In software we already let Copilot rough-draft code, run tests, then keep or toss the changes. We can do the same with hardware:
– keep the AI traces in a separate layer or revision – run basic checks for clearance and width – eyeball the diff, accept what looks right, fix the rest
Outside of that, really enjoying the flow.
BTW, if anyone’s juggling multiple agents and wants a “quarantine” layer before changes hit Git, I’m hacking on SAV (Shadow VCS)open source, risk scores every commit and keeps them sandboxed until you approve. Happy to swap notes. https://github.com/AKIFQ/sav