Comment by zmmmmm
11 hours ago
there is hypothetically impossible and practically impossible
if AI tools get good enough that they can exploit ingrained hardware properties such as subtle timing and race conditions and swarms off them become cheap enough to deploy so that they can persistently exploit low probability events we could reach a state where I think it's actually practically impossible.
Then we will simply make better hardware too. There is nothing impossible about that, it's simply not something that has been prioritized amidst a culture of shortcut-taking. So far, this culture has been predicated on gambling that the one taking the shortcut will not be the one paying the consequences when the toll eventually comes due. If immediate exploitation of flaws becomes a certainty, the behavior will correct itself.