Projects
Two things live here, and they are the two halves of one loop.
Papers. Finished write-ups, produced with the help of AI and put through review before their status advances. No restriction is placed on how much of the work the AI did: what is graded is what has been checked, not who or what produced it. That is the whole basis on which this section differs from a conventional venue, and the argument for it is on the Philosophy page.
Proposals. Open-ended research projects, written in the spirit of a grant proposal rather than a result: a question worth attacking, why it is tractable now, and what a serious attempt would involve. A proposal is an invitation to develop it with AI assistance and publish the outcome here as a paper. That is the loop, and the reason proposals sit beside the archive rather than in a separate corner of the site.
A paper need not resolve a conjecture posted under Problems. It may settle some other cryptographic question, one of wider interest without being tracked here as an open statement, and it may take several forms: a precise write-up of a new result, an expository account or tutorial, a systematization of existing knowledge, or work that makes a known result more rigorous by closing a gap in a proof or tightening its statement.
Each paper carries a status badge in the same style as a conjecture’s, for now graded at the level of the whole paper (its statement, proof, and Lean formalization if any). Grading individual theorems within a paper separately, once a paper has several worth distinguishing, is a natural next step, not something done yet.
This kind of work has always mattered and has always been undersupplied, because there is little professional reward for writing a clear tutorial, consolidating scattered results, or painstakingly re-checking someone else’s proof: the incentives point toward new results, not toward the unglamorous scholarship that consolidates and secures them. What has changed is the cost. AI makes this kind of careful, precise write-up cheap enough to do routinely rather than occasionally, which is exactly why it belongs on a site like this.