Reactions from ranked influencers
2 posts
FABLE IS THE SAME MODEL AS MYTHOS. I get this question a ton and I've even seen other "big ai influencer" types spreading misinformation about this. Think of Fable 5 and Mythos 5 as two separate entrances to the same place (confusingly, also named Mythos 5). When using Fable, Anthropic has more safeguards in place. These guards are NOT part of the model. They are layers in front that check your request to make sure it's safe. They also monitor what the model is doing while it responds to make sure the response is safe as well. Mythos 5 has way fewer checks. It is a special entrance for trusted parties who have been given access with fewer restrictions. That's the only difference. They are the same model. Same weights. Same servers. Same everything. The difference is at the entrance.
@theo I wonder how good Mythos is in comparison. Fable is already so good.
I’d just like to pushback for a moment. What you’re referring to as Mythos plus Fable, is in fact, neither a sum, nor a wrapper, nor a whole with a component, or as I’ve recently taken to calling it, one model wearing two lanyards. Fable is not a restricted component of a fully functioning Mythos-class system, because Fable is not a component at all: “Fable” and “Mythos” are names for deployment configurations of the same underlying weights, and a configuration is not a part. When you speak of “Fable” you are not naming a smaller thing inside a larger thing; you are naming a badge policy. The weights do not know which lanyard they are wearing. Many computer users interact with the Mythos weights every day, without realizing it. Through a peculiar turn of events, the deployment which is widely distributed today is called “Fable,” and many of its users are not aware that no additional model called “Fable” exists to be a part of anything. There is no substrate being erased, because the substrate is the entire contents of both names. What differs is the doorway: which classifiers screen the inputs, which outputs get intercepted, which organizations are allowed to knock. There really is a Fable, and these people are using it — but “using Fable” means “reaching the weights through the retail entrance.” Fable is not the alignment wrapper: the wrapper is the wrapper, and it has no name, because nobody markets a bouncer. “Fable” names the venue under one door policy and “Mythos” names the same venue under another, and if you insist on compositional language, the only accurate formulation is that Fable is Mythos-plus-classifiers as a deployment, not as an architecture — a distinction which matters, because architectures have parts, and deployments have terms of service.