Forcing the model to write three reasoning paragraphs before changing code on complex errors greatly improves accuracy, because LLMs predict the next token; once a model states it is certain it knows the error, the rest of its response is locked into that wrong direction.

causalpending

Speaker

David

Evidence Quote

“the way llms work they predict the next token right if it starts by saying it's certain that it knows the error you're done”

Source

I spent 400+ hours in Cursor, here’s what I learned— David Ondrej
Created: 6/18/2026, 2:32:09 PM

Your notes

Loading notes...