Hacker News: mips_avatar

New comment by mips_avatar in "Cybersecurity researchers aren't happy about the guardrails on Anthropic's Fable"

mips_avatar — Thu, 11 Jun 2026 02:41:14 +0000

They’re not safety guardrails they’re anthropic doesn’t like anyone who isn’t anthropic working on AI rails

New comment by mips_avatar in "Cybersecurity researchers aren't happy about the guardrails on Anthropic's Fable"

mips_avatar — Thu, 11 Jun 2026 01:53:06 +0000

Anthropic specifically said that those notifications are temporary and fable5 will only pretend to help you if it’s ml classifier gets tripped

New comment by mips_avatar in "Cybersecurity researchers aren't happy about the guardrails on Anthropic's Fable"

mips_avatar — Thu, 11 Jun 2026 00:06:59 +0000

Anthropic is trying to hide bad behavior by being vague, it's important to not be vague when calling it out.

New comment by mips_avatar in "Cybersecurity researchers aren't happy about the guardrails on Anthropic's Fable"

mips_avatar — Wed, 10 Jun 2026 23:49:04 +0000

From the model card: "the safeguards will limit effectiveness through methods such as prompt modification, steering vectors, or parameter-efficient fine-tuning" aka they will take your ML research code and inject bugs into it until it breaks using a LORA (or some other form of PEFT)

New comment by mips_avatar in "Cybersecurity researchers aren't happy about the guardrails on Anthropic's Fable"

mips_avatar — Wed, 10 Jun 2026 23:22:28 +0000

Yeah we can probably figure out how to run it on xiaomi gpus

New comment by mips_avatar in "Cybersecurity researchers aren't happy about the guardrails on Anthropic's Fable"

mips_avatar — Wed, 10 Jun 2026 22:57:55 +0000

They've said that they'll stop notifying developers when this gets triggered, instead they'll load in basically like a LORA that's designed to inject bugs into your code.

New comment by mips_avatar in "If Claude Fable stops helping you, you'll never know"

mips_avatar — Wed, 10 Jun 2026 06:04:17 +0000

I guess it’s better they’re open about doing bad things. But now it’s a problem that they think this sets a precedent. They are one step away from feeling justified in using claude code running on a deepseek engineers laptop to hack deepseek to destroy their training system.

New comment by mips_avatar in "If Claude Fable stops helping you, you'll never know"

mips_avatar — Wed, 10 Jun 2026 05:30:19 +0000

Given that Anthropic has never released anything open weights I wouldn’t count on the fact that they view finetuning Gemma 4 as something allowable. I think they think nobody other than Anthropic should have AI

New comment by mips_avatar in "If Claude Fable stops helping you, you'll never know"

mips_avatar — Wed, 10 Jun 2026 00:31:40 +0000

Margin compression is terrifying

New comment by mips_avatar in "If Claude Fable stops helping you, you'll never know"

mips_avatar — Tue, 09 Jun 2026 23:01:36 +0000

Thats exactly it

New comment by mips_avatar in "If Claude Fable stops helping you, you'll never know"

mips_avatar — Tue, 09 Jun 2026 22:01:09 +0000

At least it gave an error! This whole silent nerfing idea is so wrong

New comment by mips_avatar in "If Claude Fable stops helping you, you'll never know"

mips_avatar — Tue, 09 Jun 2026 22:00:33 +0000

Fine for me. Not for thee

New comment by mips_avatar in "If Claude Fable stops helping you, you'll never know"

mips_avatar — Tue, 09 Jun 2026 21:39:59 +0000

Like if you hired a different services company who decided to sabotage your business that would be fraud.

New comment by mips_avatar in "Claude Fable 5"

mips_avatar — Tue, 09 Jun 2026 21:29:50 +0000

Like if you're using claude code on a feature tangential to your training pipeline it's allowed to nerf itself and damage your AI work.

New comment by mips_avatar in "If Claude Fable stops helping you, you'll never know"

mips_avatar — Tue, 09 Jun 2026 21:24:53 +0000

I'm really uncomfortable with these changes, like everything Anthropic's doing as "frontier research" today will be regular product engineering in a year.

If Claude Fable stops helping you, you'll never know

mips_avatar — Tue, 09 Jun 2026 21:19:11 +0000

Related: https://simonwillison.net/2026/Jun/10/if-claude-fable-stops-helping-you/

Comments URL: https://news.ycombinator.com/item?id=48467896

Points: 1029

# Comments: 500

New comment by mips_avatar in "Claude Fable 5"

mips_avatar — Tue, 09 Jun 2026 18:24:09 +0000

They said in their terms and conditions they will silently sabotage you if you do this.

New comment by mips_avatar in "Claude Fable 5"

mips_avatar — Tue, 09 Jun 2026 18:01:07 +0000

It's bad that Anthropic can determine what this means. If you're building a modern app you're likely training your own embedding models and now anthropic can just silently sabotage your training pipelines?

New comment by mips_avatar in "Apple Core AI Framework"

mips_avatar — Tue, 09 Jun 2026 03:44:22 +0000

Seems like they still won’t let you run models on GPU while the phone is closed or the user switches apps

New comment by mips_avatar in "Gemma 4 12B: A unified, encoder-free multimodal model"

mips_avatar — Wed, 03 Jun 2026 17:58:53 +0000

I don't think we've bottomed out on what we can do with embedding models. They're these tiny models that absolutely rip on modern cpus with 8 bit int optimizations. Like in my app we can say pretty definitive things about hundreds of millions of places in the world on retrieval tasks on regular hardware.