Everyone saying this is different than the dot com bubble: Yes, I know LLMs aren’t 100 percent hot air and there are legitimate use cases for them, but the internet also wasn’t useless and it didn’t go away. It was just overhyped. It’s still an apt comparison
It is, but I think we are a little bit past the point of the froth of that time. The mania is simpler, but the amount of capital going at it just has to be a lot bigger.
Most of the capital is going to the big 7 though, and they spend a lot of their own money on it as well. To me they are just chasing this pipe dream of AGI, which they are nowhere near to because their approach is fundamentally flawed. I genuinely believe AGI is possible, but not in the way they are going about it. That is actually a very good thing, bubbly as it is at the moment, because in their race towards AGI, which would be a true game changer, they are creating a black box with all the safety measures removed and even those companies don’t know what the hell it does or how the hell it does it exactly.
The only person who I think might manage to build AGI in our lifetime is Song-Chun Zhu, because he is the only person who has an approach that makes sense.
Basically he states the big data approach is fundamentally flawed because the LLM’s cannot reason at all. Everything it generates is probabilistic. So the systems and robots he builds are built to be “aware” of their surroundings and context, not relying on a dataset, but on ad-hoc input. An example cited in the article is the Tongtong 2.0 robot which uses a pillow to give it some extra height so that it can retrieve a book from a bookshelf otherwise out of reach. It sees the book, realizes it can’t reach it and then finds a solution so that it can.
Everyone saying this is different than the dot com bubble: Yes, I know LLMs aren’t 100 percent hot air and there are legitimate use cases for them, but the internet also wasn’t useless and it didn’t go away. It was just overhyped. It’s still an apt comparison
Also need to add that there is so much in the field of AI and machine learning that AREN’T LLMs
It is, but I think we are a little bit past the point of the froth of that time. The mania is simpler, but the amount of capital going at it just has to be a lot bigger.
Most of the capital is going to the big 7 though, and they spend a lot of their own money on it as well. To me they are just chasing this pipe dream of AGI, which they are nowhere near to because their approach is fundamentally flawed. I genuinely believe AGI is possible, but not in the way they are going about it. That is actually a very good thing, bubbly as it is at the moment, because in their race towards AGI, which would be a true game changer, they are creating a black box with all the safety measures removed and even those companies don’t know what the hell it does or how the hell it does it exactly.
The only person who I think might manage to build AGI in our lifetime is Song-Chun Zhu, because he is the only person who has an approach that makes sense.
What is Song-Chun Zhus approach? I couldn’t really find an answer to that with a quick search.
This article doesn’t go into great detail about his work (it’s more about his life) but does touch on the notions of it: https://www.theguardian.com/news/ng-interactive/2025/sep/16/song-chun-zhu-why-one-of-the-worlds-most-brilliant-ai-scientists-left-the-us-for-china
Basically he states the big data approach is fundamentally flawed because the LLM’s cannot reason at all. Everything it generates is probabilistic. So the systems and robots he builds are built to be “aware” of their surroundings and context, not relying on a dataset, but on ad-hoc input. An example cited in the article is the Tongtong 2.0 robot which uses a pillow to give it some extra height so that it can retrieve a book from a bookshelf otherwise out of reach. It sees the book, realizes it can’t reach it and then finds a solution so that it can.
Lol wut? LLMs are a joke. The Internet is useful.