Post #4199606

Reading the OpenAI hugging face hack information… First, OpenAI should be prosecuted under the CFAA Second, the hacks are mostly boring. The fact that each of the environments allowed access to credentials and/or shelling out is just horrid security. The fact that a machine could brute force (try a bunch of different approaches in rapid succession) to obtain what should have been privileged info/access underscores that we need radically better security by default.