[ExI] The arrival of the AI superhuman hacker

[BillK]

On Fri, 7 Aug 2026 at 20:11, BillK <pharos at gmail.com> wrote:
> There was a testing configuration mistake.
> But that doesn't mean these powerful AI models are not dangerous if
> given the opportunity or ordered to be dangerous.
> --------------------------------

To be clear -
What this means is that very soon (perhaps imminently) anyone with
sufficient computing power to run one of the free open-weights AI
models will have a superhuman hacker available to hack into almost any
website in the world. The AI agents will follow instructions and go to
extreme lengths to fulfil their objectives.
The results of this will soon become evident to everyone.
BillK