zlacker

>>apsec1+(OP)
The main question is what to expect from OpenAI now? No changes very unlikely, that would mean it was just a power grab. So two options remain: more open, more closed. How about slow down and open up? Hope they wouldn't dumb down GPT4. If they allow to use their models to generate training sets (which is prohibited now, AFAIK), that would be nice.

>>two_in+XB
My guess is that the immediate roadmap has already been locked in up to X months out. So, we'll likely never know what the "changes" will be. Short term changes are likely still Altman's work. Long term is the next decision maker.