It takes about 2 months of coding with an LLM to get reasonably good at it
The neophyte
- doesn't have a fine tuned system prompt
- has a wildly untuned feel for what the bot can do? Is it stupid? Is it a silicon omniscient god-in-a-box? Neophytes gravitate to the extremes
- needs a greenfield project (human written code has context window problems)
- needs tests or some other "objective function"
For humans, "worked once on my machine" is often good enough. For bot written code, it is not.
So the articles that says, "people wrote code with bots and it was bad" I'm still reading it as "people who didn't know how to write code with a bot demonstrated that they didn't know how to use a bot"
♡ 2 ↻ 1