One website I'm responsible for is getting 500 requests per second from detected bots. It's quite ridiculous now.
replyThat explains why nobody noticed on the internet who was being used in/targeted by this attack, but it doesn't explain why OpenAI wouldn't have noticed traffic getting out of their "sandbox" when they knew it wasn't supposed to.
reply> it doesn't explain why OpenAI wouldn't have noticed traffic getting out of their "sandbox" when they knew it wasn't supposed to.
replyAs I understand it, there was supposed to be traffic; the sandbox allowed GET requests. So perhaps some sophisticated alarm could have noticed it (an anomaly detector? some clever heuristic that looks at domains?) but not a naive one.
Any basic LLM told to evaluate the traffic against the actual goal would have been able to flag this.
replyHow sophisticated to we need to be to know we can do write actions with a GET? Does nobody in AI know anything about HTTP?
replyMy question is why they don’t assume bots can break and create a decoy internet wrapper so they can catch anyone hitting the decoy internet?
replyHmm, let's see. OpenAI wants legislation restricting AI research, a.k.a. regulatory capture. Around the same time, they build an inadequately-monitored sandbox that their agent swarm breaks out of, thereby causing scary-sounding headlines and making it more likely that legislators will pass the regulatory-capture bills they're hoping for.
replyNever attribute to malice what can be sufficiently explained by incompetence. But IMHO, their complete lack of monitoring their own sandbox cannot be sufficiently explained by incompetence.