💾 Archived View for idiomdrottning.org › butlerian-jihad captured on 2023-03-20 at 17:58:36. Gemini links have been rewritten to link to archived content

-=-=-=-=-=-=-

Butlerian Jihad

mnl wrote:

Most importantly, having a wide array of tools at your disposal is what allows you to be pragmatic. And pragmatically, large language models represent the biggest paradigm shift in programming that I have personally experienced. It feels an order of magnitude more life-changing than discovering Common Lisp, and I’m only 3 months into using these things intensely.

(And a few weeks back we had Tom Scott’s embarrassing and self-shaming video.)

My experience with the current generation (ChatGPT), for programming specifically, is that:

it suggests impossible things that can not ever be made to work even with tweaks
sends you down a rabbit hole of wrongness when what you would’ve needed instead was a blank slate and a clear perspective
it lies and says that it has tested things (even giving the specific version of the compiler it’s supposed to “work” on) without having done so

Now I don’t wanna base my anti-LLM sentiment on “the current generation doesn’t work very well” so I’m being very careful and deliberate in saying how those issues are specific to that one version.

(More generally, code is law and I don’t wanna be ruled by an ouroborus of law-generated law.)

But that said, my three issues with the present version are pragmatically & currently speaking enough of a showstopper for me to nope out for now.

Clogs in the cogs! is my rallying cry.