When a departing OpenAI researcher in May discussed the need to learn to control ultra-intelligent AI, LeCun said: "It seems to me that before 'urgently figuring out how to control AI systems much smarter than us' we need to have the beginning of a hint of a design for a system smarter than a house cat."
He likes the cat metaphor because cants have a mental model of the physical world, persistent memory, some reasoning ability, and a capacity for planning. None of these qualities exist in today's "frontier" AIs, including those made by Meta.
LeCun shared a Turing Award with Geoffrey Hinton and Hoshua Bengio (who hopes LeCun is correct but adds, "I don't think we should leave it to the competition between companies and the profit motive alone to protect the public and democracy. That is why I believe we need governments involved.")
LeCun still believes AI is a potent tool — even as Meta joins the quest for artificial general intelligence.
He cites many examples of how AI has become enormously crucial at Meta and has driven its scale and revenue to the point that it's now valued at around $1.5 trillion.
“AI is integral to everything from real-time translation to content moderation at Meta, which, in addition to its Fundamental AI Research team, FAIR, has a product-focused AI group called GenAI that is pursuing ever-better versions of its large language models. "The impact on Meta has been enormous," he says.
At the same time, he is convinced that today's AIs aren't intelligent in any meaningful sense—and that many others in the field, especially at AI startups, are ready to extrapolate their recent development in ways that he finds ridiculous. OpenAI's Sam Altman said last month that we could have artificial general intelligence within "a few thousand days."
But creating an AI this capable could easily take decades, [LeCun] says — and today's dominant approach won't get us there.
He bets that research on AIs that work in a fundamentally different way will set us on a path to human-level intelligence.
These hypothetical future AIs could take many forms, but work being done at FAIR to digest video from the real world is among the projects that currently excite LeCun. The idea is to create models that learn in a way that's analogous to how a baby animal does by building a world model from the visual information it takes in.
In contrast, today's AI models "are just predicting the next word in a text," he says. Because of their enormous memory capacity, they can seem to be reasoning when, in fact, they're merely regurgitating information they've already been trained on.
