News Nvidia's business model 'may fall apart' if corporations can't make AI pay, says SK Group boss

[slightnitpick]

Who are you quoting?

I would point to self-driving cars. If AI can drive a car, then it doesn't seem like a reach to suggest that it would be able to power a humanoid robot. Same with drones/airplanes.

There's a reason why bulk freight prefers rail to roads for any real distance travel. AI might be able to power a humanoid robot, but an AI powered humanoid robot is going to be really inefficient compared to a programmed, obstacle-avoidant, non-humanoid robot in most use cases.

 

[ttquantia]

AI has been controlling robots for a few years now, and is already doing quite a bit of work for programmers. I'm not sure how much of it exists in airplanes, but they pretty much already run themselves.

AI is pretty good at replacing other forms of heavy computing and also at replacing humans. AI is as good as the best doctors at detecting some things.

While there are a lot of misconceptions, some of them are with people like you, who just don't realise how much AI is being used, or, as has been done repeatedly over the years, classify anything that already works as "not AI".

I know exactly how much AI is being used. The AI behind LLM and AI image generation is completely different from the AI in many other applications, including the AI used in robotics. AI was already "just about to" drive cars autonomously, but it did not happen, and the AI hype would have died if LLM hype had not started. AI is being used in lots of things, but there is no AI breakthrough happening now anywhere else except generating text with LLM and generating interesting images. No breakthroughs in robotics, self-driving or anywhere else. There is very slow gradual progress that has been happening over the last 30, 40 and 50 years, and the last couple of years LLM and "funny image" (and now also "funny video") AI fad is nothing but a small footnote in the progress of AI, and won't be changing much anything very much.

 

[ttquantia]

I would point to self-driving cars. If AI can drive a car, then it doesn't seem like a reach to suggest that it would be able to power a humanoid robot. Same with drones/airplanes.

That is the flawed thinking I was talking about. If a person can do X then that person can do Y does not mean that if a computer program can do X then a similar computer program can do Y. The mistake now is that LLM is "almost like human" because it can produce text that is often indistinguishable from texts produced by humans (and in some cases the text is actually identical, as LLMs are known to plagiarize), and this skill does not have much to do with other human skills. LLMs are not going to drive a car or control a robot. Ever.

 

[bit_user]

That is the flawed thinking I was talking about. If a person can do X then that person can do Y does not mean that if a computer program can do X then a similar computer program can do Y. The mistake now is that LLM is "almost like human" because it can produce text that is often indistinguishable from texts produced by humans (and in some cases the text is actually identical, as LLMs are known to plagiarize), and this skill does not have much to do with other human skills. LLMs are not going to drive a car or control a robot. Ever.

I never said anything about LLMs. I related self-driving cars to AI-controlled robots and drones/planes. I'm well aware that LLMs aren't involved in any of those applications.

BTW, Nvidia does support using generative AI for parts of the simulation infrastructure to train & test self-driving AI:

Advancing Autonomous Vehicle Simulation with Omniverse and Generative AI | GTC 24 2024 | NVIDIA On-Demand

Learn how NVIDIA Omniverse brings together the AV ecosystem to address the challenges of achieving scalable, high-fidelity simulation

www.nvidia.com

 

[slightnitpick]

there is no AI breakthrough happening now anywhere else except generating text with LLM and generating interesting images. No breakthroughs in robotics, self-driving or anywhere else.

Some nice things are going on in protein design.