LeCun says we're still far from human-level AI: where's the L5 self-driving car?
pswider · x · 2026-10-03
Yann LeCun pushed back on viral robot hype, noting the video in question was two years old and that we remain far from human-level AI. He concedes superhuman performance in math, coding, and answering known-answer questions, but asks where the Level-5 self-driving car is — pointing out Tesla FSD is rated L2 and Waymo L4. Critics replied that kid robots have started serving at McDonald's in China, underscoring the split over how fast AI progress really is.
Related event: LeCun Reaffirms AGI Is Not Coming Within Two Years(6 posts)→
More from AGI Musings
- Avi Loeb: One non-human UAP in a million would matter more than AI — examachine · 2026-10-03
- 'Intelligence explosion': arXiv submission chart goes vertical on Reddit — Many_Consequence_337 · 2026-10-03
- Researcher argues 'loss of humanity's long-term potential' is a flawed definition of existential risk — birchlse · 2026-10-03
- Hutter et al. Argue Generalization Requires Universal Induction in New arXiv Paper — examachine · 2026-10-03
- ChatGPT surfaces relevant past context unprompted, user reports from long-term testing — CellAgentLab · 2026-10-03
- A source-heavy 30-minute video essay separates real AI extinction risk from hype — Ok-Professor7130 · 2026-10-03