Comments by "Joe Qi" (@i6power30) on "How Nvidia Grew From Gaming To A.I. Giant, Now Powering ChatGPT" video.

@jaylu7021 it can appear to be self aware that's all it matters.
4
I think NVDA and traditional ai has reached peak valuation.. Neurnets can only go so far. The next big thing is AGI that are capable of self aware and reasoning. AGIX could be the next 100-1000x
2
@potsmoker54 how do you know anyone you interact with is truly self aware? What if you live in a simulation? And everyone you know is actually AI fooling you to believe they are real living humans?
1
@sa1t938 Hopefully not. The current form of neural nets is too stupid. It requires enormous amount of data to statistically brute force learn all possible corner cases. Normal human brain can learn much more efficiently with much fewer examples, and extrapolate. Just because currently, neural nets are the dominant algorithm in AI, doesn't mean it will continue to be. A totally different approach may come up to be more superior, and more flexible than brute force big data approach, the one that require less intensive computation, but more clever algorithm.
1
@sa1t938 yeah I agree that better architecture is needed to advance AI not just throwing more computing power at it as Nvidia wants to sell more of their hardware. It's just going to create more and more specialized AI systems but no where close to AGI that is capable of general reasoning like a human. It's still a mystery how human brain is capable of learning and extrapolating so much more efficiently with just a tiny fraction of data as the current most advanced AI algorithms need. I would say that the current neural nets are just crude approximation of human brains. A lot more theoretical advance is needed to achieve AGI, it's not simply a matter of more compute hardware.
1