Musk hails Tesla FSD V15's new architecture: 25x better critical intervention rate
elonmusk · x · 2026-11-22
Elon Musk calls Tesla FSD V15 a "banger" release with a completely new end-to-end architecture he claims drives better than humans.
Per the full release notes shared in the quote:
- Upgraded vision encoder shared across FSD, Robotaxi, and Optimus, trained on millions of internet videos for better generalization and data efficiency, improving the critical intervention rate by 25x.
- Tesla Reality View replaces all explicit model signals with a single end-to-end visualization using less total compute.
- Improved navigation via an end-to-end learned model trained on fleet route preference data.
- Personalized parking location selection, with maps now showing multiple parking options.
More from Embodied
- VR-teleop-kit demo: clamp robot to table, record teleop data with lerobot — DominiqueCAPaul · 2026-11-22
- Building a bilingual Stack-chan desktop bot with local LFM-3-Audio — alexcovo_eth · 2026-11-22
- Tesla's overhead parking view is generated video, not stitched camera feeds — PTrubey · 2026-11-22
- Sentdex's LLM robot harness solves unseen toy task on first attempt — Sentdex · 2026-11-22
- Frontier LLMs solved robot in-context learning as a side effect, says researcher — YouJiacheng · 2026-11-22
- PhD intern's trick: carbon fiber TPU turns a Fin-Ray gripper into its own touch sensor via flex-dependent resistance — burhop · 2026-11-22