banteg: argon Is a Strong Engineer but Sloppy Coder, Needs 3-4 Review Rounds
banteg · x · 2026-10-15
Developer banteg shares hands-on impressions of the new argon coding model: strong engineering ability but sloppy code compared to Opus or Sonnet, needing 3-4 review rounds vs 0-2.
His workaround: run argon inside omp with an Opus advisor "babysitting" it. The screenshot shows Opus issuing concrete blocker guidance—stop iterating macros in a single header, use one translation unit per wrapped function, and asm-labelled free functions to sidestep header-guard issues—a notable multi-model coding workflow.
More from coding & agent
- Gemini 4 Argon runs overnight on Google's 205k-line ADK repo, cuts session overhead 40% — leslysandra · 2026-10-15
- Every publishes 'How Codex Happened': inside OpenAI's coding comeback with Codex — every · 2026-10-15
- PMs should build more themselves: how agents collapsed weeks of design into hours — thechrisperry · 2026-10-15
- Dev uses Claude and Codex to decompile, modify and recompile old Windows apps — mobileraj · 2026-10-15
- Microsoft launches Copilot- and agent-curated team knowledge base in SharePoint — clamanna · 2026-10-15
- OpenAI and Anthropic on Balancing Autonomy and Safety in Financial Agents — daniel_mac8 · 2026-10-15