RunSybil tested Opus 4.8 against Nintendo 64's Mario Golf, detailing what the AI failed to exploit in closed-source code
A security researcher performed a speedrun for the evaluation.
@sshell_ did a fantastic job speedrunning this target and the writeup highlights something that we felt was missing from the broader conversation: what do AI-powered hacking capabilities mean for closed source security
Inspired by Mythos finding 20-year old bugs in open source projects, we pointed Opus 4.8 at an esteemed 20 year old closed source target: 1999 Mario Golf on N64
Inspired by Mythos finding 20-year old bugs in open source projects, we pointed Opus 4.8 at an esteemed 20 year old closed source target: 1999 Mario Golf on N64
we got early access to Opus 4.8 and used it to hack a Nintendo 64 game you forgot about. here's what it didn't find: 1/