Is that fair? it does exceed

our human context window and that's for sure and you know if you look at like from the like finding zero days and and these models certainly are finding zero day vulnerabilities um and it's not that it has some magic capability beyond compute capacity we have you know a data center you know, running at full capacity, figuring this out.

And that's now outputting, you know, discovering zero days in minutes and hours versus weeks, if at all.

And a good example of that is like, you know, A, the exploit market, you used to be able to, if you found something like an iOS, you'd be making a million bucks.

And that's because the level of effort was substantial.

But now when you're having, say, thousands and thousands of computers all trying to solve one problem and getting to that reward hacking, which we should dive into after.

we are now brute forcing um an application and finding those vulnerabilities and you know we're finding vulnerabilities that are 20 years old that just like we just didn't have the the human uh you know the appetite the bandwidth to keep digging to find those like every software is vulnerable given enough time to attack it and you know if you know rewind like looking at like we've made passwords so strong that it's going to take till the like the death of the universe to crack this password well it's been a paradigm shift now that we have the compute capacity to do a lot more parallel computing to try to get to that point so It is impressive.

I guess what I'm asking is, how do you see that this stuff is happening? And is it more kind of an after the fact? Or can you actually register this type of an attack in real time in a more traditional attack versus something that's sort of high velocity, high capability, like an AI model?