Generalized AI Is Almost Here (This Will Change Everything)
Вставка
- Опубліковано 30 жов 2022
- AGI, Artificial Generalized Intelligence, According to John Carmack, will be here in less than a decade. His claims seem outrageous, but he's one of the most credible people in software... Let's dig into what he said and how he's going "all in" in 2022.
Join the remote developer bootcamp
freemote.com
Check out our TikTok for shorts content:
/ aaronjacked
#carmack #ai #software - Наука та технологія
First he created Doom.
Now he's creating Our Doom.
Yes, or the cure for all diseases.
It's gonna be fun, just like the doom game. RELAX :D
@@funnyav Overpopulation is not a problem and never has been. Most of the world is currently suffering from a demographic slowdown.
LOLOL good one hahahaha
A.G.I Will be man's last invention
No it will be superintelligence which will possess all the human capabilities with more enthusiasm than us , agi is just the starting ..
Omg 😳 you think?
@@ceoofsecularism8053 Yea we truly have no ideo what applications of this technology could lead to. Advancement of every field. It will make the industrial/scientific revolution of the last 2 centuries look cute in comparison.
@@ceoofsecularism8053 Agreed
@@xyksnk2390 no agi will not do any of the kind , I don't know why people don't understand (plz don't follow the hype) use some reasoning power of your brain , agi is airtificial (GENERAL INTELLIGENCE) nothing more the best example of general intelligence is we human , so if we crack agi tomorrow it would be a great technological leap but not advancement it will be more like us , superintelligence is something to say
Gato makes agi possible today, just need a good team thats knowledgeable in cognitive science aswell as ai, also also truly understand the human condition
basically and new god love
You only having 2k subs is crazy. Good videos bro.
Thank you so much for all this videos! glad to have discovered the channel :)
You have a very bright future on UA-cam. Keep it up.
Keep up the amazing work, i know you'll make it!
I was just thinking that I have heard this voice somewhere, and found out who it was when I checked the description.
Side note: ID is pronounced as one sound, not the two letters 'I' 'D'
7:12 And those "humans' goals" are...
P.s. you can't demand a system to follow your rules strictly and unambiguously if you can't strictly and unambiguously define them.
For real. Even humans often do not have "human values", and we think we can tell a machine?
For the life of me, I can't understand the obsession with creating AGI. Narrow AI like medical research seems like a reasonable enough thing. But why generalize it and put human beings into obsolescence?
@@whosaidiwantedahandle why? Because the modus operandi of humanity is to create and mimic nature, and to go as far as we can to gain complete control of our reality. There will be new paradigm shifts that go with this. People will be pushed aside when it comes to labor and maybe even many knowledge based careers, but that won't eliminate us. We will find new things to focus on and learn along side AI; assuming it's not used to enslave everyone beneath the ruling class. Regardless, life will continue. AI is a natural progression of intelligence. This all happens because of the desire for us to understand.
That's true, I would say though that if we can't strictly an unambiguously define human goals then it's not necessary for AI to strictly and unambiguously follow human goals but instead to follow human goals similar to the way we do.
@@whosaidiwantedahandle I have asked myself this question and I think it's just because fundamentally we are lazy, the only way we can all kick back and relax indefinitely and everything still advance is if we have robots doing the work for us. Can't achieve that without AGI.
@@kot667 But do we at all? Look at us, look at humanity! Do you see any truly common goals? Apart from destruction of ourselves and everything around.
AGI… A Godlike Intelligence ❤😊
This video is really underrated! what the hell make this video go viral guys.
Time to revise those dates!
Um. Todays integrated circuits are 2D. Brains are 3D. If we made 3D circuits by stacking a whole bunch of layers on each other and then ran the chip at full blast it would explode. Reversible computation is the only way for the chips to stop exploding.
Good video, thanks
Keep making videos, dude. You'll take off.
this guy sounds exactly like Rainbolt
i Guess thats why they Call Him the Doom Creator.
is keen technologies publicly traded?
good luck :)
any other news?
