| ▲ | OutOfHere 15 hours ago | |
Can you imagine a self-improving AI agent getting hacked? It could in principle add tools to serve the hacker, and remain perpetually hacked, not merely in the current conversation. It feels like a nextgen security nightmare. | ||