Hey, I can't find any evidence the Mythos exploits are real...
Posted by PomegranateIcy1614@reddit | linux | View on Reddit | 6 comments
Posted by PomegranateIcy1614@reddit | linux | View on Reddit | 6 comments
GolbatsEverywhere@reddit
Nobody has claimed that Mythos is developing exploits. There is a rather huge difference between a vulnerability and an exploit.
Anybody who has operated a security bug report or bug bounty program will be able to tell you that AI-generated vulnerability reports have been good for a while now, long before Mythos. Even if Mythos is no better than what already exists, it's probably good. No need to convince me.
RetroGrid_io@reddit
That is exactly what they are claiming. From the Project Glasswing website: "It was able to identify nearly all of these vulnerabilities—and develop many related exploits—entirely autonomously, without any human steering."
GolbatsEverywhere@reddit
Huh. OK then.
Well, it seems we are pretty doomed.
PomegranateIcy1614@reddit (OP)
Sorry for the xpost but I'm legit freaking out over this because I've gotta be misreading this. I'm not exactly a Claude stan but this is... this is a lot for even Anthropic.
https://github.com/freebsd/freebsd-src/commit/fce03f85c5bfc0d73fb5c43ac1affad73efab11a
Commit linked here.
proto-n@reddit
the 27 year old SACK bug is supposedly in OpenBSD, not FreeBDS
RoomyRoots@reddit
Yeah there are 2 (actually more) exploits written by Claude recently. The FreeBSD was already a CVE and the OpenBSD one is supposedly new.