Hi Andrew, amazing work and done for the best of reasons!

I see you are replying in the thread, in your closing notes you wrote:

> I would not have been able to get past the initial stages without ChatGPT.

One of the coolest uses of LLMs for me is indeed exploring a problem domain where I am not a subject matter expert.

The problem is if it leads you down a dead end or hallucinates a solution. Then, at the edge of your knowledge, you actually have a very small foothold with which to challenge these chat replies.

Did you also face this on the more HW 'heavy' part of your project? Any tips on how to navigate terra incognita with LLMs?