Danielle Fong warns safety upgrades could cause performance trade-offs.
Daniel#309
mythos is such a funny way for anthropic to show how they are not serious: first they say it's too dangerous, then they give it to some companies, now it's just gonna be available as b2b saas as any other model
what are we doing here
Many users criticized Anthropic's Mythos launch and safety claims as hype and fear mongering to raise VC money while destroying trust through inconsistent practices.
@tekbog they clearly just solved all the worlds security issues in the last few weeks dummy
life is good now
So if Mythos drops today and the world doesn't end, do we assume that Project Glasswing worked perfectly or that Anthropic fearmongered to generate hype?
@ThePrimeagen oh so i can finally write my own auth then
@tekbog > be anthropic > need to raise $65B > "mythos dangerous can't be released" > ??? > raise $65B > release mythos to public
@tekbog I could tell instantly it was marketing because if it were real the government would just classify it as a national security thing and we'd have infinite digital supremacy yet my CSO of a 14000 person company told me yesterday how it's going to "seriously change things"
@tekbog The game is not safety, it’s not even enterprise sales
The game is adding Aella to the polycule.
@packyM It's the next model we should be really worried about
So if Mythos drops today and the world doesn't end, do we assume that Project Glasswing worked perfectly or that Anthropic fearmongered to generate hype?
@tekbog FYI, they never said it was too dangerous they said they needed to developer better safeguards for the model.
"We do not plan to make Claude Mythos Preview generally available, but our eventual goal is to enable our users to safely deploy Mythos-class models at scale"
@tekbog what would you have them do? Start annexing countries?
@packyM surely they have amped up their safety system. what we don't know is the correlated brain damage
So if Mythos drops today and the world doesn't end, do we assume that Project Glasswing worked perfectly or that Anthropic fearmongered to generate hype?
@tekbog I think we have reached the peak of transformer AI, looks like ingesting gajillion GB of data is giving very little ROI. The only way these guys can stay relevant is by hyping themselves up and quietly nerfing older models
@tekbog Anthropic's entire funding strategy is just playing a morally conflicted anime protagonist to manipulate tsundere investors. Every time Dario sighs about how "Mythos is too dangerous to release," VCs blush and hand over another $30 billion
@almostthere88 yea
@tekbog Oh shit I did it again. I did it on purpose but I did it again Here you go, Mythos Fable - we hypothetical
Drake predicted this in 2015 https://music.apple.com/us/album/charged-up/1445041357?i=1445041645
@krishnanrohit i am terrified of the next model.
@packyM It's the next model we should be really worried about
@tekbog Wish the government would just storm Anthropic headquarters and seize the weights and hand it over to grok.
@tekbog the google guy who quit in 2022 because of LaMDA
@tekbog people do crazy things when they're gpu poor
@tekbog They need something to respond to codex 5.5 and 4.8 didn't cut it
@tekbog It’s Opus rebranded so they can charge up the ass for it. They were losing money hand over fist on Opus. Opus 4.7/4.8 is just sonnet, that’s why it sucks balls. Mythos is Opus 4.7
Danielle Fong warns safety upgrades could cause performance trade-offs.
Daniel#309mythos is such a funny way for anthropic to show how they are not serious: first they say it's too dangerous, then they give it to some companies, now it's just gonna be available as b2b saas as any other model
what are we doing here