But is it materially different? Or is it a difference in quantitative increase of likelihood of the exploit?

The thesis "everything hackable" will be hacked, and with the advent of hacking LLMs, the boundary between criminal hacking and civil hacking is being blurred, it's no longer attributable to a criminal intent, hacking commands are just something that happens through no criminal intent.

Practically, the hackers, whether human or syntehtic, are a catalyst, they accelerate the rate at which a bug that causes damage will manifest, and perhaps the magnitude of the manifestation, but the essence of the negative manifestation of that bug is the same. Whether it was triggered by a solar ray flipbit or by a rogue LLM or by a russian hacker, the essence of the issue is the same.

▲

namenotrequired 20 hours ago | parent [-]

Yes? “one of us made a mistake” vs “someone is attacking us” is a difference that people care about.

▲

croon 18 hours ago | parent [-]

Not only care about; a mistake is unlikely to be repeated, a motivated attack is very likely to be attempted again.

▲

lostlogin 18 hours ago | parent [-]

> a mistake is unlikely to be repeated

If it was a mistake, it happened at many facilities. That seems like a repeated mistake to me.

	▲	croon 13 hours ago \| parent [-]
		That's not what logically follows from my statement. It's a conclusion you can make if you believe my premises are false though. It could also be a third possibility of a bug/issue affecting many of these specific batches.