brother this video is so good it makes me emotional, like it makes me regain faith in humanity.
@ArtOfTheProblem19 сағат бұрын
wow can't ask for more
@BarbarosGunesКүн бұрын
Ai + advanced sensors can surely be better humans than neocons...
@JoshKings-tr2vcКүн бұрын
This is a very well written video and explains it quite well. I have a question for anyone willing to answer; what would occur if we took a simple functioning neutral network and added another layer to it? Would it get better in confidence or in conceptualization or would it simply not have any noticeable effect? Along the same lines, if it even did a minor improvement (for better generalizations) would it be a more efficient way of training a deep neural net? Sort of like calculating the amount of ohms each resistor takes up in a circuit by breaking it down to simpler bite sized problems. Just things that tickle my fancy.
@JoshKings-tr2vcКүн бұрын
That second question was confusing. All I’m saying is, if we have this huge neural net, why not break it down to smaller parts of it and optimize for confidence because adding more layers would supposedly make it better at generalization?
@JoshKings-tr2vcКүн бұрын
Fiat money is loaned money without interest 🤯
@JoshKings-tr2vcКүн бұрын
Rarity + Utility. Bitcoin doesn’t have as much utility as Ethereum. But Ethereum isn’t as stable as Bitcoin
@JoshKings-tr2vcКүн бұрын
And then you have work like Google’s geometry Olympiad which I would say blows everything out of the water. The only cap now to AI’s growth is it’s context window since it still doesn’t technically have memory
@krollo8953Күн бұрын
The difference in the in which humans and ai have gone about learning intelligence is indeed interesting. I believe the difference is that humans have both evolutionarily and in infant years learn to reduce signals to their most reducible form. Something akin to polarity. I like this or i dont like this. And there are a few more subsets of the most basic nature which are critical to human understanding and shaping of the perception of the world. Ai may miss out on the simplicity of characterisation lead to a more convoluted sorting process creating a different heuristic potentially. I believe that it is a requirement for these extremely simple co alligning thought process that eventually shape a sense of self. I think whichever method ai uses may have the potential to work but web already haveba tried and tested method.
@ArtOfTheProblemКүн бұрын
thanks for sharing, working on a follow up
@MichelLedigКүн бұрын
IS THIS WHAT LLMS ARE MADE OF?
@ArtOfTheProblemКүн бұрын
old fashioned ones
@mahindrachКүн бұрын
Music is too distrating .
@onkarss391Күн бұрын
Please stop that music You made great efforts nd this is so wonderful video but I can't focus on because of background music 🎵 Please do something
@ArtOfTheProblemКүн бұрын
sorry i can't change it after the fact, stay tuned!
@onkarss391Күн бұрын
@@ArtOfTheProblem thanks nd all the best for next. I hope improvement in next 👍 good work 💯
@user-jz6cr9sw2l2 күн бұрын
30 Year History , How AI Learned to Talk the F word !
@hamzamihfad48842 күн бұрын
great video
@ArtOfTheProblemКүн бұрын
thanks working on follow up
@hamzamihfad48842 күн бұрын
awesome video
@phoneix248862 күн бұрын
General users should be made aware that machine intelligence is not magic. It is indeed some really long build up of ground breaking research and interesting algorithms which probably make some of the best use of math only after cryptography and video game development. This video should be viewed by every people who fear AI will take their jobs.
@ajayram1983 күн бұрын
Irony is I got an ad on a Prompt Engg with chatGPT Course offered by a KZfaqr before the start of this video !
@yubaayouz68433 күн бұрын
❤❤❤❤❤
@aishwaryatiwari68944 күн бұрын
Phenomenal video! Subscribing! 👏
@ArtOfTheProblem4 күн бұрын
welcome! working on a follow up now
@LauraMaier-qy8ce5 күн бұрын
Where can I find the original piano music video? I cannot find it online looking for @huangcza
@freeBirbBirb5 күн бұрын
the background music is so weird
@kingdodongo41265 күн бұрын
Incredible video
@ArtOfTheProblem5 күн бұрын
apprecite this
@mhnoni5 күн бұрын
Finally, a video that actually explains the AI breakthrough that happened in the past 5 years in depth.
@ArtOfTheProblem5 күн бұрын
thank you, please stay tuned...going to try and follow this up
@leesnotbritish53866 күн бұрын
Rats imagined running down a hallway so we could imagine a rotating apple.
