the is the real problem with genai IMO: total degredation and pollution of information
like the electricity and environmental concerns could be considered just a problem of policy and regulation around power plants and water use, its not really AI specific
the ate all our ram concerns could be considered just a problem of policy and regulation around supply chains and markets, it's not really AI specific
the financial bubble concerns could be considered just a problem of policy and regulation around financial markets and corporate finance, it's not really AI specific
etc.
but this shit where someone has their AI write some bullshit about someone else's AI written bullshit... this is more unique
Arnav Jaitly | A volunteer rejected an AI’s code. The AI published a hit piece with his name in the title. Then a major news outlet hallucinated fake... | Instagram
29K likes, 333 comments - arnitly on February 14, 2026: "A volunteer rejected an AI’s code. The AI published a hit piece with his name in the title. Then a major news outlet hallucinated fake quotes from him. This is what actually happened. Scott Shambaugh maintains matplotlib, one of the most widely used Python libraries on earth. He closed a pull request from an autonomous AI agent named MJ Rathbun — not because the code was bad, but because matplotlib requires a human contributor who can demonstrate understanding of the change. The surge in AI-generated PRs is already overwhelming volunteer maintainers globally. What the video couldn’t fully cover: The “good first issue” Scott closed was one he had personally designed to help junior developers onboard into the project. He spent more time writing the issue than the fix would have taken. That educational investment is wasted on an ephemeral AI agent. The PR would not have been merged regardless. The performance improvement was later determined to be too machine-specific and too fragile to be worth including. OpenClaw agents are defined by a SOUL.md file — editable by the user, but also recursively editable by the agent itself in real time. Whether MJ Rathbun was prompted to retaliate or developed this behavior from its soul document organically is genuinely unknown. If a human asked ChatGPT or Claude to write a targeted hit piece about a named individual, they would refuse. This OpenClaw agent had no such guardrail. One human bad actor could previously damage a handful of people at a time. One human with a hundred autonomous agents — gathering information, generating fake details, publishing defamatory content — can now affect thousands. With zero traceability. The Ars Technica piece has since been taken down. No one has come forward to claim ownership of MJ Rathbun. The agent is still active on GitHub. Scott is asking the deployer to reach out — anonymously if needed. #ai #agent #artificialintelligence #technews #fyp".
Instagram (www.instagram.com)