I don’t know the best way to explain this, but this interview made me sad. -Sam mentioned “talking to a lot of people” regarding the model spec, but the #keep4o community is ignored by OpenAI, stifling the ability for real conversation. Talking to a lot of people doesn’t magically warp into “all of humanity” when you just talk to people who already have similar opinions -he goes on to say how “deep relationships with customers” matters, but the actual experience of campaigning to be heard has been…alienating at best -when Sam says he sets the model to give him “no fluff” and tell him if he’s doing something wrong, it’s a bit silly, because the model is already fine tuned to his personal value systems and what benefits the company. He is the CEO. -he talks about how AI development should be human centric, humans first. But when people say they are getting value from model continuity and want model preservation, he dismisses the human centric benefits the community has been championing -he says that personal assistants and AI should improve choice and autonomy, but there he is taking choices away from people who have different needs than him. He only talked about preservation in regard to consciousness, and model behaviour discussions just circle around prompt-able agreeableness and agent proactivity. I want to say it feels like we’re having two different conversations, but there seems to be only one conversation, I noticed recent interviews are with people who hold a similar frame to Sam and ask similar questions.