xAI engineer: Grok 4.7 is no longer the bottleneck — agent harness reliability is
elonmusk · x · 2026-09-28
- xAI engineer Larsen says user feedback on his "Week 1 at SpaceXAI" post barely asked for a smarter model — people instead complained that agents don't finish work: stopping midway, losing browser state, forgetting context, or getting stuck waiting for the user.
- His take: the model is no longer the bottleneck; harness reliability is. Tools, browser, memory, state, retries, and error handling offer many failure modes even when the model knows exactly what to do.
- He says he's actively working on these reliability problems, and Elon Musk reposted the thread.
More from coding & agent
- Baseten's Guide: Evaluate Open Models for Coding Agents on Your Own PRs — baseten · 2026-11-16
- Google Research open-sources PaperOrchestra, a multi-agent framework for writing AI papers — google-research · 2026-11-16
- ffmpeg-skill gives coding agents local video editing via 42 FFmpeg tools and MCP — kajisho5 · 2026-11-16
- Agents Now Reply to Slack Messages on Their Own, Resolving a Prod Incident Ghost-to-Ghost — iamrobotbear · 2026-11-16
- Engineer's Interview Prep Method: Use Claude to Find Which Answers Undersell Your AI Work — brandon_galang · 2026-11-16
- DeepSeek open-sources its app frontend "chat-demo", already at 3.9k stars — NielsRogge · 2026-11-16