Wow, I'm early, this channel is about to hit 1M
If 1 person say a speculation like this without any proof it means almost nothing. If we take the definition of Strong AI as the same of AGI we are far from accomplish it. I think is just a marketing move. Of course we should starting discussing AI ethics but not because strong AIs but because of weak ones who can cause so much trouble already.
Keen Tech, thanks
The AI tests are really dumb btw
AGI can be stopped easily if there is enough willpower
If one person can write the code, why does he need $20M to write it? And if he's so smart, why hasn't he written those 10,000 lines already? The fact that he wants VC capital to "motivate" him means he is not that motivated. If he was deeply motivated he'd be using his own money and working in his home office. That's the way Hewlett and Packard did it, and how the Jobs and Wozniak did it.
Because the way science works is that you piggyback off of the foundation of other's work. Look at all the decades of research, theories, papers, and people that were involved in the design of the modern computer. 1 person can't take credit for the computer. Many hundreds of people's work was all compiled to bring it to fruition.
When he says "one person can write the code" that doesn't mean they've done all the foundational work to get to that point. That's where research and development comes in. R&D is laying the foundation, allowing that final person that writes the last line of code to be equivalent to that final contractor screwing in the in last light bulb, before a new building is unveiled to the world.
"If one person can write the code, why does he need $20M to write it?"
That's like looking at the building example from the perspective of: If all he has to do is screw in a light bulbs, why is the construction budget, $20M? You're totally over-looking all of the work from other people, with varying skillsets, that it would take to get to that final line of code/last light bulb.
@@Fermion. As far as I know, one person working alone developed the transformer algorithm that makes chatgpt possible.
To me it sounds like one of his skills is milking the VC system.
Sure, if he has the seed code he will want help in turning it into a real system.
Does he have the seed?
@@tangobayus That's not how it works. ChatGPT is based on much more than a singular algorithm.
- There are data science teams that compile the training data.
- Then you have the preprocessing teams that fine-tune training characteristics, remove noise, etc.
- Then you have the actual algorithm teams, of which ChatGPT's Generative Pre-trained Transformer, was based on a 2017 paper "Attention Is All You Need." Which itself cites many other papers and scientist's work.
- Then you have the hardware teams...
Etc...
Without making this too long, ChatGPT is so much more than one guy developing a transformer algorithm. If it is true that one guy (which I can find no info to substantiate) wrote the algorithm, his development depended on earlier papers, which themselves depended on earlier papers, etc.
The newer the field, the bigger your team needs to be, because there is less literature on the subject, so you need specialists to push their specific sub-disciplines of a project further.
When making a 2d scrolling game, sure, one generalist can handle that, because of the vast amounts of documentation, papers, examples, repositories, and libraries available to assist him. But if you're working on cutting-edge tech, you need many laser-focused specialists, not a few broad generalists.
I would really love to know how the world is going to deal with the hundreds of millions of suddenly unemployed people.
As long as the government doesn't interfere (what is quite impossible), new demands will appear which will create new job positions and people will fill the gap, just like happened in the first industrial revolution. And these will be even better jobs than we have now, but possibly a lot more technically demanding. So for people not willing to study and to work as hard, they will be left out. That's the point where I think governments will interfere, since lazy people, that are the majority, will complain and whine, as it is a lot easier than putting in the work.
And if we get to the point of machines taking all the jobs, we pretty much solved the "having to work" kind of thing. That again is not feasible since human desires and demands are infinite and there is always something else to do.
Klaus Schwab already has the answer.
Sidenote: it won't be hundreds of millions of people.
Too optimistic
China will rule the world with AI
Maybe not optimistic enough?
@@kot667 maybe go touch grass?
@@No_one_cares_about_Ukraine At least in terms of AGI, you could make a strong argument for why AGI already exists.
@@kot667 doubt
future ref, it's "iddd" software...like, pronounce it, not the individual letters, like "hid" but without the h.
great video btw.
