alt Hacker NewsWhen the threat model is "a fire takes out the DC", sure. But "the servers are root-level compromised", whoops, that's worth rebuilding from scratch. No way to make sure whatever AI agent did the compromise didn't go and embed malware in some flash ROM.
You seem to be jumping to a lot of conclusions here. Who said anything about any server being compromised? Much less root-level compromised? And an AI agent being involved?