Tesla has been testing four autonomous cars on California roads, totaling 550 miles between October and December, reporting 0.33 disengagements per mile
Google/Waymo so far ahead it's ridiculous Timothy J. Seppala / Engadget : Google's self-driving cars are getting better at autonomy Roger Fingas / AppleInsider : Alphabet's Waymo sharpens self-driving car tech, expands testing lead over rivals like Apple Steve Kovach / Business Insider : Waymo's self-driving cars are performing way better Alan Ohnsman / Forbes : Tesla Has Little To Show In California's Tally Of Autonomous Car Test Data Andrew Krok / CNET : Tesla is now testing autonomous vehicles on public California roads Fred Lambert / Electrek : Tesla had 4 self-driving Model X prototypes last year and drove over 500 autonomous miles on public roads Tweets: E.W. Niedermeyer / @tweetermeyer : *Extremely Elon Musk voice* Orders of magnitude... Tesla: 330 disengagements/thousand miles Waymo: .2 disengagements/thousand miles http://twitter.com/... Anne Riley Moffat / @a_riley17 : Do these CA self-driving car disengagement metrics matter? “It doesn't tell you everything” -analyst https://www.bloomberg.com/... via @danahull Mark Bergen / @mhbergen : Elon is testing four self-driving cars in Cali. But Larry's cars are performing at a better rate. http://www.bloomberg.com/...
Context & Ripple Effects
Tesla's first appearance in California's autonomous-vehicle reporting is a modest one: four Model X prototypes covering just 550 miles between October and December, at 0.33 disengagements per mile. The state's DMV disclosure regime is what makes this visible at all — it turned a private testing program into an annual public scoreboard.
That scoreboard has since become the industry's yardstick. A year later, Waymo and GM led the same report with tens of thousands of miles each, and Apple's filings — from roughly one disengagement per mile to one every 145 miles by 2020 — show how closely rivals watch these numbers.
First-order effects
- Tesla enters the DMV data pool with a sample size two orders of magnitude below Waymo's, so its 0.33 disengagements per mile reads as a starting point rather than a competitive result.
- The filing confirms Tesla is running public-road autonomy testing in California alongside its existing programs, putting it under the same mandatory disclosure rules as Google, GM, and Apple.
Second-order effects
- Rivals now have a concrete number to measure against: Waymo's coverage lead frames every subsequent Tesla filing, and press comparisons like Forbes' 'little to show' tally become an annual ritual.
- Apple's trajectory in the same dataset — improving from once-per-mile disengagements toward one every 145 miles — shows entrants using the metric to signal progress, forcing everyone in the pool to publish year-over-year deltas.
Third-order effects
- Disengagements per mile hardens into the de facto public benchmark for autonomy readiness, even though tiny fleets like Tesla's make the ratio statistically fragile — a structural tension that persists as long as the DMV publishes raw counts.
- The gap this filing reveals never fully closes on the metric that matters commercially: nearly a decade later, Tesla's Austin robotaxi service still trails Waymo's driverless fleet by a wide margin, suggesting early test-mileage deficits foreshadowed deployment-scale ones.
The trend: California's DMV disclosures have turned autonomous-vehicle testing into an annual public ranking, where mileage scale and disengagement rates increasingly predict who can actually deploy driverless services.