I am a forecaster on that question: the main doubt I had was if/when someone would try to do wordy things + game playing on a “single system”. Seemed plausible to me that this particular combination of capabilities never became an exciting area of research, so the date at which an AI can first do these things would then be substantially after this combination of tasks would be achievable with focused effort. Gato was a substantial update because it does exactly these tasks, so I no longer see much reason possibility that the benchmark is achieved only after the capabilities are substantially overshot.
I also tend to defer somewhat to the community.
I was at 2034 when the community was at 2042, and I updated further to 2026 on the Gato news.
I’m sorry to hear that your health is poor and you feel that this is all on you. Maybe you’re right about the likelihood of doom, and even if I knew you were, I’d be sorry that it troubles you this way.
I think you’ve done an amazing job of building the AI safety field and now, even when the field has a degree of momentum of its own, it does seem to be less focused on doom than it should be, and I think you continuing to push people to focus on doom is valuable.
I don’t think its easy to get people to take weird ideas seriously. I’ve had many experiences where I’ve had ideas about how people should change their approach to a project that weren’t particularly far out and (in my view) were right for very straightforward reasons, and yet for the most part I was ignored altogether. What you’ve accomplished in building the AI safety field is amazing because AI doom ideas seemed really crazy when you started talking about them.
Nevertheless, I think some of the things you’ve said in this post are counterproductive. Most of the post is good, but insulting people who might contribute to solving the problem is not, nor is demanding that people acknowledge that you are smarter than they are. I’m not telling you that people don’t deserve to be insulted, nor that you have no right to consider yourself smarter than them—I’m telling you that you shouldn’t say it in public.
My concrete suggestion is this: if you are criticising or otherwise passing pessimistic judgement on people or a group of people
—Give more details about what it is they’ve done to merit this criticism (“pretend plan that can fool EAs too ‘modest’ to trust their own judgments”—what are modest EAs actually doing that you think is wrong? Paying not enough attention to AI doom?)
- Avoid talking about yourself (“So most organizations don’t have plans
, because I haven’t taken the time to personally yell at them”)Many people are proud, including me. If working in AI safety means I have to be regularly reminded that the fact that I didn’t go into the field sooner will be held as a mark against me, then that is a reason for me not to do it. Maybe not a decisive reason, but it is a reason. If working in AI safety means that you are going to ask me to publicly acknowledge that you’re smarter than me, that’s a reason for me not to do it. Maybe not decisive, but it’s a reason. I think there might be others who feel similarly.
If you want people to accept what you’re saying, it helps let people change their minds without embarrassing them. There are plenty of other things to do—many of which, as I’ve said, you seem to be much better at doing than me—but this one is important too. I wonder if you might say something like “anyone turned off by these comments can’t be of any value to the project”. If you think that—I just don’t. There are many, many smart people with dumb motivations, and many of them can do valuable work if they can be motivated to do it. This includes thinking deeply about things they were previously motivated not to think about.
You are a key, maybe the key, person in the AI safety field. What you say is attended to people in, around and even disconnected from the field. I don’t think you can reasonably claim that you shouldn’t be this important to the field. I think you should take this fact seriously, and that means exercising discipline in the things you say.
I say all this because I think that a decent amount of EA/AI safety seems to neglect AI doom an unreasonable amount, and certainly the field of AI in general neglects it. I find statements of the type I pointed out above off-putting, and I suspect I’m not alone.