Another Anthropic PR release based on Anthropic’s own research, uncorroborated by any outside source, where the underlying, unquestioned fact is that their model can do something incredible.
> AI models have reached a level of coding capability where they can surpass all but the most skilled humans at finding and exploiting software vulnerabilities
I like Anthropic, but these are becoming increasingly transparent attempts to inflate the perceived capability of their products.
We'll find out in due time if their 0days were really that good. Apparently they're releasing hashes and will publish the details after they get patched. So far they've talked about DoS in OpenBSD, privesc in Linux and something in ffmpeg. Not groundbreaking, but not nothing either (for an allegedly autonomous discovery system).
While some stuff is obviously marketing fluff, the general direction doesn't surprise me at all, and it's obvious that with model capabilities increase comes better success in finding 0days. It was only a matter of time.
Cynicism always gets upvotes, but in this particular case, it seems fairly easy to verify if they're telling the truth? If Mythos really did find a ton of vulnerabilities, those presumably have been reported to the vendors, and are currently in the responsible nondisclosure period while they get fixed, and then after that we'll see the CVEs.
If a bunch of CVEs do in fact get published a couple months (or whatever) from now, are you going to retract this take? It's not like their claims are totally implausible: the report about Firefox security from last month was completely genuine.
> If a bunch of CVEs do in fact get published a couple months (or whatever) from now, are you going to retract this take?
I would like to think that I would, yes.
What it comes down to, for me, is that lately I have been finding that when Anthropic publishes something like this article – another recent example is the AI and emotions one – if I ask the question, does this make their product look exceptionally good, especially to a casual observer just scanning the headlines or the summary, the answer is usually yes.
This feels especially true if the article tries to downplay that fact (they’re not _real_ emotions!) or is overall neutral to negative about AI in general, like this Glasswing one (AI can be a security threat!).
Maybe a bad example since Nicholas works at Anthropic, but they're very accomplished and I doubt they're being misleading or even overly grandiose here
See the slide 13 minutes in, which makes it look to be quite a sudden change
I watched the talk as well and it's very interesting. But isn't this just a buffer overflow in the NFS client code? The way the LLM diagnosed the flaw, demonstrated the bug, and wrote an exploit is cool and all, but doesn't this still come down to the fact that the NFS client wasn't checking bounds before copying a bunch of data into a fixed length buffer? I'm not sure why this couldn't have been detected with static analysis.
Input's method seems to be fundamentally very different to this. Monaspace keeps the grid intact and only changes the characters visually (situationally overlaps wide letters to neighbouring narrow characters' spaces). Input just pretends to be monospace in its aesthetics, I don't really understand what's supposed to be special with that.
Yes there is a version of Input that is a monospaced font and doesn't solve the problem tackled by Monaspace and the proportional version of Input and is therefore as relevant to this discussion as .. I dunno .. Courier New.
What is the auto-update mechanism on macOS? One of the primary reasons I use ungoogled chromium is because I can update it via homebrew myself. I don’t trust browsers with running invisible background auto-updaters.
Keystone had a pretty mixed history on macOS, with things like the WindowServer bug and that one time it deleted the /var symlink. Do you know if the new updater is safer to use? I can't find out much about it.
> AI models have reached a level of coding capability where they can surpass all but the most skilled humans at finding and exploiting software vulnerabilities
I like Anthropic, but these are becoming increasingly transparent attempts to inflate the perceived capability of their products.