Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Claude Fable 5

anthropic.com

Post reply on HN
Print view View original

411–420 of 1001 posts

Re: Claude Fable 5

#411

Post by hashmap »

bkjlblh wrote: post #220

> In light of the ability of recent models to accelerate their own development, we’ve implemented new interventions that limit Claude’s effectiveness for requests targeting frontier LLM development (for example, on building pretraining pipelines, distributed training infrastructure, or ML accelerator design). Using Claude to develop competing models already violates our Terms of Service, but enforcing this restrictio…

3 months before asking for what to eat before a linear algebra exam trips the machine learning topic ban is my guess. I got flagged immediately asking why my JEPA thing breaks weird.

Re: Claude Fable 5

#413

Post by ravila4 »

Fable's ridiculous. It's flagging basic biology research questions as a security risk. I'm talking basic fundamental genetics topics that make working on any genetics-adjacent codebase unusable.

Re: Claude Fable 5

#414

Post by seemaze »

bkjlblh wrote: post #220

> In light of the ability of recent models to accelerate their own development, we’ve implemented new interventions that limit Claude’s effectiveness for requests targeting frontier LLM development (for example, on building pretraining pipelines, distributed training infrastructure, or ML accelerator design). Using Claude to develop competing models already violates our Terms of Service, but enforcing this restrictio…

Ah, so this is why raw Mythos was too "dangerous" to realease..

Re: Claude Fable 5

#415

Post by unfunco »

I tried running a simple security review on a Terraform module I made and after some thinking, it responded:

> ● The model returned no content because the response was blocked by content filtering.

> Blocked? We are performing a defensive security review on a Terraform module I made, what's blocked by content filtering? This is a legitimate use-case.

> ● The model returned no content because the response was blocked by content filtering.

A waste of money. I'm not going to just hope that the model returns a response, I'm already for paying for wrong responses, I'm not going to pay for no response, especially when I'm paying per token.

Re: Claude Fable 5

#417

Post by grumbelbart »

alvis wrote: post #87

Another thing to note: 30-day retention for all traffic on Mythos-class models Is it good or bad? 30 days is a long time for anything bad to happen

It's bad. I believe them not to use it for training, but t means relevant data can and will be exfiltrated by US agencies or through court orders (see NY Times vs. OpenAI, where only traffic without any rentention was safe).

Re: Claude Fable 5

#419

Post by AAYALAG »

eggbrain wrote: post #24

For those of us on subscription plans: * From today through June 22, Fable 5 is included on Pro, Max, Team, and seat-based Enterprise plans at no extra cost. * On June 23, we’ll remove Fable 5 from those plans. Using it after that will require usage credits. If capacity allows, we’ll extend the included window. * After this point—when sufficient capacity allows us to do so—we aim to restore Fable 5 as a standard part…

[dead]
Post reply on HN