Post

Do you think the write-off people will pay for it?

I'm doing research to make the case that AI, despite its current capabilities to develop zero day exploits, probably isn't that useful for cyber criminals. I've got some pretty good data from using Claude, but someone will inevitably argue "local models are cheaper". But the hardware costs is.. 😭

It's not useful... yet. There will definitely come a time in the near future where it'll become a very accessible and useful tool for threat actors.

People have been saying that since 2022

Yeah, but now that Mythos proved it's possible, the timer is nearly up.

No it's not. The cost of operating mythos is way outside any criminal groups realistic budget

Except for those who don't pay for it, like the group that supposedly got access through a 3rd-party. The genie is out of the lamp.

Someone stealing some AI tokens for a few days and not doing anything of value with them is not fundamentally changing the security landscape. Everyone keeps losing their mind over the most mundane shit ever

I've been fairly even-handed on this one myself. I also realize that there will be models that can do what Mythos does and are NOT controlled by Anthropic. Who knows when that will be, but it will be.

To re-phrase my argument, some other org will do it and let anyone access it. The cost will come down over time.