systems will become better and better simulants of human cognition eventually supplanting the need for them - that day isn't today nor even years from now, but it is coming in the next 30 years and the fact that it will happen in our lifetime should tell you how you should position yourself.
Glad to see you make this - there's far too much "AI hype" going on, and the media remains predictably happy to parrot corporate press releases and marketing, without review. Tech hype cycles just keep repeating and getting bigger and bigger.
REALLY REALLY Good video! I don't think AI is causing layoffs as yet... at least not in terms of the most recent progress in generative AI over the last 2 years but I think we might be seeing more subtle efficiencies brought about by process/job analysis and software solutions to simply get stuff done better. A classic one (which is actually AI) would be the call center AI's that take your name and account number. Likely these get your details up on the screen of the agent when you speak to them, likely cutting a minute of say an average 8min conversation. This means something like a 13% efficiency in call throughput for agents available, meaning it's possible to cut 13% of staff from that section. This is a fairly extreme example though. I genuinely think most of it will be around simply the adoption of tech for efficiencies rather than AI.
“Just because something isn't a lie does not mean that it isn't deceptive. A liar knows that he is a liar, but one who speaks mere portions of truth in order to deceive is a craftsman of destruction.”
― Criss Jami
It's just lies by omission
@@whannabi is right. Sounds like this philosopher has never heard of a “lie of omission” before. There’s lying through commission, which is the typical lie (committing an untruth) and then lying through omission which obviously is disregarding details to sell a narrative.
Aka sales people
I still remember just a few years back when everybody was all crazy about "cloud computing"
I just watched a documentary here on YT about clockwork automata.
Even back then, the inventors claimed their machines could think like a human.
For example, a clockwork chess player automata.
Turns out, there was a guy under the table operating it.
Also, people back then called each other NPCs.
That is, an insult for a person who is more machine instead of human.
Cloud computing is real. But it is not much more than moving code onto managed servers.
Do you know how many moving their applications back to premises these days😂?
How many ??? @@dtsh4451
@@k.h.6991 never used paas then?
I remember listening to an earnings call where the CEO proudly stated: "we are now using generative AI to predict inventory levels".
You are asking a chatbot how many boxes of Kleenex to buy for the quarter, also machine learning has been used for inventory management for a looong time now, nothing wrong with being late but don't try to use it as a marketing gimmick to prompt up your stock price.
Been following along that better part of the decade! THX a lot Dagogo!
the layoff come mainly from the very high number of hires in the previous 5 years.
0% interest rates made it dirt cheap to hire anybody for any R&D project.
Now that interest rates increased, those companies fired the non-essential workers in 2022 and 2023. And now, they also stopped a few R&D (or moonshot) projects to focus on money making projects resulting in more layoffs.
AI has nothing to do with that. Even more, those big companies hired a lot of people to work on AI related projects because that could result in so much revenue.
Also, if you've ever used those AI language models to code or to do anything other than basic stuff, you'll realise that it cannot replace even part of the jobs of evn junior employes
Don't forget the forgivable PPP loans that most companies took out to buttress uncertainty were actually used to over invest on the expectation of further loans, those have likely dried up by now
Anyone who worked in manufacturing during the "downsizing" years can spot the hole in the plot is actually the tech firms are recording massive profits because they are laying off highly paid human workers en masse. At a time when profits are not presentable to stock holders, this is a typical quick fix to make it look profitable. Hiring back workers for half the salary keeps the ball rolling just fast enough to make it look promising. And of course all profits continue to buoy the CEOs.
I think you are mostly correct, but as far as coding goes, you can replace a small percentage. For example, you would need less junior devs to do the basic tasks. Which is actually a bad thing long term unless AI completely takes over. It's bad because jr devs need to come in and do the trivial tasks to learn.
Agreed in my company also most layoffs are from divisions that were dreamed up in 2010s and now are not viable enough like AR and Gaming.
They haven't even done a single layoff in money-making divisions, by little but they are even hiring.
Most of those big companies are overstaffed with mostly useless people, for an example the case of X ..
My job is rolling out AI, but the outpost has to be manually reviewed and the AI will have to be constantly updated anytime time there’s a change (we have changes all the time). The older leadership teams just get so excited about AI not understanding all the manual work required.
Maybe it will get better, but it’s not ready at the moment.
This has been my feeling from the start i work in tech and i found zero use cases outside of writing python scripts i had no need of.
