I fall for accidental engineering. Every engineer does from time to time. It goes something like this: After an hour long unsuccessful debug session I don't really know anymore how to continue. I start to use Google or my currently favorite LLM, which gives me (partially hallucinated) commands that "fix the smoking gun". I run them. After all, I don't really know what to do anymore, right? With some luck the problem disappears. Which of the actions taken did actually resolve the issue? And is it really resolved or did I just treat the symptoms? I don't know.

In the pre-LLM this happened occasionally. However, lately I overhear much more conversations that contain some form of "Claude gave me some commands, I ran them and it fixed the problem.". But did it really?

LLMs are powerful tools. But with great power comes great responsibility. Don't let your hard-earned engineering skills be watered down by the allure of a "quick fix" LLM solution. Do you really understand what solved the problem? If not, you are on the best track to become an "accidental engineer".