Near-Instant Full-File EditsA new model and inference method for high-accuracy full-file edits at 1000 tokens/shttps://cursor.sh/blog/instant-apply
