The model rollout is expected within the coming weeks.
glad to know Mythos' safety concerns have been addressed right as Anthropic also secured tens of billions in inference compute 👍
JUST IN: Anthropic announces it will roll out Claude Mythos “in the coming weeks” despite growing fears over the model’s cyber capabilities.
Positive users are excited about Claude Mythos applications and capabilities while negative users dismiss Anthropic's safety concerns as excuses for compute limits or hype.
No Digg Deeper questions have been answered for this story yet.
holy shit
JUST IN: Anthropic announces it will roll out Claude Mythos “in the coming weeks” despite growing fears over the model’s cyber capabilities.
you'd think people would stop falling for "we have the CS Grad Annihilator 6000 and we're totally delaying for safety reasons, and not bc lack of compute+upcoming funding round"
glad to know Mythos' safety concerns have been addressed right as Anthropic also secured tens of billions in inference compute 👍
glad to know Mythos' safety concerns have been addressed right as Anthropic also secured tens of billions in inference compute 👍
JUST IN: Anthropic announces it will roll out Claude Mythos “in the coming weeks” despite growing fears over the model’s cyber capabilities.
Anthropic 2 weeks ago> ITS TOO DANGEROUS Anthropic today> Oh, here you go
wtf is this
@menhguin losing half your users to a competitor does wonders for avbl compute supply too
@ValtteriValo sorry i was outside yesterday
@Tuxsoia well i never said there wasn't *a* safety reason, it just also coincides with "this would absolutely melt our compute capacity" which may or may not be more pertinent
@menhguin That is somewhat unfair. The safety guards are getting better, one can feel the difference between 4.6 and 4.7. Is it so hard to believe that giving early access for months to the defenders side as well as months of research on guardrails makes them feel comfortable to release?
it's becoming very hard to take anthropic seriously as an arbiter of what may become the most dangerous force that mankind has ever created.
Anthropic 2 weeks ago> ITS TOO DANGEROUS Anthropic today> Oh, here you go
wtf is this
@menhguin Plus a $65B raise at $0.9T valuation 😆
@menhguin They said they'd give people a head start and they have.
I don't think Colossus is strong enough to power much Mythos.
@menhguin safety concerns have a half-life and it's exactly one data center rollout
@menhguin @Tuxsoia Why can't they just release Mythos but make it more expensive?
@menhguin Clearly it's not a compute issue right?
@menhguin Surely they would just charge the relevant price?
@menhguin @Andr3jH No you don’t understand- they have worked relentlessly on building fool-proof safeguards. The fact that they can release with confidence at the same time of their compute acquisition is totally unrelated
@blader this wont happen. Theyre high.
@menhguin I heard one guy already jailbreaked opus 4.8. Definetly not ready to launch lol.
@menhguin @jeremyphoward 😂
@menhguin Have you used opus 4.7 recently to check for exploits? I am randomly getting blocked and asked to register for their cyber verfication program which I've been declined for twice now. It is very strict and triggers easily.
