zlacker

To me it's not surprising because of the background to their formation: individually multiple orgs could have shipped GPT-3.5/4 with their resources but didn't because they were crippled by a potent mix of bureaucracy and self-sabtoage

They weren't attracted to OpenAI by money alone, a chance to actually ship their lives' work was a big part of it. So regardless of what the stated goals were, it'd never be surprising to see them prioritize the one thing that differentiated OpenAI from the alternatives