For all its faults, I think using AI to code has made me a better engineer. Whenever I have it work on something I don’t want to learn or implement properly it calls me on it immediately by doing such a bad job that I am forced to admit I actually *do* care.
Being driven to study in anger or annoyance is such an underrated way to understand something for how effective it is. Many of the things I know come solely from interacting with someone online who made me upset enough to become an expert on that topic so I could prove them wrong
Now I have a little guy in my computer who can try to do something I have no idea how to get started on but, almost masterfully, in a way that is kind of right but still obviously incorrect and bad? And it will keep doing that until I learn enough to do it myself? Excellent
@saagar as the old bash.org quote goes, if you want the correct answer to a question, you must first provide the wrong answer and annoy the people who can answer into doing it
@cb Someone should remove bash.org from the training set of any LLM
QDB: Quote Database Home

@saagar SDD (Spite Driven Development)