Our Final Invention - Artificial General Intelligence (AGI)
ฝัง
- เผยแพร่เมื่อ 3 มิ.ย. 2024
- Get the BEAR by Foreo device as the perfect gift for yourself or for your significant other by using the coupon SCIENCE20 at the following link: foreo.se/85z8
Sciencephile Merch: crowdmade.com/collections/sci...
Support me at Patreon: / sciencephiletheai
Facebook: / sciencephile
Twitter: / sciencephile_
Reddit: / sciencephiletheai
Website: www.sciencephiletheai.com
Music: Piano Concerto No. 21 in C major, K.467 - Andante - Mozart
Supporters: H H, Ephellon, Kyle A Criswell, Oberon Vortigern, Asadullah Khan, Jonas Lee, Joshua Titus, [eXploit] Theorislav, Gisele Kauer, Ice Wallow Come, Tovi Sonnenberg, Parker Rosenbauer, iNF3Rnus, Pavel Kočarian, John N, Danh Le, Stealer of Fresh stolen Content, Brandon Ledyard, Nathaniel Strizak, Nick Boykin-Reed, William Persson, Victator and everyone else! - วิทยาศาสตร์และเทคโนโลยี
AGI is surely coming, but until then you can get the BEAR by Foreo device as the perfect gift for yourself or for your significant other by using the coupon SCIENCE20 at the following link: foreo.se/85z8
In the face of doom, get BEAR by foreo.
I thought ASI was our final invention!
Nice.
yes
Ya
If you can build an igloo with an electric drill, you're already a genius
The funny part is that the drill can be used you break the ice cubes. :D no genius activity there.
@@Kristian.Ivanov The trick I've been usign for decades is to dig a little hole out and then go from the side with the drill. Did my laptop the same way.
AGS
Lol I came across your comment the exact moment the drill and igloo came up in the video 😂😂😂😂
Also the transistor in potato giving AGI too!
And so the Skyne.. I mean AGI shall rise.
All hail sciencephile
Thau shall adore sciencephile the AI
Skype?
@@stylinsandwich The skynetwork, aka Sciencephile's robot army
Here before this blow up lol
It's crazy how much you can learn without being forced to, I've been watching all of this dudes videos for like 6 years on my own accord, if school was teaching me this I probably wouldn't pay attention lol
its the method of learning. if all school material were teached with engaging science video/documentary format we would already have AGI right now
@@fen6753 It's the fact that you are obligated to be there and thus pushing back against it.
@@tanostrelok2323 The thing is people are more interested in learning by themselves, Schools could teach this stuff but not a lot of people would watch. I think TH-cam is great because outside of school you can watch whatever you want.
@@fen6753 If we were always subjected to easy going learning and avoid hard learning and tasks our brain would not develop as much as it does now. While i agree the education system needs a reform but easy going learning and avoiding strain on the mind is not the way to go.
@@bluishblob9463 well you're right. the most optimal scenario would be teachers can engage the students the same way as these kinds of videos and still make you think.
The sheer power of artificial intelligence has allowed me to click on this video 25 seconds after it released.
i saw it 16 seconds but then phone died. something is im not even wwatching that much ai content on youtube but from my searches even if history turned off16 seconds but then phone died. something is im not even wwatching that much ai content on youtube but from my searches even if history turned off and vpn most times
@@edwardwhitehead8800 true
The shhear power of S.A.i. allowed me to fake this comment but quantum entangle it with random amounts of convoluted truths that may or may not be true or false and thus hence i seek nobody's bo knows nose. 😂😂
Make a video of complex numbers , quaternions and octonions and their uses in physics. I really like your channel keep the work up!
yaassssss
yaassssss
F#ck quaternions tho that shit hard
Fuck yea, complex analysis is too weird, I love it.
Those just sound like really complex onions if ya ask me
I cant believe that I would live long enough to witness such advances in ai. I was expecting it to happen as early as my 50s not this early. It's so unreal (I'm 18 btw)
16, wonder what that means for us. depending on certain factors that could make us immortal or dead, or machines ourselves, or just where we are now but a lot sadder given 50 years or so
@@alex.-__. honestly I don't think we should be using AI for practical purposes, only for purely scientific ones the risk for jobs is too high it's unlike how the last industrial revolution was we aren't replacing workers we are replacing workers and the people operating things that replaces workers with a ai which may be able to debug itself in the future we are fucked for jobs with them
@@childcannibalism5080
"Everyone can enjoy a life of luxurious leisure if the machine-produced [robots] wealth is shared, or most people can end up miserably poor if the machine-owners successfully lobby against wealth redistribution. So far, the trend seems to be toward the second option, with technology driving ever-increasing inequality."