Anyone involved in integrating this stuff within a business realises how few use cases there really are. A lot of the big productivity gains through things like code generation are already realised and from working with that I can tell you competent devs are going to have a lot of work in the future cleaning up the mess this leaves behind.
Its amazing that there are any use cases.
Any minute of manpower saved is a gigantic accomplishment.
I feel like people have been wrongfully conditioned to expect insane breakthroughts.
This technology will improve incrementally, like everything else.
Its inching up toward human-level ability, and every small progress is a mindboggling miracle.
@@carlpanzram7081 I agree. Its still mind blowing to me to be able to “talk to” a computer using natural language and I wouldnt downplay the technology at all. It is the companies producing the technology that are creating these expectations though.
@@Hiroprotagonist253 thats just to lure investors and hype the hype
Yeah I tried hard integrating it but it's rarely helpful. I find it very useful to come up with excel formulas.
@@darkstepikit's gonna crash soon
When I found out that Amazon A.I. store was a lie & just 3rd world tech support watching through cameras I was the least shocked person on earth.
Living in a place that would never have one, I thought it was real for a limited market that would never scale much. I like the true version better.
3rd world tech support!?
More like labelling data for ML.
They were probably data scientists or engineers.
Most of LLM still need human intervention for proper learning.
None of so called AI is completely autonomous, they're made to learn on large amount of data under human supervision and still can make huge amount of mistakes.
Humans were used as a back up system, not as a system.
@@iscifion7122 Damn, someone here with an actual brain. You are spot on!
@@TheManinBlack9054 70% intervention rate my guy. After what, 4-5 years of operation. Only ever in intent were humans not the system.
In the late 19th and early 20th century, there were a lot of products riding the wave of this new fangled electricity thingy.
My personal favorite is «electric underpants»
"Devin" has made quite the splash in the management community. When developers checked the site code it appears to be much closer to a high end scam. Access is quite limited.
Time will tell.
I went to a Google event last week, and Gemini gave this as a response 'Nike is the sponsor of Real Madrid'. There was a spanish guy in the room who almost choked to death laughing.
I don't get it...
@@MarvinPowell1 they've worn Adidas since the 90s and as far as I'm aware have never had a kit made by Nike. Their rivals Barcalona have the Nike deal
@@MarvinPowell1Real Madrid is a football team (aka soccer in usa) in the Spanish league,which is sponsored by Adidas not Nike.
@@MarvinPowell1 The AI mistook Madrid for Barcelona. In Spain, this is heresy of the highest order.
@@MarvinPowell1 wtf dude? you didn't get the joke, must be an American 😂
"Instead of using junior devs or interns, use a neural net that requires thousands of video cards." - CEO of a video card company
"Our cars will be completely self driving in two years" - CEO of an AI car company, every year for the past 15 years.
In reality, current neural nets are just a fancy search engine paired with a randomizer pulling pieces from its training data.
In a US university a PhD scholar submitted her university PhD thesis , after a month she was fined and lost her degree for plagiarism, She had used Chat-GPT to write her thesis
What’s incredible to me is that every company AI service I have encountered so far have been the most frustrating unhelpful tools ever. They don’t know what you want and they can’t problem solve at all. There’s no humanity there and personally I find myself getting much angrier when there’s no person on the other end bc I can cuss it out with no consequences. And then when you do make it to an actual employee they have no training and can’t help you either. I don’t think AI is the future. It takes too many resources we already have too few of. The tech industry has one of the most corrupt and morally bankrupt work cultures I’ve seen. It takes a lot of background infrastructure to make these systems work and truly I don’t know how much longer until the whole system collapses in on itself.
That's because the technology you're being sold doesn't at all do what the media hype is trying to convince you that it's capable of doing. It's not some magical thing where you ask it a question and it pulls an answer for you out of its "database of knowledge." Seriously, just look up how this tech works and you'll come to the same conclusion that it's "just a more sophisticated version of autocorrect."
It could be more impressive if funeral homes implemented this "just walk out" service!
It's in Beta. It went through human trials at the tomb of Lazarus. Still waiting for the full release.
Do you want a zombie apocalypse? Because that's how you get a zombie apocalypse!
Hey, What bout dis
we implement this "just walk it off" healthcare system.
