i’ve been watching/listening to this podcast episode and it’s really interesting what Simon does with LLMs. using it for fun experiments, proving that “vibe coding” isn’t all bad and can actually make pretty great results when harnessed with tests, etc.