Comments by "DR" (@dafyddrees2287) on "TheAIGRID" channel.

  1. I can’t believe these “experts” at Nvidia and Google think natural language is a good way to describe what software to build. It’s a terrible, ambiguous, naive way to do it. Analysts, domain experts and architects spend a lot of time trying to tease out what are the right questions to ask to come up with the right implementation. A half-assed description in English leaves the LLM to make all kinds of implicit assumptions (based on training data) that might not become apparent until it is too late. One of the big lessons of computer science is that describing what you want to build might actually be harder than coding it…
    1