[janalsncm]

The issue with LLM driven development is that it’s often as hard to verify the outputs of the model as it would’ve been to write it myself. It’s basically the programming equivalent of a Gish gallop.