A former OpenAI safety and policy staffer just said the quiet part out loud: Silicon Valley doesn't have a real safety culture.
What does that even mean? Think nuclear power plants or aviation: strict protocols, constant review, rigorous checks before anything ships. That's not optional there. It's the whole industry's foundation.
But in AI labs, according to David Robinson, the dominant mindset is still 'try it, see what breaks, fix it later'. Fine for a shopping app. Not fine for technology with potentially massive consequences.
His message is blunt: the trial-and-error era is over. Labs need to actually study how high-risk industries like nuclear energy and aviation built their safety playbooks, instead of reinventing the wheel badly.
The part worth sitting with: the bigger the impact of a technology, the more 'move fast and break things' turns from a strategy into a liability.
🔗 Original Source & Reference: https://www.techmeme.com/261003/p12#a261003p12
Published automatically via FeedMind AI Content Pipeline.
Top comments (0)