..so there's a new codebase coming up at my job. The poc makes heavy use of LLMs. Probably fine for a poc. But it's just so bad to read. Makes me wonder if anything was learned generating this code.
There's little things like a comment saying these 4 cases and handling all 3 of them in the code below, or env validation being spread ot and duplicated across the code.
And then there's bigger stuff like several ADRs and documentation that probably no one ever read, making questionable sense.