If it’s very good at finding security vulnerabilities, I would assume that the code it generates is much more hardened than anything your average developer can put out.
Mythos aside, frontier LLMs can already be used to find exploits at faster pace than humans alone. Whether that knowledge gets used to patch them or exploit them is dependent on the user. Cybersecurity has always been an arms race and LLMs are rapidly becoming powerful arms. Whether they like it or not LLM providers are now important dealers in that arms race. I appreciate Anthropic trying to give “good guys” a leg up (if that is indeed their real main motivation which I do find credible but not certain). But it’s still a scary world we’re entering and I doubt the fierce competition will leave all labs acting benevolently.
Dario is big on beating china, and no doubt he believes cyber security is how to do that. You can tell, but anthropic is sht at everything else. Nobody uses it for real research.
Yeah, I'd pretty pissed at my doctor for finding cancerous cells that probably wouldn't have been a problem for quite some time, either. Ignorance is bliss, security through obscurity, whatever.
You may joke, but this is a genuine issue in certain screening tests. e.g. most cancerous cells found in PSA prostate screening are so slow growing that they never cause any symptoms during a person's lifetime, so the treatment is almost always worse than the disease. It's similar for some sorts of thyroid and breast cancer tests. This is why a lot of countries are heavily reducing these sort of tests
The doctor analogy is more like you're grateful that your doctor found cancerous cells before they became a problem, but at the same time his other business is selling cigarettes.
That's exactly not what they're doing. They aren't creating operating system vulnerabilities. They're telling you about ones that already existed.
Well, in a slightly indirect manner. Claude is writing a ton of code, and therefore creating a lot of security vulnerabilities.
That's not what's happening here. This announcement is about the velocity with which Claude finds vulnerabilities in already-existing software.
5 replies →
If it’s very good at finding security vulnerabilities, I would assume that the code it generates is much more hardened than anything your average developer can put out.
Mythos aside, frontier LLMs can already be used to find exploits at faster pace than humans alone. Whether that knowledge gets used to patch them or exploit them is dependent on the user. Cybersecurity has always been an arms race and LLMs are rapidly becoming powerful arms. Whether they like it or not LLM providers are now important dealers in that arms race. I appreciate Anthropic trying to give “good guys” a leg up (if that is indeed their real main motivation which I do find credible but not certain). But it’s still a scary world we’re entering and I doubt the fierce competition will leave all labs acting benevolently.
Dario is big on beating china, and no doubt he believes cyber security is how to do that. You can tell, but anthropic is sht at everything else. Nobody uses it for real research.
Yeah, I'd pretty pissed at my doctor for finding cancerous cells that probably wouldn't have been a problem for quite some time, either. Ignorance is bliss, security through obscurity, whatever.
You may joke, but this is a genuine issue in certain screening tests. e.g. most cancerous cells found in PSA prostate screening are so slow growing that they never cause any symptoms during a person's lifetime, so the treatment is almost always worse than the disease. It's similar for some sorts of thyroid and breast cancer tests. This is why a lot of countries are heavily reducing these sort of tests
The doctor analogy is more like you're grateful that your doctor found cancerous cells before they became a problem, but at the same time his other business is selling cigarettes.