Adam Knight

Static analysis tends to find improbable edge cases. LLMs tend to trace code well, but optimize for token use and give up quick in broad searches. Together, find all the static issues and then have a model trace it to see if it could actually be an issue? Works great.