A coding agent may still be a bad writer, and some say it has gotten worse
dbreunig · x · 2026-07-27
The thread asks whether a strong coding agent can also become a good writer — or even a merely passable one.
The reply argues that today’s most capable models often write worse than models from a year ago, pointing to the same over-dramatic flourishes, canned phrases, and choppy structure across systems like Fable and Opus. The poster suggests this looks like RL has flattened stylistic diversity.
Related event: Top models now write worse than a year ago, critic says(2 posts)→
More from coding & agent
- Tokyo Agent Forge hackathon shipped production-ready AI agents in one day — DavidBennett__ · 2026-07-27
- Long-running agents will need immutable event logs, this thread argues — sebpaquet · 2026-07-27
- Agentic Data Science in Practice: Agents Write Code but Answer Wrong Questions — hugobowne · 2026-07-27
- A VS Code extension adds Markdown-style highlighting to Alchemy string templates — samgoodwin89 · 2026-07-27
- Claude’s Stripe MCP connector is being called unusable after repeated disconnects — evielync · 2026-07-27
- Reddit user says Vecel’s Eve looks like “Next.js for agents” — richie9830 · 2026-07-27