For those wondering this cryptic post is a copy/paste of a Reddit post, stating that over 135K OpenClaw agents are connected to the internet, reminding how fundamentally unsecure OpenClaw is. The bad idea referenced in the post title:

The (bad) idea: Breach all machines and transform them in the first ai-assisted autonomous tor-enabled self-growing botnet Target OpenAI infrastructure Sit back and laugh