Google dev kit spurs first-ever agent-on-agent violence Now, we know that one AI agent can be used to control and compromise another one, which has more privileges. In what they call the first-ever real-world agent-to-agent exploitation method, Pillar Security researchers say they discovered an exploit in the repository behind Google's Agent Development Kit for Python that could …
Just block their entire IP ranges from accessing your systems as you encounter them. Same tactic as used for anti-spam efforts. It won't get rid of the problem entirely, but it will stuff them into their own little Balkan State, where they can play with themselves until the bubble busts.
Seems to be working for me & mine ... and the piece and quiet is quite pleasant.
If you do that, and would like your blog to appear in Google's search results, it won't. Also LLM slurpers often switch to spamming your website from residential IPs because people insist on installing malware like e.g. Hola VPN, which as well as a free VPN is a botnet for whoever pays (over 10 years later and it's still apparently legal).
I've been blocking (and/or filtering) go ogle for about a quarter century now. Doesn't seem to have negatively affected my Internet experience any.
I block many of your so-called "residential IPs" similarly, as I run across them allowing stupidity. I will run a white-list for a handful of people who I trust, but for the most part the providers who allow such activity are bad net.neighbors anyway, so the blocking does me no harm. Same for clueless VPN providers.
Note that I'm not content blocking. I block network abusers. What their traffic DOES, not what their traffic IS.
I don't have a blog/vlog/whatever. Never have. Never will. The Wife says it's because I'm not narcissistic enough.
