Software Engineer, Chip Design Tools (Google DeepMind) · Google · 2026
Evidence reviewed
Software engineer who resigned from Google DeepMind on September 24, 2026. O'Callahan said his team built improved tools for hardware-chip design whose principal effect would be to make AI cheaper, faster, and more pervasive. He said AI progress was already far too rapid, that moving to another Google project not accelerating AI was impractical from New Zealand, and that he no longer believed continuing the work was good for people.
O'Callahan wrote on September 24, 2026 that he was resigning from Google that day.
“My team’s goal is ultimately to make AI much cheaper and lower-latency, and I don’t think that’s good for people right now.”
O'Callahan said he valued his colleagues and the work itself, but concluded that even an indirect contribution to AI capability acceleration carried moral responsibility. He had publicly argued since 2023 for slower AI development and stronger regulation, and wrote in 2025 that he regularly considered resigning while weighing his ability to influence Google from within against his role in accelerating AI.
O'Callahan explains his support for a pause in advanced AI development. He identifies risks from economic displacement, malicious use, and loss of control, argues that society needs time to adapt, and calls for government regulation and industry self-regulation.
Writing while employed at Google DeepMind, O'Callahan describes conversations with colleagues about mass unemployment, human irrelevance, and extinction. He says he believed AI would most likely be disastrous for humanity and recounts his internal debate over whether to keep trying to influence the company or resign.
O'Callahan argues against treating advanced AI development as a race to be won. He calls for cooperation, verification, and negotiated limits between laboratories and countries rather than accelerating toward systems whose consequences and control mechanisms remain uncertain.
0 open forecasts · 1 warning and contemporaneous claims
AI will most likely be disastrous for humanity
“I believe AI will most likely be disastrous for humanity.”
Get an email when a new, source-verified profile is published.
Confirm your address by email. You can unsubscribe at any time, and we never share your address with third parties.
O'Callahan explains why he resigned from Google DeepMind. He says his work on chip-design tools would ultimately make AI cheaper, faster, and more pervasive, while AI progress was already moving too quickly for people and institutions to understand or adapt. He also discusses existential risk, power concentration, economic disruption, cybersecurity, and accountability.