Fully Ehh and Aye
"Self -burial" checkout services 😆
Dark Humor 😂😂😂
Speaking of false advertisement; robot taxi, hyperloop, solar roof tiles, bricks, thermonuclear resistant glass, and so on
.. It baffles me how some apparently are totally resistant to it...
Pizza shops on mars etc
solar roads ...
Thunderfoot!
Bricks!? I'm insulted on behalf of Half as Interesting, the premier brick-related YT channel.
Got it, bricks don't exist
Well done mate! Brilliant! You have a kanck of forming a view that is quintessentially Aussie. Love ya work, keep it up.
They used to be a term that my parents used when things like this happened and it was referred to as FALSE ADVERTISING! HOW IS THIS NOT A SUBJECT THAT CAN BE USED TO SUE COMPANIES AND HOLD THEM ACCOUNTABLE.
so AI stands for “All Indians”，huh？😅
Only two comments? 😂Best comment I’ve seen
Man this comment deserves a few thousand more likes
@@SimonsRandomRants You know the youtuber is Black?
😂😂😂
Whatever happened to Google's "Don't be evil" mantra?
Money
99% of peoples are evil non vegans
They were always evil; they just stopped being in denial.
Or stopped trying to convince us
That's gone since October 2015
The black-box nature of AI makes it worse than closed source software in that nobody can truly comprehend how the model really works, leaving a huge potential for errors especially when it encounters novel inputs. We should still be treating it as a hobby grade technology interest like 3D printing or fpv drone racing but because the powers that ought not to be want to exploit this particular technology to subjugate us, it’s being heavily underwritten and hyped. Machine learning still struggles to drive cars, a fairly specific task compared to what humans accomplish in general… Leaving anything to AI at this point is definitely premature and will eventually expose points of failure. At best it’s ready to be shadowing humans and being evaluated and trained, that’s the smarter approach.
nah it'l be iterating endlessly and self testing it's code until it begins developing reliably, then that's when thinks will go a bit crazy
The black-box issue concerns deep learning algorithms for the most part. Standard statistical algorithms such as linear regression, K-means, Decision Trees, Logistic regression are interpretable.
Funny that Amazon actually DOES (or did?) have a service called Mechanical Turk.
Which is human powered (and yes, very low paid).
In a big (like one of the top five biggest) IT company we used to say:
- "AI" is for the general public
- "machine learning" is for investors
- "statistics" is actually to hire people who know how to code it
(same company ended up firing 10% of its worforce because investors were scared of its potential lateness in "AI"...)
A big difference with the "first wave" of automation in the 19th century is that this time it WILL affect white collars too : doctors, lawyers, architects etc. Also creative jobs : why AI is not replacing dangerous, uninspiring jobgs but the most creative ones (composer, cg artit etc.) is beyond me.
Makes you wonder what 'innovation' really means anymore.
Well, the biggest innovator is actually: the government.
They fund the research in R&D and the corporations try to do the D part to make it into products.
While trying to pay as little taxes as possible.
Elon Musk trying to re-invent the train making it more expensive and complicated (hyper loop) for minimal or no gains at all in speed or service and efficency and just backing it up with tons of hype.
Basically that's modern innovation by tech bros 😂
It means charging you a monthly subscription for rubbish
@@mafiousbj Making highly complex, expensive and difficult to scale or replicate technologies whose net effect and net impact is insignificant, is not an innovation at all. Such techs and companies are basically modern day scammers.
@@mafiousbj Elon Musk in African. Not surprised.
Thank you! Everyone talks about AI as if it'll replace 100% of every task, while others point out that it can't. But like you said, it doesn't need to do everything, just enough that downsizing is possible. E.g. a team of 3 instead of 10... and of course, AI will continue to work toward 100% of everything, but it doesn't have to get there before it starts causing havoc.
The havoc will start when we get cheaply manufactured humanoid robots that can repair and build each other. Before that it will be contained in a virtual world, not interesting for the average person. But it won't stay there long.
Great video! I had a similar suspicion regarding the rational for current tech layoffs…
It’s also a casualty of AI actually meaning any algorithm. A simple, step by step, procedure for choosing or predicting something is AI, deep learning, LLMs, neural networks, robotics are all a specific kind of AI.
