Three prevalent themes in the discussion
| Theme | Supporting quotes |
|---|---|
| 1. AI‑assisted PCB design shows promise for simple/hobbyist work but struggles with complex, high‑performance boards | “I used claude and gemini to edit/debug kicad schematics. And it was pretty good.” – tadasv “Copperhead isn’t just a wrapper around claude or gpt. it combines model intelligence with its own hardware IR, deterministic engines and continuous verification.” – animeshchouhan “I made 2 PCBs with Claude… both work. I’m pretty impressed, especially because I’ve never done a PCB myself.” – ecesena |
| 2. Skepticism about replacing skilled hardware engineers; NP‑hard nature and gatekeeping attitudes limit full automation | “The one thing I will say, is that my experience of hardware engineers is very gatekeepy.” – Aurornis “NP-hard problems are still NP-hard even with magical LLM pixie sprinkles.” – Joel_Mckay “Auto‑routers have been around far longer than even smartphones. Somewhat functional in the trivial problem domains, and always useless where design choices mattered.” – Joel_Mckay |
| 3. Emphasis on end‑to‑end toolchain integration (KiCad IR, DFM, manufacturing, benchmarks) to close the loop from design to assembled board | “we’re building tool plane integrations around the same. stay tuned!” – animeshchouhan “we’re also building a design house around it to close the loop with fabs, handle DFM questions and eventually let you hit ‘order’ and receive fully assembled boards!” – animeshchouhan “we’re developing https://copperbench.org - an open‑source benchmark for evaluating ai agents on real, verifiable hardware design tasks.” – animeshchouhan |