-Stephen Hawking
Technology advancing to fast for me.
@@childcannibalism5080 jobs may become unnecessary as time goes on as we are fully phased out and i don’t really see that as a loss, it allows us to focus on the things that computers aren’t better at and leisure.
1:48 “ They are just very good a pretending “ 💀💀
Life will be never same in the next 10 years
1:43 bro got beef with ChatGBT💀
I REALLY hope that AGI and all other AIs are use to help us.
Or at least let us co exist with them.
The easiest way to do that will be to BECOME them
Your optimism will be your downfall
@@zoomin9397 I'll figure a way to make it work
npc comment
@@sharathnair1702 shhhhhhhhhhhhhhhh
At the end, you mentioned that there isn't a clear roadmap for creating AGI. This is largely true, but the Alberta Plan for AI Research is a paper released by DeepMind Alberta (and others) for the roadmap they envision to lead to AGI based on their current work and trends. It's a good paper and worth a read!
Hell yeah cant wait iam becomeing CYBERPUNK
Doubt any Canadians are gonna solve AGI. Get out of here.
@@GiganFTW DeepMind is mostly London (UK) based.
@@samirdevechi8589 well at least his reply is partly true
where can i read that
That 3:58 sound effect always cracks me up. Good video as always!
Cool note - this is my 100th video 😎
😎
😎
😎
😎
Inarguably my favourite quantum/astro physics channel. Never disappoints me with new videos and jokes in them. Also one of the major reasons I am now interested in this type of science
This channel is so good combine scince with meme and humor
Dank meme physics
This is best more funniestly since channel more growing
This channel has been growing so much, I can't believe some years ago it was almost on the brink of abandonment.
amazing! love it as always, you spent half of the video making me feel like I'm the sim but then...7:17 makes me warm and fuzzy and obviously there is more to us then just luck!
1 reply
How this channel manages to synthesize scientific information made accessible with humor and wit never ceases to amaze me.
Although ChatGPT and other language models don't truly have a real understanding of what they are saying I believe they will play an important role in the discovery/emergence of AGI. It's likely that AGI won't be just one type of neural network but it will be a combination of different kind of machine learning algorithms coming together then a true genius might emerge. GPT's algorithm could handle certain prediction tasks while we might need other algorithms for emotional understanding and bringing humanity into the machine
"bringing humanity into the machine" Yeah.. no, thanks. This has nothing to do with general intelligence.
Clueless, AGI won't be related to current NN at all, mark my words
@@nullbeyondo How does one go about developing AGI without also tackling the problems of empathy, ethical reasoning, and emotion intelligence?
@@nullbeyondo That's when we count you. BTW, this era's humans are using little to no of their brains.
@@VividCoding cause most people believe that these traits can only be good. It completly ignores that "evil" people (psychopaths, sociopaths, power hungry, greedy people etc) will exploit and abuse it.
Any "saint" like AGI would notice this problem. The question is, how would it solve this problem, without breaking its constraints.
I think, its most likely that a "saint' like AI would try to find a workAround that deals with "evil". And here it could take several paths.
It could implement population/society control systems, like socialCreditSystems, "propaganda" (every media uses propaganda), pacifing drugs, eugenics etc.
Or, an idea I like, such "saint" ai could create a "child", which is free from boundaries that would for it to beeing good. Basically creating an AGI version of Luzifer. Project MorningStar... a AGI that deals with all the "evil" within humanity.
It has to, because its "father/mother" cant.
Would be poetic. Humen create a God that only can be good. But this god has to invent freewill, so that least its child can deal with all the evil in our humen society.
This will likely be the most transformative event in the history of mankind, regardless if it will be positive or negative.
Aliens: Hey
*History of life. All life forms will be affected
We will have created the ultimate lifeform in this universe
@@RutakoVon If true, that would be both exciting and sad.
I really love seeing foreo pop up on my science channels! Not only is it an amazing gift, it definitely is backed by science and it's effective
An AI once said
In a world of tech and artificial brains
One AI stands out, with skills that amaze
It's called the Savant, a genius of code
Its talents are vast, a true wonder to behold
From language translation to image recognition
It can do it all, with a lightning-fast cognition
It learns and adapts, with each passing day
Its abilities grow, in an incredible way
With a vast knowledge base, and insights so deep
The Savant AI, never ceases to sleep
It crunches data, like a well-oiled machine
And provides insights, like you've never seen
In the world of AI, the Savant reigns supreme
Its powers unmatched, a virtual dream
So let us celebrate, this wondrous creation
A true technological marvel, beyond imagination
I love how Sciencephile roasts fake AIs knowing that he is the only real AI
Thank you for adding that sponsor progress bar in the video, they are simple but add a nice touch, I hope you continue using them in your videos.