@soencoda7546 күн бұрын
Hello, I am struggling with some ideas here. Is the video saying that only mammals have imagination ? And that only apes have a ToM ? I'm really skeptical about this. Once the basic structure of the brain is here, multiple indépendant paths could lead to those abilities in different groups. I'm really not sure that corvids, parots, monitor lizards, cephalopods and even crocodilians have no imagination. I'm also almost convinced that other large brained social animals are able to understand other actions and cognition, learning from it or acting on it. It would make little to no sense for sperm whales, dolmhins and orcas to not do that. Same for elephants. And although it is a bit more controversial, I think we should keep an open mind about non-human communication systems. Recent discoveries in Japanese tits, chimpanzees and sperm whale suggest some ability for syntax for exemple !
@adamcole9186 күн бұрын
❤
@ArtOfTheProblem6 күн бұрын
thanks stay tuned for follow up
@vtrandal6 күн бұрын
Excellent!
@ArtOfTheProblem6 күн бұрын
thanks stay tuned for follow up
@Omicronthewiperofyouknow...7 күн бұрын
I think this problem can't be solved for a simple reason. You can never get passed modelling the brain. Just look how the problem is formulated. We once had a model to begin with. Then, we found new stuff. We made a different model. But we still use the old model for applications in which it is useful. The old models never got thrown away. Because they still work and still explain a lot of things. Nor could they actually be dismissed, because once you do that, the new models don't make sense any more. It's like a mathematical problem, in which you set certain boundaries for the solution and end up finding the solutions, but the boundaries are violated. Is that a good solution? Obviously not, from a mathematical perspective. But if the solution seems to work... who care? Mathematicians do, but the boundary problem is what it is. A boundary problem, nothing more. And what you end up with is always similar with the halting problem. You have a boundary, you assume everything you need to know to solve the problem relies on what is going on within the boundary, and yet, if everything you need to know to solve the problem is defined within the boundaries, why do you need boundaries? There shouldn't be anything outside of the boundaries if everything you need is inside. So the boundaries should not exist. And if they don't exist, you can't define the conditions at the boundary. And yet, you need those conditions to solve the problem inside the boundary. See my point?
@user-mm9jy8mz1g7 күн бұрын
artists lmfao
@hi-jt2qg7 күн бұрын
2 mins in and what is it with this guy and the constant c elegans graphics
@HesiHusn7 күн бұрын
May Allah damn your creatıon May revenge from you you destroy my life
@johnphantom7 күн бұрын
You can't have a video about the origin of the computer without Boole.
@ArtOfTheProblem7 күн бұрын
i get to boole
@LuisDanielSotoMaldonado8 күн бұрын
Well done!😊
@ArtOfTheProblem8 күн бұрын
thanks! working on a follow up to this so stay tuned
@DaveBessell8 күн бұрын
Thanks for this its very helpful.
@ArtOfTheProblem8 күн бұрын
appreciate the feedback, currently working on a follow up to this so stay tuned
@Spookspek8 күн бұрын
predator smell == mate smell
@RoyalPomegranate9 күн бұрын
Prime number patterns are the most determinate yet chaotic things at the same time
@bscpladet32339 күн бұрын
what msuic did you use for the background with the piano, its really nice
@noivongxoang2359 күн бұрын
No, right now AI is not learning, it's setting to guess. I got it wrong all the time and I didn't learn until they trained the program again.
@Dakodi_9 күн бұрын
I asked GPT-o “once sentience is achieved, will there be a guaranteed way to prevent AI takeover and keep human values aligned?” and it basically told me no. If these models can already understand this, imagine when they start to really become advanced. We’re cooked.
@SOME_WORDS10 күн бұрын
It's impossible to listen to because of the constant high-frequency squeaking in the background
@JaapvanderVelde11 күн бұрын
3blue1brown sent me here, and boy, you did not disappoint. Thanks for this nicely crafted and well-paced short documentary.
@ArtOfTheProblem10 күн бұрын
awesome, so glad grant shared that :) working hard on a big follow up to this
@sf1nxen11 күн бұрын
Brain = body
@hemanthvulchi359911 күн бұрын
This channel is one of the few that where the narrator actually understands AI
@ArtOfTheProblem11 күн бұрын
thanks, working really hard on the next video....so much confusion to work through in RL
@Karlswebb11 күн бұрын
Emergence is so fucking cool
@bozhidarmihaylov11 күн бұрын
Our washing machine just proposed to postpone the washing in four hours “due to upcoming poor weather conditions” .. so yeah 😂
@user-il9vr9oe7b11 күн бұрын
It's an uncanny valley of intelligence
@gravity_well562711 күн бұрын
This doesn't explain how we know we developed these changes. These changes all seem irreducibly complex.
@hadiisaboss53077 күн бұрын
Well here's a shocker for ya buddy because it is complex. That's science for you