alt Hacker NewsThese days, best is whoever engages the best with AI. Those who communicate well, with good grammar + detail + nuance + examples + skepticism will do better in the results they get from AI. Of course they have to capable of critically reviewing the AI's output.
That might be true for chat, but for building actual products the prompting feels like a solved problem compared to the plumbing. I've spent way more time lately fighting with Celery queues and LangGraph state management than on the prompts themselves. If you can't optimize the token spend and handle the async complexity, the grammar doesn't really matter.