Is the Claude Code harness improving faster than its models?
charango · reddit · 2026-09-16
A Reddit post links a Substack essay arguing that Claude Code's harness has visibly improved over recent weeks, even as the underlying models may feel flat or slightly worse. The author speculates this could mean more resources thrown at the tool, an internal new model being tested, or access to Astra. Commenters echo the observation and debate whether perceived model decay is real or psychosomatic.
More from coding & agent
- Microsoft's MemSteer turns unlabeled user histories into agent memory supervision — MicrosoftResearch · 2026-11-02
- She fired herself from her own Instagram: AI agent runs the account, views up 82.3% — heyshrutimishra · 2026-11-02
- Sourcegraph: AI writes code faster than teams can review it, driving more outages — AI Engineer · 2026-11-02
- Tsinghua & Microsoft's HarnessMM cuts agent token use by up to 87.93% while raising success rate — rohanpaul_ai · 2026-11-02
- GitHub launches Agent HQ: Anthropic, OpenAI, Google coding agents in Copilot — emmanuelvivier · 2026-11-02
- DeepSeek open-sources Super-Agent, a coding agent that reads team chats to minimize human input — emmanuelvivier · 2026-11-02