I have first hand experience with a company that I was trying to contact a representative as when I called it connected me to their AI representative Emily after repeating my request over and over I waited until a real rep could help me which wasn’t much better at this call was directed to a call center in the Philippines and because the reps English was not good they redirected me to another rep which was in Africa and once again the reps English was poor so this time I stated I would only speck to someone based in the U.S. - The picture if perfectly clear these companies are looking for the lowest cost to handle business and when AI fails they transfer you to countries with lower labor costs so the layoffs are only about profit
Well, at least back in the day companies actually cared about the customer experience and their workers and not just profits.
If you're going to criticize a non-native speaker's ability to speak English, you should at least figure out how to use punctuation.
@@TheManinBlack9054when was that? Big Companies have never cared about customer experience unless it affected profits in my nearly 30 years of life.
@@DarkGob He's not criticizing the non natives English speakers ability to speak English. He is criticizing a system that forces him to have to work with non native speakers, who in turn provide a worse service because they are non native.
If you're going to feel the need to white knight and defend people you don't know, you should at least figure out how to understand the argument at hand.
This has happened with several companies when IT services Cognizant replaced established professionals with cheap 8$ labor. After destroying many businesses companies asked for their old workers back.
The wallpaper is misleading! I thought Sathya Nadella was arrested related to some AI stuff😂
Great video! Nice to see a channel as large as this one talk turkey
According to Amazon AI stands for Actually Indians
😂 ☠️
@@tehcheezman7471 You know the youtuber is Black?
To be fair. That comment from 2021 is completely right. Absolutely none of these AI's "think" or actually try to emulate the human thought.
+1
What is a thought ?
That bot which started to cuss at the Customer when inquired about his package delivery demonstrates that the Algorithms/Machines/Neural Networks are beginning to actually feel like humans since a low paid Customer Service Agent working for a giant conglomerate or an abusive employer can become frustrated and lose it all if he decides to quit his job or get fired so he will begin cussing around at anyone whosoever comes in his way, be it an employee, a customer or his employer.
AI = Anonymous Indian
What concerns me most is who will control AI, where the profit will land, and how the power in the workplace will shift. My fear is that the concentration of all the power and all the money will accelerate exponentially, so a handful of people who do not answer to anybody and who are more powerful than states, will control everything. We are already halfway there.
Bingo, when they talk about how much AI will increase productivity one has to ask, okay but for who? Do a cabal of billionaires really want to make your daily life easier & more productive? Or do they want something good enough to slash thousands of jobs and add to their treasure trove through the savings? I think the answer is fairly obvious.
We’ve been there for a while
Do not fear my child. We're already there
I was stuck on a codewars problem. I had solved it and passed all tests except it would timeout... ie. my solution wasn't fast enough on large datasets. I gave my code to Copilot, explained the problem, and it proceeded to give me wrong answers 6 prompts in a row and none even passed the basic tests. Then on my 7th attempt it gave me back the exact example that I had given in the first prompt. Ultimately, it didn't solve the problem. I have seen it do some amazing things but not this time.
There probably just wasn't a good example to steal from github or some other source.
I love working with AI LLMs, reminds me of the days of flashing different ROMs on HTC One M7
First time here.This is whats missing from the A.I conversation, a balanced view. Thanks.
Most of what marketing people call AI, I call computing. There are no thinking machines out there, least of all at your local call center lol.
Nothing is more satisfying than coming home from work and seeing that cold fusion has uploaded a new video
Yes, while you have work, before AI takes it over and does it better.
One of the most important questions in my mind with the future of AI and where it will end up is down to cost and benefit; how will companies make enough money to justify the monumental cost of running the infrastructure behind this? So much money has been sunk into these technologies and the time needed to recoup these costs is quite significant and some of these core Gen AI companies are running out of money and new data to scrape. They constantly need more and more from investors. Is it just a matter of cooking the books and writing off the 260 million dollars worth of the newest Nvidia chips you need to run the models every two years?
This time is different. Well put. The bubble is going to be a whole bunch of bubbles popping and growing simultaneously.
Ayyy!! Thanks for the Joe Blogs shoutout, love that channel
I had this theory that it does give me bubble vibes - seeing all these companies stock rising by silly amounts makes me question how does that translate into revenue/ profits? does it boost their revenue by 200%+ or save 50%+ on expense? - I'm personally not convinced
Yus! Awesome video! I always look forward to your stuff! How do you make such informative and graphically rich videos? I'd ask AI....but...
We had the same bubble with the blockchain and crypto. Even though I have Bitcoin. We need it to do more before than just “limited amount of nothing”