World model
An internal representation of how the world is structured, built from data, that lets an agent predict and explain observations.
As defined in cwt_alison-gopnik
Other definitions
A simplified mental representation of the physical world and how it works, allowing prediction and manipulation of physical objects and scenarios
Metas AI Boss Says He DONE With LLMS...In AI, an internal representation of how the physical and social world works, including intuitive physics, object permanence, and causal relationships. Necessary for AI systems to plan and act effectively in complex environments.
Sir Demis Hassabis on The Future of Knowledge | Institute for Advanced Study
Where it comes up · 35
- Metas AI Boss Says He DONE With LLMS...11 claims
- The Shape of AI to Come! Yann LeCun at AI Action Summit 202511 claims
- Yann LeCun Says LLMs Have 2 Years Left…10 claims
- The future of intelligence | Demis Hassabis (Co-founder and CEO of DeepMind)10 claims
- Yann LeCun: World Models: Enabling the next AI revolution9 claims
- Yann LeCun: Special Lecture on AI and World Models8 claims
- Inference, Diffusion, World Models, and More | YC Paper Club7 claims
- The Future of Intelligence: A Conversation with Jeff Hawkins (Episode #255)6 claims
and 27 more
Most important claims · 10 of 129Search all claims for “World model” →
- Human memory is a reconstructive process, not a videotape—it is reassembled from familiar components—and imagination is the same process using the same semantic components recombined in novel ways for purposes like planning; this idea of pulling together world-model parts to simulate something new for planning is probably still missing from current AI systems.Demis Hassabis — Scaling, superhuman AIs, AlphaZero atop LLMs, AlphaFold
- Hierarchical planning allows AI systems to decompose complex goals into multiple levels of abstraction: high-level planning works with macro actions and abstract representations over longer timescales (e.g., 'go to the airport'), while lower levels plan detailed actions over shorter timescales (e.g., 'walk to the elevator, push button, descend')—humans do hierarchical planning constantly, but learning machines have almost no capability to do this; nearly all robots use hand-crafted representations at each hierarchy level rather than learned abstractions.The Shape of AI to Come! Yann LeCun at AI Action Summit 2025
- There is a trade-off between model quality and search: AlphaZero plays beyond world-champion level using only tens of thousands of evaluated positions versus millions for brute-force systems like Deep Blue, while human grandmasters use only a few hundred—because richer, more accurate world models let you pick which nodes to expand, so improving models makes search more efficient and lets you go further.Demis Hassabis — Scaling, superhuman AIs, AlphaZero atop LLMs, AlphaFold
- The new LayWM research proves that efficient world models do not require infinite computational resources; instead, they require better mathematical design to achieve the same physics understanding on consumer-grade hardware.LeWorldModel: Stable End-to-End JEPA World Models From Pixels
- The historical problem with training world models through prediction is 'representation collapse'—where an AI can achieve perfect prediction accuracy by mapping all inputs to the same blank representation (e.g., zero) without learning any actual physics, because the loss function only penalizes prediction error, not learning quality.LeWorldModel: Stable End-to-End JEPA World Models From Pixels
- The primary safety concerns for advanced AI systems fall into two categories: malicious actors (individuals, rogue nations) repurposing general-purpose AI technology originally developed for beneficial uses like medicine, and inherent risks from increasingly autonomous and agentic AI systems that may become difficult to control as they approach artificial general intelligence.Sir Demis Hassabis on The Future of Knowledge | Institute for Advanced Study
- Large language models are fundamentally limited to language manipulation and reasoning only in domains where language itself is the substrate of reasoning, making them incapable of building world models that predict physical consequences of actions.Yann LeCun's $1B Bet Against LLMs [Part 1]
- Humans can learn a new word like 'blanket' from a single example and immediately use it grammatically and semantically across infinite novel sentences, because they map language onto a rule-governed world model that supplies syntax, morphology, and semantics—getting off to the races with one training example, unlike deep learning.184 | Gary Marcus on Artificial Intelligence and Common Sense
- Language models encode values and ways of thinking about the world — an early Llama translated to French sounded like an American speaking French — and models out of China have certain values baked in that can't be removed with a light fine-tune, which is why building standards around American models matters.Mark Zuckerberg — AI will write most Meta code in 18 months
- The next era of AI will involve agents that can accomplish things more autonomously, similar to game-playing programs but more generalized with world models, raising questions about how to control and deploy such systems.Sir Demis Hassabis on The Future of Knowledge | Institute for Advanced Study