← Back to live feed1 story
Former OpenAI safety lead David Robinson stated that the company used over 100 times more agentic compute per day in September compared to nine months prior. Robinson resigned after concluding that OpenAI is unready to align or control systems that could out-think humans, and he characterized the firm's safety culture as broken.
Sign in to suggest
Key sources
- SOURCE@stephencouncil“OpenAI firing three researchers over allegedly mishandling information”x.com
- SOURCE@adriennelaf“among the longest-tenured employees at OpenAI—and oversaw safety reports on 12 frontier launches”x.com
- SOURCE@ezraklein“on track to build something that could think circles around us — and we were this far from being ready for that at a safety and alignment level”x.com
- SUPPORT@_nathancalvin“OpenAI researchers used more than 100x more agentic compute per day in September than they did nine months prior”x.com
- SUPPORT@miles_brundage“regret helping spread the idea of “iterative deployment” which maybe briefly made sense in the GPT-3 era but makes no sense at all after many deaths have been tied to AI”x.com
- SUPPORT@jachiam0“professional safety engineering practices from other fields haven't taken enough root in AI safety”x.com
- SUPPORT@matthewstoller“Don’t trust AI safety people from the big firms, trust *actual* cybersecurity experts”x.com
- SUPPORT@ryan_t_lowe“we need layers of safety like in other high-risk fields like nuclear energy”x.com