Waymo Calls Tesla's Path to Self-Driving a 'False Summit' — What It Means for Your Car Today

Waymo just published a list of "10 AI lessons" drawn from more than 200 million fully autonomous miles, and according to Electrek, several of them land as a direct rebuke of Tesla's self-driving strategy — without ever naming Tesla once.

The sharpest line is lesson 10: trying to upgrade a driver-assist system into full autonomy is a "false summit." That is, almost exactly, the path Tesla is walking with Full Self-Driving.

What Waymo Actually Said

The post came from Srikanth Thirumalai, Waymo's head of AI foundations, and every lesson is framed around safety and the company's driverless mileage record. Lesson 10 is the one aimed at the L2-to-L4 playbook:

Simply improving a driver-assist system (L2) for full autonomy is a false summit. True L4 maturity can only be safely achieved by a purpose-built system, validated on closed courses and hardened by the uncompromising experience of driving without a human in the car. — Waymo

Waymo's argument is that neither billions of simulated miles nor millions of "supervised" miles with a human present can teach a system what it learns once there's no human backup at all. That's a pointed contrast with Tesla, which trains FSD on data from customer cars whose drivers remain legally responsible for the vehicle.

Two other lessons hit Tesla just as squarely:

• Lesson 1 — cameras alone aren't enough. Waymo states that camera-only sensing can't deliver safe, full-scale autonomy, and says it combines cameras, lidar, and radar for redundancy. Tesla runs a camera-only "Tesla Vision" stack.
• Lesson 4 — no black boxes. Waymo warns against pure end-to-end neural networks that turn raw camera pixels directly into steering commands: "You can't build trust with a black box." Tesla has spent two years selling exactly that end-to-end approach as the core of FSD v14 and v15.

The Numbers Behind the Jab

Waymo has scale to back up the confidence. Per Electrek, it's now serving more than 500,000 paid, fully driverless rides every week across its US markets, with a target of one million by year's end.

Tesla's Robotaxi service, by comparison, launched in Austin with human safety monitors in the car and by some counts has shrunk rather than grown. Elon Musk has promised widespread "unsupervised" FSD by year-end — a target that has moved repeatedly. A year into the program, Tesla confirmed last month it has driven 380,000 unsupervised miles total. Electrek notes that's roughly a single day of driving for Waymo.

One Austin resident in the comments summed up the street-level view: he sees a Robotaxi maybe once a month inside the service area, and a few dozen Waymos every day. His conclusion — Tesla is winning the driver-assist race but isn't close on true autonomy — is a fair reading of where things stand.

What This Actually Means for Your Tesla

Here's the part that matters if you own the car rather than the stock: nothing about this changes what your Model 3, Model Y, Model S, Model X, or Cybertruck does tomorrow morning. FSD remains a supervised system. You are the driver. The "false summit" debate is about whether that ever stops being true — not about whether it's true today.

So the practical takeaway is to keep treating your car as a very capable driver-assist vehicle, and set it up accordingly:

• Keep your cameras clean. This is the single most underrated FSD tip, and it's the direct consequence of a camera-only stack. A filmy B-pillar camera or a road-salt-crusted repeater degrades the system's inputs with no lidar or radar to fall back on. A microfiber and a small detailing kit in the door pocket does more for FSD behavior than any software setting.
• Protect the front camera housing and pillar glass. Bug splatter, hard water spots, and winter grime all sit right in the sensor's field of view. Cheap to prevent, annoying to diagnose.
• Add a real dash cam setup, not just Sentry. If you're supervising an end-to-end system that even its critics call a black box, you want your own record of what happened. A high-endurance drive sized properly for your car, and a USB hub that keeps Sentry and TeslaCam writing reliably, is basic insurance.
• Don't defeat the driver monitoring. Steering-wheel weights and similar gadgets are exactly the wrong response to a supervised system. They're also a fast way to lose FSD access.
• Charge on your own terms. Whatever happens with autonomy, the daily reality of ownership is charging. A properly rated Level 2 setup at home, plus the right adapters in the trunk for public stations, is the accessory decision that pays back every week.
• Keep roadside gear on board. Supervised means you're still the one handling a flat, a dead 12V, or a stranded stop. A tire inflator, a jump pack, and a reflective triangle belong in every EV, autonomy debate or not.

You can match cabin gear, camera-cleaning supplies, charging adapters, and roadside kits to your exact model and year in our full accessories catalog.

The Bottom Line

Waymo didn't name Tesla, but it didn't need to. Camera-only sensing, an end-to-end black box, and the belief you can incrementally upgrade a consumer driver-assist product into a robotaxi — those are three specific Tesla decisions, and Waymo just called all three of them out.

Whether Waymo is right is genuinely unsettled. Tesla's bet is that generalization arrives suddenly and its fleet advantage takes over; there's no clear evidence yet that's happening, but there's no proof it can't. What's not unsettled is today: your Tesla still needs a supervising driver, clean cameras, and a trunk that's ready for the boring problems autonomy hasn't solved yet.