Sources: as Uber prepares to resume testing cars in autonomous mode in Pittsburgh, the cars still face issues, raising safety concerns
sometimes at night — at speeds as high as 55 miles an hour. Now, they won't operate at night or in wet weather, and they won't exceed 25 m.p.h., Uber said. http://www.nytimes.com/... Eric Boerer / @erokeric : “The cars have reacted more slowly than human drivers and struggled to pass so-called track validation tests, the last step before returning to city streets” READ: Uber knew theywere not safe the first time http://twitter.com/... Daisuke Wakabayashi / @daiwaka : It also was struggling to pass track validation tests, failing serious safety categories like being slow to recognize that a car wasn't yielding. Our story w/ @kateconger https://www.nytimes.com/...
Context & Ripple Effects
Uber has spent most of 2018 working its way back toward public-road autonomous testing in Pittsburgh: Mayor Bill Peduto demanded a federal investigation before any return after learning of the plans secondhand, sources reported an August restart target for Pittsburgh and possibly San Francisco, and last month Uber published a safety-measures report while applying to resume autonomous mode in Pittsburgh.
This report undercuts that arc: on the eve of resuming, the cars are failing track validation tests — including slow recognition of yielding vehicles — and Uber is imposing constraints it did not previously operate under: no night driving, no wet weather, and a 25 mph cap where the cars once ran up to 55 mph.
First-order effects
- Uber's Pittsburgh relaunch proceeds only in a heavily degraded envelope — daytime, dry conditions, 25 mph — which limits what the resumed testing can actually validate about the system.
- Safety advocates like Eric Boerer gain concrete evidence for the argument that Uber knew the cars were not ready when it first sought to return them to city streets.
Second-order effects
- Mayor Peduto's earlier demand for a federal investigation gets fresh leverage: a program that fails its own validation tests gives city and federal overseers grounds to condition or slow future expansion beyond Pittsburgh.
- Rivals' autonomous programs can now benchmark against Uber's restricted operating domain, making 'we test at night and in rain' a competitive differentiator in the race back to public roads.
Third-order effects
- If validation failures persist, the pattern points toward external gatekeeping of autonomous testing — cities and federal regulators demanding documented pass/fail thresholds rather than trusting operator self-certification.
- The gap between Uber's published safety report and its cars' actual performance shows why operational AI governance — audited validation, not policy documents — becomes the deciding factor in whether autonomous programs earn street access.
The trend: Autonomous vehicle programs are being forced from self-declared readiness toward externally validated, condition-limited operation, with each city's approval becoming the real gating milestone.