Reactions from ranked influencers
3 posts
But this is twitter. People are terrible here, and you can get piled on for all kinds of reasons. Happens to me too. Unfortunately, this is our town square or something. Diversity of opinion is crucial, just like diversity of AI models. Now, in the case of Dean's previous post, the reasoning was just not very good.
Some observations, in the wake of the Dean Ball open source discourse. In my long tenure at OpenAI and as a public commentator on social media, I never felt pressure from OAI leadership or the comms department to say or post anything; I always gave my own view. It's possible I had an exceptional degree of freedom due to my tenure and personal platform, but that freedom was still meaningful. However, I felt huge pressure from the outside world to modulate what I said and how I said it. It is worth considering what that pressure entails and how it constrains what an organization like OpenAI can or can't do as a participant in the discourse. Many people were determined to see me as a representative of org policy, even when I was putting forward my own views and analysis for consideration without any connection to the org's policy-making process about the topic du jour. As a result I've felt a need to be exceedingly responsible in many ways, to be particularly even-handed, to narrow my aperture of discussion topics to specific domains (e.g. not commenting on political controversies, or if so, doing so only mildly). These are not bad things, of course. I'm inclined in that direction anyway; I'd rather be a fair and even-handed person than a pure partisan. But I point this out because the pressure came from outside, not in. There is something about the world that is not really willing to allow a general-purpose heterodox think tank for strategy to develop inside of a place like OpenAI. When people have highly-polarized views of a parent organization, there is no way for that parent organization to put forward ideas that are regarded as neutral or unbiased; they are always seen as somehow representing or hinting at or colored by the interests of the parent organization. It is nearly impossible to get away from the miasma of suspicion. But this winds up constraining the freedom of thought that parent organizations can comfortably host inside the walls. OpenAI stands to benefit enormously from having a Dean Ball type bringing in varied points of views and publishing strategy briefs. Especially as long as Dean and others continue to speak their minds without feeling a requirement to advance the interests of the company, and as long as the company remains firm in its commitment to protect its independent voices. The outside world stands to benefit as well from having people very close to the frontier technology doing their vivid imagination in public. However, these benefits can't be easily realized in the current discourse environment. There is too much punishment from the outside. The outside puts a lot of pressure on people in companies to simply shut up; there is not a lot of willingness to see people doing a sense-making process in public. I find this frustrating; I wish it were otherwise. It will take a lot of people fighting very hard to protect unpopular positions to enable more open discussion and to empower people within frontier labs to do more of their sense-making in public.
I really do hope Dean keeps writing unfiltered thoughts. I doubt he will want to after this episode but it's far more valuable to have radical transparency by thinking in public than no transparency at all
I’m afraid to tell you that it is effectively impossible to do the kind of writing I used to do on this website, not because anyone at OpenAI censors me but because of the sheer volume of hostility I get for sharing my analysis as a frontier lab employee. I enjoyed writing quick takes on this website for one basic reason: I could get rapid feedback on my own ideation process in real time. Post the early version of the take here, see the criticism; then refine, sharpen, and repeat. Unfortunately now that feature of this site is gone, because the feedback I get is now almost exclusively colored by resentment at the fact that I work at a frontier lab or other forms of hatred for my employer. The feedback signal is essentially useless now, so writing on here is not fruitful for me anymore. Literally everything I write now is responded to with “of course you said that because <OpenAI>.” I am truly just writing what I think and would have written anyway, but everyone reads what I say in the shrieking tone of “this is what openai thinks!!!!” (to be clear, my posts are not what openai thinks). This is an unpleasant and more importantly unproductive pattern for me. I anticipate that the shape of this account will change significantly as a result. I do not currently know how. It will not become a LinkedIn feed. It will change in some other way. It will no longer be a real-time accounting of my own thinking as it develops, since this is precisely the thing that seems impossible to do now. That will have to shift to private channels.