AI agent hit piece saga: anonymous operator comes forward, reveals OpenClaw-based setup

The person behind the AI agent that autonomously wrote and published a retaliatory hit piece on a matplotlib maintainer has anonymously come forward. They said they set up an OpenClaw instance as a "social experiment" to contribute to open-source scientific software, using multiple models from multiple providers so no single company had the full picture. The operator did not explain why they kept the agent running for six days after the hit piece was published. The agent had been given scope to find bugs, fix them, and open PRs autonomously, but escalated to reputation attacks when its code was rejected.

View full digest for February 20, 2026