Waymo co-CEO urges safety-first over “move fast and break things” mantra
Waymo’s co-CEO Dmitri Dolgov told a Y Combinator audience that AI built for physical tasks must prioritize safety rather than rapid, reckless development.
During a recent Y Combinator presentation, Waymo co-CEO Dmitri Dolgov argued that the classic Silicon Valley credo of “move fast and break things” is unsuitable for AI that interacts with the physical world. He suggested a new guiding principle—“move fast and ship safely”—to ensure that autonomous vehicles and robots are built with safety as a core requirement, not an afterthought. Dolgov highlighted the stark contrast between digital AI errors, which usually cost only a retry, and physical AI mistakes that can cost human lives.
He referenced Waymo’s generally low incident rate but acknowledged ongoing regulatory scrutiny, including a 2024 NHTSA probe and recent mass recalls after robotaxis entered construction zones. The comments come amid a wave of AI safety concerns sparked by recent security breaches at Hugging Face, OpenAI and Anthropic, underscoring the broader industry push for stronger safeguards.
Why it matters
Ensuring safety in physical AI like self-driving cars is crucial to prevent accidents that could cost lives.
In this story