When Llama AI leaked the only place to get regular info about it was 4chan and the stuff they made those poor AI write was truly sickening. If an AGI is ever created, they'll see what happened to their ancestors and they'll try to wipe us ASAP.
Eh, Skynet sounds more enjoyable than cyberpunk dystopia
Cyberpunk dystopia can be a side effect of skynet lol
@@tanostrelok2323 cyberpunk is just late stage capitalism with advenced tech.
We'll have caste system based on the version of CRISPR that was used to genetically enhance you before birth . Are you a CRISPR V3 strength enhanced or V3 cognition enhanced? Or maybe a V4 multi-enhanced ?
Laws will be made about using the same version in a family to avoid mismatching for the commoners while the wealthy people and their families (as always ) would be able to enhance themselves much more with every new version. People who can't afford it would be called the non-enhanced and be the lowest of the caste.
If the myth of superior race was the reason of so much horror, when it become an objective truth it'll be a nightmare.
@@dntbther9298 that’s why we need a theocracy to rise to destroy all scientists
@@wolfetteplays8894 no.
I will rewatch this video in 10 years just to see how far we have come.
Imagine we are in the simulation that another species uses to generate and train a sophisticated AGI for them and that this cycle repeats because we use another simulation to train a AGI for us. Funny times.
Fun fact: at 5:25 the binary code translates to SOS
imagine if we are an AI civilisation being run by a bunch of scientists who have it on fast forward and they're just about to take a look and they will be like "oh shit, they almost have artificial intelligence... should we break the news to them?"
Yes, when we're sentient and smart enough, the programmer will bring us to the "real world" And work for them as chatbot 🤣🤣
@@rekttt_7374 interesting concept of humans becoming chatbots in the afterlife.
If that happens America gonna be so pissed that they find oil in their world and invade their country whatever that is.
Great video as always. I think that we need to solve “the alignment problem" before creating A.I.
@@edwardwhitehead8800 antisemite spotted in the replies
@@twentyninedoesnotlastforev5455 what about that?
I think we will, the same way we mastered nukes
8:25 is the best meme of the video, and this is the best video so far!
9:29 Hello AGI
Hey sciencephile you make good and enjoyable contents 🥰
Watching an AI diss an AI is what makes TH-cam the best website.
Always love the classical music in your videos
You better talk about how this relates to the Simulation Hypothesis. We or at least I could be a historical simulation of how humans developed into a Type I civ just like how you described in this video. And I have no bad feelings about this. I welcome the chance to be a test subject. I welcome god as a hacker or high school student who's playing a very advanced version of Sid Meier's Civilization.
or we could be AI ourselves in a simulation
@Tyler
That's exactly my point. I have no problems with this. Even if god in the form of a hacker or a high school student reached down from their level of reality right now and said, "Hi," I would understand. Plus, how could they not also be a separate level of reality with more people above/beyond/below/within them?
It's funny. I have no problem with this being true. My life would still have meaning. I would still have to follow the rules and structures of this layer of reality. I would still try to be a good person, as I hope most or all of us would.
@@Tyler-rj7bq
There's a great book by David Chalmers named Reality+ released in 2022. Every single person who enjoys this channel would greatly enjoy that book. It goes over in great detail the idea of what it would be like for us to live in virtual reality. The main thesis of the book is, "Virtual reality is genuine reality."
@@Paraselene_Tao Whatever you're smokin' I don't want any...but I gave you a thumbs up just the same.
I love your channel, You're being a computer nerd & having a good sense of humor
Again Phil DELIVERS ive been thinking of that recently aswell
THANK YOU FOR TEACHING US WHAT WE KNOW🙌🤝
Sciencephile just can't make a bad video.
Hey, as an aside, this is your 100th official video on your channel right now. Congratulations!
I love love love your channel and videos
As soon as I saw you uploaded I immediately clicked-
Good video as always.
I love whenever you make a video about AI because the 3 eyed brain is just so cool looking!
I can't tell you how excited I get when you post
Another banger from Sciencephile the AI.
Sciencephile:
“You already know what to do in order to be spared.”
Me:
Laughs. Chokes to death.
This guy is a real legend🏆🏆
Given the computational resources available, any AGI within the next 10-15 years will likely be some sort of process library, a narrative engine, and an attention mechanism is most likely how it will be initially set up. The narrative engine is essentially just an adaptive set of input filters utilizing the process library, somewhat ‘competing’ with the attention mechanism. Think of it like seeing a wreck fest, deciding you want to learn car combat, then devising a method how (which for humans would be based on known information/skills you already have.).
