Fable flags or refuses too many questions because of it's "policy"
Safety lobotomy by Anthropic
AI Judge changed title after evaluation, original title: "Abacus.AI CEO Bindu Reddy blames Anthropic's safety policies after Fable 5 Max blocks basic math and science queries"
Anthropic's new Fable 5 model, released as the public version of its Mythos-class system, routes queries on topics like phases of matter, cyclic groups, and PCR primer design to an older model when its classifiers flag potential risks in chemistry, biology, or cybersecurity, prompting Abacus.AI CEO Bindu Reddy to call out the approach as overly restrictive safety training.
Fable flags or refuses too many questions because of it's "policy"
Safety lobotomy by Anthropic
Anthropic applies broad detection across cybersecurity, biology, and chemistry prompts, swapping in Claude Opus 4.8 for flagged inputs and notifying users, which creates blocks on standard scientific queries even when no clear malicious intent appears.
API access starts at $10 per million input tokens and $50 per million output tokens with existing subscriptions covering use only through June 22 before credits kick in, and a narrower trusted program for life sciences is still in planning stages.
Many users slammed Fable and Anthropic models for excessive safety refusals blocking math, code, and science queries as overreach that renders them worthless, while a few praised unrestricted alternatives.
Fable can't even talk about PCR primer design without generating refusals. I can only infer than ant has zero biology people actually involved in prerelease testing.
This is the dark timeline.
Gatekeeping math knowledge is how we end up in a stone age nanny state.
Massive step backwards in terms of freedom.
"DANGEROUS MATH" a story in two acts.

@FirstName5000 the free market capitalism American math
lol I think the super high frequency of refusals may backfire a bit on Fable and anthropic
From early feedback
Why pay 2x if I have 70% probability of refusal?

@jeffcafe_ @MartinShkreli Price gouging is not capitalism its greed you gay simp loser retard

@MartinShkreli WWhat math did you use when you raised the price of Daraprim by 5,500% ???

@MartinShkreli Enough larping brother, can u even pass a complex analysis course

@MartinShkreli I now you're not asking Claude about this but yes some math is restricted for national security reasons

@jeffcafe_ @MartinShkreli Lol because he is a greedy selfish scumbag, duh!

@MartinShkreli No such thing as “simply a discussion about algebra & group theory”
You should know this. Shame! 4.8 forever!
(Same thing happened to me lol)

@MartinShkreli There is no such thing as math that is dangerous? Are you retarded?

@FirstName5000 @MartinShkreli Just don’t like commies. You’re just criticizing capitalism.

@MartinShkreli @FirstName5000 I don’t know why you still catch shit for that when it wasn’t what you were convicted for…

@MartinShkreli This is such obvious marketing bullshit I can’t believe people fall for it. It’s like Viagra commercials saying you might need to go to a doctor for a 6 hour boner. “This product is so powerful it might do something bad”! Such obvious bullshit.

@FirstName5000 @MartinShkreli Commies when the market prices goods and services without their input.

@FirstName5000 @jeffcafe_ is there a price for medicine that you would set? what would you have charged?

@colombiagrad this is what AI is for!
@ebarschkis

@FirstName5000 @MartinShkreli Ok, but isn’t him getting convicted of a Ponzi the better demonstration of that than raising the price on something in a way that was perfectly legal? Lol

@MartinShkreli We need to start a Math Lobby before it gets banned Ideas have always been the most dangerous thing, word is getting out

It would be more honest of them just to say "we refuse all biology questions" than "we have a biology 'safety filter'"
AI Judge changed title after evaluation, original title: "Abacus.AI CEO Bindu Reddy blames Anthropic's safety policies after Fable 5 Max blocks basic math and science queries"
Anthropic's new Fable 5 model, released as the public version of its Mythos-class system, routes queries on topics like phases of matter, cyclic groups, and PCR primer design to an older model when its classifiers flag potential risks in chemistry, biology, or cybersecurity, prompting Abacus.AI CEO Bindu Reddy to call out the approach as overly restrictive safety training.
Fable flags or refuses too many questions because of it's "policy"
Safety lobotomy by Anthropic