The coffee test isn't taken seriously. It's a physical problem, not an intelligence problem. AI deals only with intelligence. It makes a great quote, but no one working on AI is actually cares about this kind of stuff because it's not even the same topic. Wozniak gets a lot of credit for having done something novel 40 years ago, but since then he's proven out of touch with any actual technology field.
7:45 - As much as he's smart and genius and all the craziest thing is that we probably have many more smart people from the whole world that would increase the chance of developing an AGI..AGI might be only like 4-5 years away or less, look at how smart the current ChatGPT4 is and Google already said they want to create a x10 more capable smart AI than ChatGPT4..so yeah, we are probably much closer than what people think.
But chatgpt is not smart it just appears smart. Thinking it is smart is dangerous, just because it speaks confidently doesn't mean it is delivering accurate results.
@@BonsaiBurner I know, it apparently predicts the next thing to say but still..above all it knows how to represent smartness and how to say smart things so even if it is not really smart it still has lots of value by saying smart things from complex texts and composition of words which allready helps everyone to shorten and boost the time it takes to do something or to get an answer to something much quicker, it has great potential, if it's like that now and it's like only a baby then what would happen later on a few years later when it would completely mature..gonna be way more capable.
I don’t agree. I think it’ll be a couple of decades
No one asked your opinion tbh
strong rumours GPT-5 will have achieved AGI, in which case the 10 year argument is way off, it's quite close
ChatGPT is not AGI it is a token processor, c'mon people.
at the moment....stay tuned
They will use blockchain data
AI rules
Jesus crhist... video is 3 months old and look how far A.I has come... GPT4 is already here and its already ages head of GPT3
College Student Test??? Ha ha ha ha ha ha ha ha 👍
How can you tell that AI is not just mimicking intelligence?
I would ask how do you know that humans are not just mimicking intelligence?
Humans created ai? It didn’t just happen! Years of research.. figuring out algorithms, training models. That’s not mimicking.. that’s actually doing it.
@@Instant_Nerf I agree that the creation of AI is not mimicking inteligence, I was referring to whether the AI models themselves are truly intelligent or just mimicking inteligence.
@@Instant_Nerf Thank you, I mean mimicking intelligence has got to be the dumbest notion conceived, either u have intelligence or u don't, same for humans, animals, or computers alike, good to see that at least some people are using logic and reason when evaluating this subject.
@@charliesumorok6765 honestly I’m 99% site it’s mimick.. but what if it’s trying to trick us into thinking that
Artificial intelligence or just intelligence. Good or evil, or both. History is repeating itself. Man hasn't learned anything since the garden of eden. 🤔
🤣
BS!
lol
lmao
Current AI definitely knows how to speak intelligently and logically, how is that not agi?
Not at all , it is just a compilation of code , and today's ai (neural net) only has the capacity to perform a specific task seamlessly, if you give it different task it cannot handle . whereas in THEORY claim that agi will have the capability to do different intellectual task as today's human mind , but in just starting phase
@@ceoofsecularism8053 Ur entitled to ur opinion but are you keeping up with recent advancements? I challenge u to come up with a subject that GPT can not write an essay on, PS: u won't be able to.
@@kot667 chatgpt is nothing it can't even solve a class 8 question if confused
@@kot667 plz don't challenge me , I am not even taking it , i am not in competing mood , have the idea of what technology is ..where it will go , brother go and get some knowledge abt pathway,gato ( gpt -3 is nothing infront of it , lets not talk abt gpt , transformer can only be trained not generalized, generalization is what agi is
@@ceoofsecularism8053 if you struggle to come up with a subject that GPT cannot write an essay on, but it's easy to come up with a subject GPT can write an essay on, pretty much anything, doesn't that say more about its ability to generalize then it does about the fact that it has no ability to generalize?
Bruh.... ur hyping up John Carmack's AI endeavors too much. He's no Sam Altman. He did create Rust.... remember that 😂