The next AI revolution is coming. But it's not AGI and not coming next year: Yann LeCun at AI Summit

AI pioneer Yann LeCun (L) at India AI Impact Summit

• Yann LeCun says there's no such thing as AGI; human-level AI is specialized
• Next AI revolution will use world models, not LLMs
• Silicon Valley focused only on LLMs, says LeCun

Did our AI summary help?

Yann LeCun, widely regarded as one of the godfathers of artificial intelligence (AI), believes there is another AI revolution coming, but it's not AGI (Artificial General Intelligence) and it's not arriving next year.

"There's no such thing as AGI. There is human-level AI, but human intelligence is extremely specialised. So calling it general intelligence is complete nonsense," LeCun said during a panel discussion at the ongoing India AI Impact Summit on February 20.

LeCun, who recently left Meta as chief scientist and started his own 'world model' venture AMI (Advanced Machine Intelligence) Labs, noted that while we will eventually build systems as intelligent as humans in all domains, progress will be slow, more difficult, and take much longer.

"It's not going to be next year, unlike what some companies in the industry are claiming," he said. "It's not like we're going to discover one secret that's going to just unlock intelligence."

LeCun argued that we should instead call it ASI (Artificial Superintelligence), systems that can surpass human capabilities.

Also Read: AI pioneer LeCun endorses India's AI leadership role, says 'Modi has India moving in right direction'

Next AI revolution will be led by world models

He explained that the next AI revolution will happen over the next few years and will be based on world models, which are systems that can understand and learn from the real world and "messy" data, rather than Large Language Models (LLMs).

"My colleagues who work on LLM and generative AI don't like me very much, and I am not very popular in Silicon Valley when I say this, but those will not be generative models," LeCun said.

He explained that current systems manipulate language so well that it "fools a lot of people" into thinking that we have systems that are human-level intelligent simply because we equate language ability with humanity.

However, LeCun cautioned that this is a mistake that computer scientists and people around them have made in AI for the last 70 years: discovering a new paradigm for AI and assuming that this paradigm will lead us to systems that have human-level intelligence.

Silicon Valley focused only on LLMs

He criticised the current industry, particularly Silicon Valley which houses the dominant AI players, for being entirely focused on LLMs.

"Everybody is stealing each other's engineers, and working on the same thing because nobody can afford to do something slightly different and then run the risk of falling behind," LeCun said "That creates a monoculture that makes the industry a little blind."

He argued that the future of breakthrough research hence remains in the hands of academia. "Preventing LLM from sucking the oxygen out of every room it gets into, I think, is the first step," he noted.

LeCun also warned that current technology is limited and the agentic systems are fundamentally flawed because they cannot predict the consequences of their actions. "Current agentic systems are not able to predict the consequences of their actions, making it a terrible way of planning actions," he said.

Why world models are better than LLMs

LeCun noted that understanding the real world is much more complicated than language, since it is continuous, noisy, and more complicated. While LLMs can accumulate knowledge about the real world, but they don't really understand the human nature, he noted.

"I've been making a joke for many years that your house cat is smarter than the biggest LLMs. In many ways that's true," he said "The next revolution are systems that learn and understand how the world works, a little bit like children who open their eyes."

These intelligent systems will be able to plan and reason by predicting the consequences of their actions before taking them.

"That's why I left Meta and created this company (AMI Labs), because I think the time is ripe for trying to make this real," he said.

Applications of world models

LeCun noted that these systems will have a lot of applications across the world.

"In the future, all of us will be working around with an AI assistant that will essentially amplify our own intelligence," he said "It's like all of us will be the leader of a staff of virtual people who are smarter than us"

LeCun added, "I'm very familiar with the concept of working with people who are smarter than you. It's the greatest thing that can happen to you. So we shouldn't feel threatened by it. It will allow people to get more knowledgeable, more educated, and make more rational choices."