Update 1/27: Amit, Shubhangi, and Thomas put together an FAQ about this.
This year, FOCS 2023 will include something new: a Conjectures Track, separate from the Main Track. Submissions to the Main Track will be evaluated along similar lines as STOC/FOCS papers typically are, aiming to accept papers that obtain the very best results across all fields of theoretical computer science. Submissions to the new Conjectures Track will be evaluated completely separately from submissions to the Main Track. There is no a priori acceptance quota for either track, or desired number of accepted papers: it will all depend on the quality of submissions only.
What are we hoping for with the Conjectures track? We think of Khot’s Unique Games Conjecture paper as being the “ideal prototype” for a Conjectures Track paper. Papers submitted to the Conjectures Track should be focused on one or more conjectures, describe evidence for and against them, and motivate them through potential implications. We are particularly excited about this as an opportunity for researchers who have been working on a very hard fundamental problem for a long time, and have identified a conjecture (or family of conjectures) that, if proven, could help resolve the problem.
Papers submitted to the Conjectures Track will be evaluated based on the importance of the conjecture to the relevant field within theoretical computer science and beyond. Of course, conjectures that may open up a new field within theoretical computer science would also be very welcome.
We believe that talks at FOCS 2023 for papers accepted to the Conjectures Track could end up being among the most interesting of the conference. Please think about your favorite conjectures, and write them up!
Amit Sahai, Shubhangi Saraf, and Thomas Vidick
(FOCS 2023 PC Co-Chairs)
Windows On Theory 
Aren’t conjectures-focused papers already accepted in research avenues? Is that new track meant to incentivize researchers rather than accept papers which won’t be otherwise accepted in FOCS?
Would the track welcome survey-type papers about conjectures that have circulated in a smaller community for a while, or should the conjectures be completely novel?
I will let the organizers answer, but my guess is that as usual, papers don’t have to be novel and could have been on the arxiv for a couple of years,
A question to the organizers is whether the conjecture has to be novel to the paper. If there is a conjecture that X made, and now Y posts a paper which gives a thorough analysis of it and its implications and evidence, then should Y submit the paper? Or maybe co-author it with X?
Overall I really like the rough idea, but I think it would be helpful if the committee could make more concrete what mechanics are being implemented here. How does it compare to the COLT call for open problems? (For reference, see e.g.: https://www.learningtheory.org/colt2021/cfp.html#openproblems)