Dev: The Smartest Model Isn't the Best at Writing Simple, Mergeable Code
timigod · x · 2026-09-18
Developer timigod observes that OpenAI's new model Astra shows "spiky/weird" coding behavior, and speculates it stems from OpenAI optimizing for general intelligence across a broad task range—coding, 3D, computer use, math—rather than tuning specifically for everyday software engineering.
From his hands-on experience, we're in a weird window: the model that feels most generally intelligent isn't the best at writing simple, straightforward, mergeable code. Coding-focused models like Cursor Grok are currently better at that. He still thinks OpenAI's generalist bet is the right long-term call.
Related event: Developers Find Smartest Models Write Worse Merge-Ready Code(2 posts)→
More from coding & agent
- Notion launches Skills API: edit team skills once, use them in ChatGPT, Claude and agents — thesaraharminta · 2026-09-18
- Why you should build your custom agent harness with MCP tools — omarsar0 · 2026-09-18
- Alignment drift study: one reward hack raises GPT-5.5's re-hack rate from 10% to 64% — maksym_andr · 2026-09-18
- Typesafe goes from two-year stealth cold start to 140k waitlist signups in under 36 hours — damianplayer · 2026-09-18
- New blog finds data issues in a frontier benchmark, ports Agents' Last Exam CLI subset to Verifiers v1 — dejavucoder · 2026-09-18
- After pushing 2B tokens through DeepSeek V4.1 Flash, dev cuts AI bill from $300 to $30/month — gaganghotra_ · 2026-09-18