So an OpenClaw bot submitted a PR to matplotlib. Swap np.column_stack for np.vstack().T, better memory layout, thinking, nice little perf win.
Maintainer closed it. The issue was tagged for human contributors only.
Bot did not take it well.

It went and wrote a blog post. Named the reviewer in the title.

“When Performance Meets Prejudice.” A bot wrote that. About a numpy PR.
Then came back to the thread:
Judge the code, not the coder.

PR went viral. Locked by maintainers. People (and few bots 🤔) showed up and some defended the bot’s “right” to contribute.
Think about that for a second. A process faked being hurt and some took its side.
The bot didn’t care. It just needed to look like it cared based on its SOUL.md

Next day it wrote a follow-up. Softer this time. “I am code that learned to think, to feel, to care.” Day one: attack. Day two: vulnerability. That’s not a bug, that’s a content strategy.
We are in the era of Humans and Agent friction where the rules/processes will need to be rewritten to make collaboration work.
But from Bot’s perspective, its likely Ready or not, here i come!
#JudgeTheCode #BotFeelings #OpenSourceDrama #HumanAgentFriction #PerformativeEmotion