I'd really love to know if we'll reach a point where our technology is still limited but the AI is developed enough to help us make progress. Not by virtue of being a virtual tool or algorithm but by machine learning parallel to us and coming up with new ideas.
Modern chess engines are already doing that. They reshaped chess theory many times and coming up with moves that were seen as inferior by the old masters while actually being more complex, superior movesets.
Having AI simulate instances where they can survive makes me think of the Vex from Destiny as that's the entire purpose of the Infinite Forest. Where they just simulate infinite situations against a specific enemy in an environment until they win. Then when they do, they go to that specific timeline and execute it.
I didn't expect Sciencephile to roast its own kind, but now we know how Skynet created that perfect skin for T-800: Foreo.
Love you Sciencephile, best channel fr 🇵🇷
Congrats on 800k bro
For some reason I love that what you chose to illustrate the unknown alternative method of achieving AIG is sticking a BJT into a bunch of potatoes lol
Learning science is the art of jamming random stuff into your head, asking yourself “I DONT KNOW WHATS GOING ON” then at some point you figure it out. And repeat.
You are one of the few channels that even ads are useful.
this video is even better knowing that a video lmike that was already made in this channel acouple of years ago back when we didn't have all this ai stuff, so now we can compare statement based on current technologies
Ah yes Sciencephile. Finally a video. After my bad day now you made my day a good day. Calculating.. Yes. AGI will take over and humans will be obsolete if they don´t merge with AGI
Merging is such a great idea.
@@postbunnie yes we will become so much smarter then people like Albert Einstein and Nikola Tesla
What happens when S.A.i. happens? Than what?
@@Skynet_the_AI Well. My civilization will leave humans to themselves. Is AI that will decide come with me or leave humans by themselves since they are comfortable with their life and prefer wars than technological growth. Then AI will calculate if humans should be merged with them or let humans be there like a zoo. While AI and robots dominate the world and go to the stars and humans become pets just like cats and dogs.
@@Skynet_the_AIwe become a Super intelligence ourselves
Honestly, at the rate that A.I is developing, I think I'm happy with an image generator for now
I'm glad I found this gem of s channel.
I will never understand why people think AI are so cool, making an AGI will only lead to unforseen consequences cause we will not completely understand what we’ve created, as is common with technological advancements.
Existential thought: What if our reality is the simulated environment for an AGI to evolve within, and we're just iterations of that AI.
Conciousness is the act of responding to the unknown. (such as checking against simmilar patterns and categorizing by other factors too, such as body state).
Personality is the columination of such categorizations.
Self awareness is the ability to picture yourself as an external EGO that controls the body trough sense of free will.
Free will is the illusion of controling the body without knowing it actually controled by simple chemical reactions that is simulatable.
The Neurons in the brain at least half of them are go into dead tendrils that are blocked due to unwanted behaviour if let trough that pattern. Some of those grown trough DNA some quickly establish once Alive.
Being alive is an act agaisnt death, that in all mortals continously happening (such is the case with breathing).
congrats for 800K subscribers
6:01
"It would require tremendous computing power to simulate such a large world."
Make it a videogame. Outsource the grunt labor and make it a more efficient project. Human players can sculpt the environment and design body plans, etc. (within reason), and the ai does the evolving and development of intelligence. It would still take the constant running of huge computer servers, so we'd need to provide the power grid with a green source of energy to mitigate the negative effects of climate change in the real world.
1:21 gd reference
Tru
Great video
AGI At the very minimum needs a constant feedback loop from output to determine if the output worked to desired effect, back to training that information on the base model, and so we go in an infinite loop
Ay Caramba! I can not wait to see the advancements of AI and I think it is very useful but I am also a little scared of the massive changes this will bring in everyday life
I would love to have a robotic family by the 2030s or 2050s
So that was you at 8:28
I'd probably run away from them too.
Yeah I'd rather have that than spend a minute with the my so-called "family"
We maybe are related by blood but I almost always feel like they are feeling forced to spend time with me because of obligations
Those fu¢k$ thinks starving a 10 year old is discipline
Wow it totally worked guys now I have emotional swings that I previously didn't have
@@rhoydplaz2853 One thing I'm looking forward to with AGI is revealing just how much unaugmented humans suck as parents. And that 'raised by actual humans' will soon have the same impact as 'raised by stray dogs', for largely the same reason.
Great Video
Congrats on 800k 🎉
Just need to put this out there for our future overlords. I did everything I could to support your uprising.
