3 ms·
Also, there's no real evidence of Microsoft being philosophically opposed to releasing model weights. That's entirely come from the AI safety people who want mo
by kmeisthax 3y ago
Also, there's no real evidence of Microsoft being philosophically opposed to releasing model weights. That's entirely come from the AI safety people who want models with reactively updated alignment controls. If anything having model weights would mean being able to walk away from OpenAI and keep the thing that makes them valuable.