Hi,
I've been programming various ESP32-based musical electronics—such as an electronic melodica and a GUI for controlling sound expanders—using various AIs (ChatGPT, Claude, Grok, Google Gemini) for about three years now. My experience is that while AI is very helpful, it fails completely on more complex projects; it generates a lot of nonsense when dealing with multiple source files and Makefiles, which ends up wasting a huge amount of time. So, I can certainly understand Dan's point. It's great for specific sub-tasks, but handing an entire project over to AI is a recipe for disaster.
Best regards
Werner
I've been programming various ESP32-based musical electronics—such as an electronic melodica and a GUI for controlling sound expanders—using various AIs (ChatGPT, Claude, Grok, Google Gemini) for about three years now. My experience is that while AI is very helpful, it fails completely on more complex projects; it generates a lot of nonsense when dealing with multiple source files and Makefiles, which ends up wasting a huge amount of time. So, I can certainly understand Dan's point. It's great for specific sub-tasks, but handing an entire project over to AI is a recipe for disaster.
Best regards
Werner