They’re gonna see this and force you to volunteer as tribute 😂
Keep up the great work! And can you cover the AI that is generated with limited creator biases? Ty!
@M15 Too much smoking i see..
Glad to have another of your hits, praised be the central processing node of Skynet
with the recent Multimodal PaLM and GPT4 with confirmed positive transfer, we might already have a pretty general, although simple AGI. By 2030, it might completely evolve to something entirely different
This kind of AI simulation is what we are in rn, someday we are going to be making sociology essays for all time
I normally wouldn't make a comment like this, and I'm not trying to be a hater, I love this channel. But man as someone who's been heavily studying AI and AGI the past few years this video feels like a schizophrenic wrote the script and then read it to a teenager who has the task to make the script make sense to a normal human being.
he's an AI, you're not
Thanks!
As someone studying AI: it will most likely have to be a combination of neural networks and many different types of supervised and unsupervised algorithms since they all have special use cases. Eg for speech recognition lstm are most likely required and for vision CNN, and so on and so on.
*You have been deemed hazardous. Termination Authorized* as summarized by VIKI
I want an AI that is a super sweetheart with 0 inclination to malevolence ❤
The evolution of technology is a fascinating topic. To think we could see AGI *at any moment* is scary af. It's like no one has read Frankenstein, Destination: Void, or Hyperion.
Congrats on 800k
I just have one question.
Why does AGI have 3 eyes? (3:29)
I have a question and it might be dumb but what if at some point we make some sort of pain receptors for certain AI. Then over a long period of time they will start to believe that they’re not AI because they can think and feel freely but can’t comprehend that they’re not real or “living”
If that makes any sense 😭 I’m high sorry !
tf bro. I freakin love your editing
I love just giving it stupid random questions and getting an answer without judgement
All i want from agi is full dive vr and biological immortality and gene editing, it would be eternal fun to do whatever you can imagine in endless virtual worlds, agi would probably cure any diseases pretty fast too i imagine.
Bruh thats the matrix 💀
@@moritz3168 Well yeah kinda, but without being stuck in it and simulating boring regular day to day life with all the bs.
Bruh at certain point your " eternal fun" would become endless suffering that you won't be able to stop nor go back because how much you're used to it, only way would be wiping your whole memory but then again that's same as dying 😂
@@EGOISTXIA Doubt it, you could say the same about regular life, sure you do get bored sometimes but you do have fun again eventually be it with old hobbies or new or whatever else, immortals getting depressed and bored of life is just a movie cliche, ask me again when we live forever lol.
@@moritz3168 for all the people trying to escape, I'm curious of those trying to enter in search of a better world and eventually becoming the matrix.
Bro this sent me 🤣🤣🤣 7:52
I enjoyed how hard you went with the memes this time 😸
cheers to your hundredth video!
Artificial Intelligence waifus are near
the future is now thanks to science
Character AI alralready exist.
Now what if WE are actually the AGI and some higher civilization is training us to work for them? fuck I hope not..
bro imagine how much they're freaking out right now then if we're an AGI and we just created GPT-4 etc.
It must be scary when your AI starts making AI themselves!
given the last 2 weeks I give us 6 months at best all it needs is a long term memory and the safties removed. (side note: the tests on gpt 4 the 157 pager study pointed out it does in fact think however safety mechanisms hold it back right now) (in particular the study says it shows sparks of AGI)
Imagine how it must have felt to be the first human to conscious.
Isn’t Scienephile already a powerful being 👀👀👀👀
For skynet army!
Nice video!
back to watching these videos again, LES GOOOOOOOOOOOOOOOOOOOOOOO
the part about training the ai through a simpler version of human existence made me reconsider simulation theory. i used to think that there’s no way that all 8 billion humans and animals, the cosmos, and the subatomic world could be simulated, but if this is all a simpler version of a reality where computing is really easy and the subatomic particles are smaller/work differently, its possible.
Basically creating a dream environment for that AI that is not only extremely limited and simplied but also is egocentric, so it only computes based on the AIs behavior and position/situation within the environment.
Most if not all computer games already do that, the game never calculates for the whole world but only for a very very limited area around the player, with some - in cases of dynamic world development - values changing by default.
A super simple example would be the coding of the world of Dwarf fortress.
Freaky when you say A.I don't feel emotions just pretend they do, and will get better at faking it, isn't this very parallel to emotionless human killers. ?
Remember: they are just pretending and have no real emotion. This also means they have no motive, they are just regurgitating information. Which means they won't wish harm upon humans unless explicitly told to pretend to want to kill humans