Top 3 Themes from theDiscussion
1. AI Expectations vs. Reality
The community repeatedly points out the “magic‑or‑mistake” paradox: when an AI succeeds it’s hailed as magical, but failures are blamed on user error.
“It’s the annoying thing about AI. If it works, the AI is magic. If it doesn’t work, you’re using it wrong.” – makerofthings
2. Monolithic Code & Agent Behaviour
Many users observe that current agents prefer single‑file, monolithic outputs and struggle when code is split across many files, even though developers often split code for readability.
“Models favor monolithic, single‑file implementations that diverge sharply from human‑written code.” – pdp
“It was some time ago, so things might have improved? … I had trouble with Cursor & Co. once the file went over ~800 lines. It repeatedly failed to edit it, until I split it up.” – BurningPenguin
3. The Future of Programming & Job Roles
Speculation about a shift from traditional coding to higher‑level model‑driven development, with developers possibly becoming “model operators” rather than coders.
“How long until AI is not even writing code but producing machine code?” – keyle > “We won’t be developers, we won’t be devops, we’ll be rollmops!” – keyle (with a tongue‑in‑cheek nod to modelops)
These three themes capture the most common viewpoints across the thread, each backed by direct quotations from the participants.