I think the point more is about advertising masking as research. Yes, OpenAI did a lot of research and there's no question about that and the quality of it and their results. But the question is if *Open*AI is doing internal research (as is common to any big company) or academic/open research. Researchers have different goals and so want to probe these models and understand them. I think the confusion comes from proprietary work looking like academic research. It is nice to peek behind the curtain, but it is unclear what the utility is.
I think a lot of the recent pushback is the social immune system going into effect. We just have to decide if this is an auto-immune disorder or not. The question comes to the advertisement-to-utility ratio. Have we crossed that threshold? What is the threshold? I think the immune response is happening because we don't know and our definition of that is as good as trying to define porn. I think there's a lot of confusion because we're not accurately codifying what the issues are. Or rather we all see different issues but are acting as if others have the same concerns; so we are talking on different pages.
I think a lot of the recent pushback is the social immune system going into effect. We just have to decide if this is an auto-immune disorder or not. The question comes to the advertisement-to-utility ratio. Have we crossed that threshold? What is the threshold? I think the immune response is happening because we don't know and our definition of that is as good as trying to define porn. I think there's a lot of confusion because we're not accurately codifying what the issues are. Or rather we all see different issues but are acting as if others have the same concerns; so we are talking on different pages.