Prof G Markets · Tuesday, September 15, 2026
Charlie O'Neal, co-head of model training at Base 10, suggests that calls to 'pace the frontier' in AI do not mean halting capability development. Instead, he posits that companies like OpenAI and Anthropic are increasing compute allocation towards safety and monitoring. O'Neal estimates OpenAI might dedicate up to 20% of its internal compute to these areas.
“The best way to view what they are actually asking for is not necessarily we're going to put a halt to capability development, we're going to put all these stringent, um, checks and like, you know, first party kind of things that are slowing us down.”
“The best way to view this is we are going to keep advancing the capabilities of the models. We are just going to allocate a little bit more compute to making those models safe.”
“There's rumors that Open AI is going to allocate up to 20% of internal compute into monitoring and safety.”