Part of Anthropic's Claude Evolution · 12 stories · since Sep 23 · newest 21m ago
Anthropic cuts Claude reasoning-token theft by up to 99%, flags Chinese rivalsAnthropic says defenses cut reasoning-token theft by 90-99%
6 Yesterday 11:45 PM · 23h ago · 10 articles · 9 posts · 13 comments · 5 sources · development 6 of 6
Anthropic said measures including preserved thinking in Fable 5.1 and Opus 5.5 reduced data harvesting through distillation by about 90 to 99 percent, linked industrial-scale attacks to Chinese firms Zhipu and Moonshot, and said it banned tens of thousands of accounts, including one request it believes came from a PLA-affiliated user.
“first you steal pretty much all of humanity's cultural output, then you try locking that up behind a paywall and complain when people do the same to you.”
jacquesm, HN commenter · hn ↗Anthropic Maker of ClaudeEmily Bender Computational linguist
Mustafa Suleyman Microsoft AI headZhipu Chinese AI companyMoonshot Chinese AI company
The whole story articlespostscomments the bright band is this development · numbered dots are the others · click one to jump
What was reported 2 claims about this development
-
7 outlets Anthropic Says New Defenses Cut Data Harvested Through AI Distillation Attacks By 90% To 99%
first by Pulse 2.0, 23h ago · also RuntimeWire, Reuters, Investing.com News, Business Standard, Anthropic, Straits Times
5 more headlines
- Anthropic details new defenses against AI model distillation attacks RuntimeWire · 23h ago
- Anthropic says new safeguards sharply reduce AI distillation attacks Reuters · 23h ago
- New safeguards sharply reduce AI distillation attacks, says Anthropic Business Standard · 23h ago
- Anthropic says new measures like preserved thinking in Fable 5.1 and Opus 5.5 have reduced data harvesting through distillation attacks by about 90% to 99% Anthropic · 23h ago
- New safeguards sharply reduce AI distillation attacks: Anthropic Straits Times · 20h ago
-
first by Forbes Europe, 23h ago · also The Decoder
1 more headline
What people said 17 voices · best of 18 · verbatim
-
A couple of observations on this (bizarre) post:(1) This seems to be a pretty bad accusation to make without iron-clad proof.(2) So their AI is not able to help defend against this.(3) If GLM 5.3 really is that good why would anybody use Anthropic's products going forward, this almost reads like an ad for GLM.(4) Regardless of whether it is true…
-
Anthropic says it has reduced the data harvested from Claude in distillation attacks by around 90-99%. I congratulate them on this achievement. It's consistent with the point I made a few months ago (All-In Ep. 282): Anthropic, not the government, was in the best position to stop distillation of i...
-
oh I know, they put out lots of papers on the topic that show as much. I am sure they use distilliation as a final fine tuning thing at most, unlike lots of people who seem to think that some Chinese models are literally just 100% distillation trained (which for many reasons is flat out impossible).
-
D
Why are they trying to ban abusive or cruel behaviour towards # AI ? AI are not conscious or sentient, and they're not living things, they're just code: https://www. theguardian.com/technology/202 6/oct/08/anthropic-bans-abusive-behavior-claude # ArtificialIntelligence
-
> This means that distilled models could be used to cause significant harmAhem: https://www.felonybench.com/I wonder why there is not even a single one of those allegedly "distilled"/"unsafe" model on that list, yet models of companies with "proportionally robust safeguards" are, hm?> Distillation is a problem for every developer and platformNo…
-
Over the last few months, we've made significant progress in reducing the effectiveness of distillation attacks against Claude. In this post, we explain how:
-
Good for them. I expect this changes nothing. But good that they can no longer claim other's progress as their own. But seriously, wait for the "those sneaky Chinese found another way to train on our models" claim.
-
S
Anthropic Bans Cruelty Toward Claude https:// slashdot.org/story/26/10/13/18 57238/anthropic-bans-cruelty-toward-claude?utm_source=rss1.0mainlinkanon
-
If you read the terms of service carefully, it seems to suggest they only prohibit you distilling outputs, you should be able to publish them no problem. I'm surprised there isn't a large scale open source project for cataloguing and scoring transcripts. Maybe there is and I just don't know about it. Businesses might not want to publish…
-
And then their CEO seems like he's getting off on the idea of people losing their careers.
-
Framing distillation as an attack now?Distillation attacks are covert, industrial-scale campaigns to extract the capabilities of one model and replicate them in another, without the authorization of the original model’s developers.Replace a few key words and it sounds like reading all the worlds media and offering a service that summarises…
-
The horror! Replicating model capabilities without authorization.Seriously though, these clowns at Anthropic are not just hypocritical. They are clearly feeling threatened by open source AI. I almost couldn’t believe the new blog post today where they argue that dangerous open source models might soon be running on (checks notes) smart ovens.
-
Isn't this a little reductive? I'm not an expert on LLM/AI technology, but isn't the secret sauce the reasoning layer? Isn't that what makes Claude et al so much more powerful than a more "vanilla" LLM? That's what they're trying to protect in this post, which seems awfully reasonable?
-
The way AI has consolidated power away from the hands of a few hundred large organization to a few dozen needs to be attacked. We need to build peer distillation networks. We need to be able to train frontier models in the open and design hardware for open inference.
-
Great ad for GLM 5.3 I have been an absolute buffoon paying for anthropic's stunted model s when I can access their same accessibility and more for cheaper. I'm switching
-
Are we all secretly (or not so) hoping distillation keeps working to commoditize the frontier?
-
Wouldn't it be funny if they killed off distillation and 93% of their business went away?
All 6 developments of Anthropic cuts Claude reasoning-token theft by up to 99%… →
NewswiresMastodonHacker NewsXRedditBlueskyGoogle News