Gary Marcus, director of Uber's AI Labs, to leave his role after four months with the company and become a “senior adviser” to Uber
Gary Marcus, a research scientist who joined Uber four months ago as director of its AI labs, is leaving the company, according to sources.
Context & Ripple Effects
Gary Marcus arrived at Uber only through acquisition: Uber bought his startup Geometric Intelligence in December 2016 and installed Marcus and cofounder Zoubin Ghahramani as co-directors of the new A.I. Labs research arm. Four months into that arrangement, Marcus is stepping out of the director role into a senior adviser position.
The departure is notable because it is so fast — the lab was pitched as a marquee research bet, and one of its two founding directors is exiting operational leadership before the unit has publicly shipped anything.
First-order effects
- Zoubin Ghahramani becomes the sole remaining founding director of Uber A.I. Labs, carrying the Geometric Intelligence research agenda without his cofounder.
- Marcus retains a formal tie to Uber as senior adviser, but the company loses the public-facing research credibility he was hired to lend the lab.
Second-order effects
- Uber's autonomous-driving ambitions lean on in-house AI research; an early founder exit raises retention questions for the team Marcus brought in from Geometric Intelligence.
- Rivals building corporate labs around star researchers — Tesla's Autopilot operation under Andrej Karpathy being the closest analog — now have a cautionary data point on how quickly acquired-founder arrangements can unwind.
Third-order effects
- If the pattern holds across the industry — Karpathy's later exits from Tesla and OpenAI point the same direction — corporate AI labs will be treated as structurally dependent on their named founders, pushing companies toward adviser-style retainers rather than long-horizon director commitments.
- Acqui-hire-based labs may increasingly be judged on whether the founders stay past the honeymoon window, making founder retention a diligence item in future AI acquisitions.
The trend: Corporate AI labs built on acquired founders are proving fragile in their first years, with companies converting departing research leaders into advisory roles to preserve legitimacy.