AGAINST VIBES: WHEN IS A GENERATIVE MODEL USEFUL

Summary from the TLDR Dev newsletter

Criticizing the "vibes"-based discourse around AI models, the author proposes a scientific model for determining their true usefulness based on the relative costs of encoding and verifying artifacts, and whether the task prioritizes the final artifact over the creation process.