Archive · 26 May 2025
Meta AI Chief: AI Lacks Human Traits Despite Advancements
Meta's AI expert, Yann LeCun, says that even though AI is getting better, it's not as smart as humans yet.
He says it's missing key things like understanding the world, remembering things, and planning.
Right now, AI is being improved with add-ons, but LeCun wants to try a different way to build AI using 'world-based models'.
These models try to imagine and understand how the world works, which could make AI smarter.
He thinks AI will take a while to be as smart as people.
Meta AI chief Yann LeCun believes current AI lacks four key human traits.
These traits include understanding the physical world, persistent memory, reasoning, and planning.
LeCun advocates for 'world-based models' as an alternative approach to current AI development.
Meta is exploring 'V-JEPA', a non-generative model, to improve AI's understanding.
LeCun disagrees with predictions that AI will quickly surpass human intelligence.
- Who
- Yann LeCun
- What
- Meta AI chief discusses limitations of current AI and future development.
- When
- Published on May 26, 2025
- Where
- N/A
- Why
- To explain the gap between current AI capabilities and human-level intelligence.
Others' Views
LeCun's Perspective
AI Threat
Others' Views
Some express concerns about AI overtaking humanity.
LeCun's Perspective
AI is not a natural phenomenon that will become dangerous.
- Key Figure
- Yann LeCun, Meta AI Chief
- Missing Human Traits
- Understanding physical world, persistent memory, reasoning, planning
- Proposed Approach
- World-based models
- Meta's Model
- V-JEPA
- LeCun's View
- AI will take time to achieve human-level intelligence