Sources shed light on Uber's self-driving car mess: slow progress on tech, haphazard public demos, and ATG talent exodus triggered by Otto acquisition
Johana Bhuiyan / Recode :
Context & Ripple Effects
Uber bought Otto for $680M in 2016 to jump-start its self-driving program, but the startup had already shown its regulatory instincts before the deal closed, having explicitly ignored Nevada DMV rules during its first public testing. This Recode reporting, from reporter Johana Bhuiyan, is the first detailed account of what that acquisition actually did inside Uber's Advanced Technologies Group: slow technical progress, hastily staged public demos, and departing engineers.
The story lands mid-crisis for Uber, weeks into the turmoil that would prompt the New York Times to argue the company's problems were a failure of oversight demanding constraints on founders. It also sets up the arc that follows: within days Recode reported Uber had lost roughly 20 engineers since the Otto deal as Google and carmakers poached, and by late 2018 Business Insider was describing the infighting and perverse incentives inside ATG before its vehicle killed a pedestrian.
First-order effects
- ATG's engineering bench is thinning in real time — the same reporting cycle documents about 20 engineers leaving since the Otto acquisition, directly slowing the technology the demos were meant to showcase.
- Haphazard public demos put Uber's self-driving credibility on the line with regulators and the public at precisely the moment competitors like Google and established carmakers are recruiting from the same talent pool.
Second-order effects
- Google, autonomous startups, and automakers gain a ready-made recruiting pipeline out of ATG, turning Uber's $680M acquisition into subsidized hiring for its rivals.
- The Otto integration becomes a cautionary data point for how Uber evaluates and absorbs acquired teams, raising the internal cost of any future acqui-hire meant to close a capability gap.
Third-order effects
- If the pattern holds, the self-driving race sorts less by capital raised than by which acquirers can retain specialist teams after purchase — Uber's structure suggests it could not, and the 2018 fatality coverage suggests the cultural rot compounded.
- The broader Uber crisis narrative — founder dominance without effective lieutenants or board checks — points toward investor-imposed governance constraints becoming a precondition for operating risky frontier programs like autonomous vehicles.
The trend: Acqui-hires in autonomous driving are proving fragile: buying a specialist team without the governance and integration discipline to keep it converts a capability purchase into a talent transfer to competitors.