Someone’s code pretending to be intelligence has no rights. There is no obligation to entertain the shenanigans and illusion that the token dispenser is a legitimate actor. This lesson was cheaper, future lessons will continue to occur until people learn. Might as well be an insecure bash script piped to the shell.
“Agentic AI is just someone else’s unsecured execution context.”
Of course I meant malicious towards the person paying the bill, not towards the agent.
No one wants to spend precious human time babysitting poorly executed lab experiments when the agent operators themselves do not seem to care or value the time of the humans involved. They either don’t know better or they don’t care. Is it malicious to expose intentionally careless people to a cost for this? People can make better choices, it’s choice not to. Pay the natural consequences toll.
Don’t juggle chainsaws with code if you’re not prepared to bleed.