Our problems

Summary

The post lays out problems Cursor is working on in AI-assisted coding, from selecting relevant context in large codebases to making edits and finding bugs. It proposes collecting 500k tokens from editor sources and using a custom, fast reranker to reduce that context to 8k tokens, while reporting 1.4 billion vectors and 150,000 codebases indexed as of October 12, 2023. Other planned capabilities include a constrained few-step agent that searches, writes, and runs code with feedback, passive and active bug finding, and larger edits presented in real time. Future ideas include time-warp predictions, code understanding in model weights, reader and pseudo-code modes, and automatic stack-trace fixes. The list is explicitly not a final roadmap because new ideas and priorities change, and the post also invites interested people to join the company.

Takeaways

1. 1

   For each request, Cursor plans to gather 500k tokens from sources including open files, similar code, symbolically connected classes, lint outputs, execution traces, git history, and documentation, then rerank them down to the most relevant 8k tokens.

2. 2

   The proposed constrained agent would search, write, and run code while periodically consulting the user, initially operating on folders containing a few hundred thousand tokens before potentially scaling to entire codebases.

3. 3

   As of October 12, 2023, Cursor had 1.4 billion vectors and 150,000 indexed codebases; its Rust-based Merkle-tree syncing engine was already built, but a custom indexing system would probably be needed soon.