Ian LeCun, an AI researcher at Meta, believes that large language models (LLMs) are useful but limited. He believes that this direction will not be able to be developed to the level of artificial general intelligence (AGI).
Ian LeCun, an AI researcher at Meta, believes that large language models (LLMs) are useful but limited. He believes that this direction will not be able to be developed to the level of artificial general intelligence (AGI).
The world’s largest companies are spending billions on artificial intelligence, including large language models like ChatGPT, Google’s Gemini, and Meta’s Llama, Business Insider reports. LeCun, who has decided to leave Meta, according to insiders, believes that’s a mistake.
«LLMs are great, they’re useful, we should invest in them — a lot of people will use them. They’re not the path to human-level intelligence. They’re just not. Right now they suck the air out of the room wherever they go — and so there’s essentially no resources left for anything else. So for the next revolution we need to take a step back and figure out what’s missing from the current approaches,» said Yann LeCun, thus criticizing his own employer.
LeCun has been a consistent critic of LLM for years, arguing that true AI will not come from language models that scrape text from the internet. He believes that the breakthrough in AI will come from «world models» that rely instead on visual data.
Previously,Yann Lecoun outlined four key characteristics of intelligent behavior that even large language models have failed to achieve. According to Lecoun, these are: an understanding of the physical world, the presence of a stable memory, the ability to reason, and the ability to plan, and «to plan complex actions, especially hierarchically.»
Reports emerged last week that LeCun would likely leave Meta and launch his own artificial intelligence startup.
