Great piece. I completely align with the idea that token budgets should be prioritized to more productive employees or at least ones who aren't going to play around with it without a good value-added use case for testing.
Similarly on the note with LLM usage, I personally use LLMs as second-Google to validate but not much for generation of content. Vibe coding seems to be too much of a pie-in-the-sky idea to really scale. Sure you can create a little toy app or something, but then what?
Great piece. I completely align with the idea that token budgets should be prioritized to more productive employees or at least ones who aren't going to play around with it without a good value-added use case for testing.
Similarly on the note with LLM usage, I personally use LLMs as second-Google to validate but not much for generation of content. Vibe coding seems to be too much of a pie-in-the-sky idea to really scale. Sure you can create a little toy app or something, but then what?