Bringing the cybersecurity capabilities of Claude Mythos 5 to more defenders
1–10 of 56 posts
Re: Bringing the cybersecurity capabilities of Claude Mythos 5 to more defenders
#2Seems like a whole lot of nothing for the average user. They have really lost the plot.
Re: Bringing the cybersecurity capabilities of Claude Mythos 5 to more defenders
#3may have been rushed by hugging face being unable to use claude to debug or fix their breach, because there isnt detail on new guardrails put into place to protect against anthropic's initial concerns of wide distribution?
separately, it is interesting they are adding an oem type offer:
>We’re working with our cybersecurity technology and services partners to integrate Claude Mythos 5 into the products and services defenders already use to secure their software.
anyone know the shape of those contracts, e.g. pure tokens/usage or more of a traditional licensing or oem type structure?
Re: Bringing the cybersecurity capabilities of Claude Mythos 5 to more defenders
#4This week I found two issues in my company codebase. After finding them I told a claude session about one and asked for a quick proof of concept demo of the exploit. It refused, including refusing simple things in the same session afterwards.
Meanwhile same model in a new tab, say I need help creating a page that hits an endpoint with a special payload and it does the same things that were too dangerous in the previous tab...
Re: Bringing the cybersecurity capabilities of Claude Mythos 5 to more defenders
#5Seems like a whole lot of nothing for the average user. They have really lost the plot.
I dont understand. Not everything has to be for everyone.
Re: Bringing the cybersecurity capabilities of Claude Mythos 5 to more defenders
#6China will beat them to general availability.
Anthropic is overconfident, and it's going to bite them in the ass.
Dario deserves it for being so hypocritical, fear mongering, and ladder pulling. Fastest hero to villain arc in recent memory.
Re: Bringing the cybersecurity capabilities of Claude Mythos 5 to more defenders
#7For those who had access, how does it compare IRL with GLM 5.3 ? iirc both models are similar in terms of benchmarks ?
Re: Bringing the cybersecurity capabilities of Claude Mythos 5 to more defenders
#8This week I found two issues in my company codebase. After finding them I told a claude session about one and asked for a quick proof of concept demo of the exploit. It refused, including refusing simple things in the same session afterwards. Meanwhile same model in a new tab, say I need help creating a page that hits an endpoint with a special payload and it does the same things that were too dangerous in the previo…
Just use an EU- or US-hosted Asian Opensource model. They will spit out everything.
Re: Bringing the cybersecurity capabilities of Claude Mythos 5 to more defenders
#9> Bringing the cybersecurity capabilities of Claude Mythos 5 to more defenders
This dance is quite annoying. When Anthropic releases a paid model, the user should control what it does and doesn't do. The other day I had a security incident that required urgent response. ChatGPT and Claude were utterly useless (I quickly attempted to get access to the former's advanced security capabilities but was met with a form I could complete - far too slow for a time-sensitive problem like a security incident!). So I used grok and it helped! I am told Kimi also helps in such cases although haven't tried it yet.
Re: Bringing the cybersecurity capabilities of Claude Mythos 5 to more defenders
#10may have been rushed by hugging face being unable to use claude to debug or fix their breach, because there isnt detail on new guardrails put into place to protect against anthropic's initial concerns of wide distribution? separately, it is interesting they are adding an oem type offer: >We’re working with our cybersecurity technology and services partners to integrate Claude Mythos 5 into the products and services d…
Hugging Face was able to sign up for Glasswing but they just didn't.