In a ’10 AI lessons’ post from 200M+ driverless miles, Waymo argued that turning driver-assist into full autonomy is a ‘false summit’ and that cameras alone can’t get there — a direct swipe at Tesla’s FSD.
Waymo just told Tesla its entire self-driving strategy is a dead end — without saying the word ‘Tesla’ even once.
Waymo just told Tesla its entire self-driving strategy is a dead end — without saying the word “Tesla” even once. In a post rounding up “10 AI lessons” from driving more than 200 million fully autonomous miles, the Alphabet-owned company argued that trying to upgrade a driver-assist system into full autonomy is a “false summit.” That is, almost to the letter, exactly what Tesla is attempting with Full Self-Driving.
The ‘false summit’
The post, published by Waymo’s head of AI foundations, Srikanth Thirumalai, frames all 10 lessons around safety and the company’s 200-million-mile driverless record. Lesson 10 is the dagger: simply improving a driver-assist (L2) system for full autonomy is a false summit, and true L4 maturity can only be achieved safely by a purpose-built system, validated on closed courses and hardened by the experience of driving with no human in the car. Waymo’s point: neither billions of simulated miles nor millions of “supervised” human-present miles can replicate what a system learns once there’s no human backup to catch its mistakes.
Cameras vs. the sensor stack
The other lessons keep the contrast going. Lesson 1 states plainly that cameras alone can’t deliver safe, full-scale autonomy — Waymo combines cameras, lidar and radar for redundancy, while Tesla runs a camera-only “Tesla Vision” stack. Lesson 4 warns against pure end-to-end neural networks that convert raw camera input directly into steering commands, another hallmark of Tesla’s approach.
Why it’s pointed
Tesla trains FSD on data from a fleet of customer cars whose drivers remain legally responsible for the vehicle. Waymo’s argument is that this supervised data, however vast, teaches the wrong lesson — a system that assumes a human will save it never learns to truly drive alone. It’s a philosophical fight with billions of dollars riding on who’s right: Tesla is betting scale and cameras will get there; Waymo is betting they’ll stall at a false peak. Two hundred million driverless miles give Waymo the standing to say it out loud.
Written for Red Robot with AI assistance and human editing. Based on reporting by Electrek.