Overview
- Elon Musk said on X in mid‑June that upcoming FSD releases will remember individual drivers’ parking preferences so cars target the spots drivers normally choose.
- Musk also said Grok, xAI’s in‑car chatbot, will be able to take natural‑language driving and parking instructions that feed high‑level intent into FSD’s planning layer in roughly three months.
- Tesla’s recent v14.3 updates — including a rewrite of its AI compiler/runtime using MLIR, faster reaction claims, improved vision encoders, and parking refinements — provide the technical groundwork for the changes.
- Company engineers and Tesla statements say safety validation is required because voice input that alters driving creates new edge cases such as misheard or adversarial commands, and regulators in Europe are already reviewing FSD safety data.
- If validated and approved, the changes could cut a leading cause of human takeovers at journey end and change how drivers interact with cars, while prompting closer oversight as Tesla expands supervised FSD in parts of the EU.