ClaimsPeople

Yann LeCun

Yann LeCun is a computer scientist who shared the 2018 ACM Turing Award, is a Jacob T. Schwartz Professor at NYU's Courant Institute, was Meta's Chief AI Scientist and founding director of FAIR, and is Executive Chairman of AMI Labs.

1,175 claims collected, 197 researched, 84 resolved. Report of 4 November 2026, method 0.1.0+c876030c.

Scores by kind of claim

Each kind of claim is scored on its own, out of 100: the share that held up, with the range it likely falls in. Fewer resolved claims give a wider range.

  • Factual claims

    Statements about the present or the past, checked against the record: did they hold up?

    78 72 83

    High confidence, 79 factual claims resolved. From 16 documents, 2006 to 2026; 19 documents sampled from 33 candidates found.

    Not scored: 43 unresolvable.

  • Predictions

    Statements about events outside the person's control, checked once they fell due: did they come true?

    64 43 83

    Low confidence, 5 predictions resolved. From 4 documents, 2010 to 2024; 19 documents sampled from 33 candidates found.

    Not scored: 66 unresolvable. 3 more fall due later.

    Does not yet allow for how likely each prediction was when made.

  • Promises

    Commitments to the person's or organisation's own future action, with something to deliver: was it delivered?

    Insufficient record: 0 promises resolved, 5 needed for a score.

    Not scored: 1 unresolvable.

All types combined: 77 (likely 71 to 82, high confidence), from 84 claims: 79 factual claims, 5 predictions. It mixes kinds of claim in different proportions for different people, so compare people kind by kind.

Yann LeCun's claims

3 claims

Back to the list

Yann LeCun

Open its page

Prediction AI capabilities

“the blueprint of those systems will be extremely different from auto aggressive LLMs.”

Yann LeCun, interviewee 8 March 2024 Transcript for Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI (Lex Fridman Podcast)

LeCun said he thinks that over the next few years there will be many AI systems with planning capability, and that their blueprint will be extremely different from autoregressive LLMs.

Not yet due The date it refers to has not come.

What would count as true or false

Judge this at the horizon. The claim gives no date, so 'the next few years' after 8 March 2024 is taken to end in March 2027. The claim has two parts: (a) many AI systems with planning capability exist and are in use or published, and (b) their blueprint is extremely different from autoregressive LLMs, meaning the core architecture is not an autoregressive next-token language model. TRUE: by the horizon, the main planning-capable AI systems (deployed or prominently published) are mostly built on architectures that are not autoregressive LLMs, such as world-model or joint-embedding predictive architectures with planning or objective-driven optimisation. MOSTLY TRUE: both parts hold, but the non-autoregressive systems are prominent rather than clearly dominant, or the difference is large but not total, with an LLM kept as a minor component. MIXED: many planning-capable systems exist and a substantial share are fundamentally non-autoregressive, but a comparable share are autoregressive-LLM-based. MOSTLY FALSE: many planning-capable systems exist, but they are predominantly autoregressive LLMs or LLM-based agents, such as reasoning models with chain of thought or tool use, and non-autoregressive designs are a minority. FALSE: planning-capable systems are almost all autoregressive-LLM-based, or there are few planning-capable AI systems at all.

Evidence

  1. Extended Summary for Policymakers International AI Safety Report 2026 (Chair-led, 100+ independent experts) · February 2026
    “The latter approach has led to the development of ‘reasoning models’, which generate an explicit step-by-step explanation (a ‘chain of thought’) before providing a final answer.”
  2. Extended Summary for Policymakers International AI Safety Report 2026 (Chair-led, 100+ independent experts) · February 2026
    “Results from a December 2024 survey of 67 deployed AI agents.”
  3. LLM-Based Agent Planning Emergent Mind · 2 February 2026
    “LLM-based agent planning is the paradigm in which LLMs serve as the central sequential decision-making module for autonomous or semi-autonomous software agents.”
  4. Introducing the V-JEPA 2 world model and new benchmarks for physical reasoning Meta AI · 11 June 2025
    “Starting from its observed current state, the robot then plans by using the predictor to imagine the consequences of taking a collection of candidate actions and rating the candidates based on how close they get to the desired goal.”

Reasoning

The criteria judge which architectures dominate planning-capable AI systems at the March 2027 horizon, and that date has not yet arrived. The current evidence describes LLM-based planning: E1 describes reasoning models that produce a chain of thought before answering, and E3 describes a paradigm in which LLMs serve as the central decision-making module for agents, though neither measures what share of planning-capable systems these make up. E4 shows a world-model planner (V-JEPA 2) that plans robot actions by using a predictor to imagine the consequences of candidate actions. The balance at the horizon is a state of affairs that could still shift, so it is not yet settled.

Cite this claim

Yann LeCun, “LeCun said he thinks that over the next few years there will be many AI systems with planning capability, and that their blueprint will be extremely different from autoregressive LLMs.” (said 8 March 2024). Verdict: not yet due. Reliability API, claim c_4daa9ddb5073, method 0.1.0+c876030c. https://epistacklabs.com/claims/c_4daa9ddb5073

Claims without a verdict

Is a verdict here wrong, or do you speak for Yann LeCun? Reply on the claim's own page, with the form at its foot (how replies work). Verdicts are made automatically; how verdicts and scores are made.