GitHub user gupta-sahil01 added a comment to the discussion: Ambient "operator 
recommender" — predictive next-operator suggestions on the canvas

I'm still looking into the auto code completion of cursor. Right now, from what 
I've understood, is that cursor uses a small custom-trained model (not GPT or 
Claude) fine-tuned specifically for code completion, with a ~272k-token context 
window over your current file, open tabs, and recent edits.

There's a neat trick called "speculative edits", which means that they feed 
your existing code back to the model as draft tokens, so it only emits the bits 
that change, about a 13× speedup, which is what lets full-file rewrites feel 
instant.

GitHub link: 
https://github.com/apache/texera/discussions/5240#discussioncomment-17312370

----
This is an automatically sent email for [email protected].
To unsubscribe, please send an email to: [email protected]

Reply via email to