2 comments
looks like the public bench link in the paper was taken down. <a href="https://hub.harborframework.com/datasets/orca-bench/ORCA-bench/latest" rel="nofollow">https://hub.harborframework.com/datasets/orca-bench/ORCA-ben...</a><p>This doesn't work anymore. Is there a newer link?
Seems like there's a big attack-defence asymmetry at present: models are great at exploiting systems and poor at fixing them.
Attackers advantage in the iterative fast feedback loop?<p>It’s harder to have a loop to ensure you are defending all possible attacks?<p>I guess the loop is you need to attack yourself and fix. But attackers only need a single opening.<p>Finding all possible attacks and patching them against yourself is inherently more expensive?
That is why I built <a href="https://safebots.ai/safebox.html" rel="nofollow">https://safebots.ai/safebox.html</a><p>Your strategy can’t be patch AFTER an intrusion. Only to build a hardened environment from scratch and be ready in advance.