Yann LeCunTuring Award laureate, NYU professor and former Meta chief AI scientist; the leading sceptic of LLM-driven AGI timelines.
I think the shelf life of the current paradigm is fairly short, probably three to five years.
World Economic Forum, Davos (reported by TechCrunch) · Jan 23, 2025 · source · archived snapshot
Admitted by the site's editor alone — Seeded by the site owner at launch; single-operator admission (ADR-0012).
Normalized claim
The current large language model paradigm will be superseded as the core of frontier AI systems by 2030.
Confidence
medium (60%)
Deadline
due Jan 23, 2030 (inferred)
Resolution rule
Hit if by January 2030 the leading frontier systems are built around a different architecture than autoregressive LLMs (for example world models or JEPA-style systems), as described by the labs themselves. Miss if LLM-centred systems still lead.
Crowd vote
Vote to see the crowd.
Your call
Locked at the deadline. No odds, no money — just your record.
More from Yann LeCun
It's going to take years before we can get everything here to work, if not a decade.
OPEN