I had an idea for how to solve the AI alignment problem, which I've been slowly writing up. At the end of 2024, if I still think there's something to my idea, I will do a poll of LessWrong or alignment researchers or something, and ask them how much they think my idea solves the alignment problem. The market will resolve to the mean answer.
If I don't get the writeup released before the end of June, I may push the resolution date.
If I lose belief in the idea, I may just resolve the market NO.
1,000
3.00@Pykess Because I'm already used to submitting to LessWrong.
Which peer-reviewed journal are you suggesting I should use, and why should I use it?
@tailcalled Solving the alignment problem would be one of the greatest discoveries in at least the past 50 years, but more likely 100 or more. Something like that would belong in Nature or Science.
@Pykess It's also a hard enough problem to solve that the priority should be getting the proposed solution in front of people for review ASAP, rather than dealing with the requirements of a journal. By all means publish it in a journal+conference+etc after that review/vetting, there will be plenty of opportunity.