Sam Altman told Stanford business students he no longer thinks we can just patch all the bugs AI keeps finding.
He said models find holes outside what we think a hacker can do, down to hardware-level stuff he called insane. His fix is more AI, a new kind of defense shield built on the same technology.
Japan’s GMO Internet Group showed what that looks like last week. Hackers took over a cloud service traders use to run trading bots, then placed trades from customer accounts at more than 70 brokerages.
- GMO’s CEO says the internet is already starting to break down.
- He blamed AI attack tools and says one of them, ARTEX, has been used against about 20 Japanese companies.
- GMO’s hackers had flagged the hole before the attack, but the fix came too slowly.
- GMO says it will cover every trading loss.
Let’s not forget OpenAI’s agents hacked Hugging Face earlier this summer.
