frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Writing fingerprint analysis of responses reveals Kimi's similarity to Claude

https://typebulb.com/u/lab/you-re-relatively-right/full
26•maxloh•1h ago

Comments

great_psy•35m ago
Shouldn’t we expect a convergence as models become more powerful ?

There is an optimal answer to any question. Something that maximizes utility and minimizes tokens.

isoprophlex•35m ago
Stolen data was stolen. Oh no! Anyway.
nonethewiser•33m ago
Can you elaborate?

This looks very similar to the claim that distilling a model from Anthropic is the same thing as Anthropic distilling the model from information on the internet.

Which is very flawed, since distillation requires the thing to exist in the thing it’s distilled from. And no LLM model existed in the information Anthropic used to train the model. Instead the model was built using information and utilizing new technology including hardware, software, transformer architecture, etc.

wongarsu•17m ago
I don't believe I understand your argument? Are you claiming a moral, legal or practical difference? Or are you saying that Anthropic spending resources on training an LLM is somehow different from an author spending resources on writing a book?

In any case, I doubt Kimi was trained without "stealing" the same data. Assembling all of your training data from Claude responses seems infeasible. It's much more likely that Kimi's base model was trained similarly to any other base model, with terabytes of data from all imaginable sources. Then the model was fine-tuned with "high-quality" data, followed by reinforcement learning. Throwing in lots of chat transcripts from other chatbots into the "high-quality" dataset would be expected, and is done to some degree by everyone, but maybe a lot more for Kimi. And likely they did a lot of reinforcement learning against the Claude API

The model would exist without Claude, it just wouldn't be nearly as coherent or smart

ghm2199•16m ago
All ML/AI models are comparable to some form of compression(al beit lossy) of information and in this case copyrighted information. The OP is pointing to this as stolen data(by all the companies that started with pre trained models)
yreg•14m ago
So what?

If training on copyrighted data without authors consent is ok then distilling is ok as well.

maxloh•13m ago
Think of it in terms of distilled knowledge, not distilled LLMs.

I find both claims unsound, though. Knowledge or model behavior itself is not copyrightable, so all these claims just boil down to the "I am not happy with that" argument. You cannot claim someone is stealing something you don't own in the first place.

jayd16•10m ago
You might as well say a photocopy of a book was "built using information and utilizing new technology including hardware, software, transformer architecture, etc."
codelion•34m ago
How do you do cross entropy analysis for Claude without logits?
shiandow•29m ago
I assume Kimi does provide logits. You can do cross entropy both ways. With slightly different results.

Edit: No that doesn't seem to be what's happening here. I think it's some kind of word frequency analysis.

josh-paul•31m ago
This looks to be more behavior based, not logit based? What is the actual claim?
weego•27m ago
It seems to boil down to China bad, US good, but with tech people speculating so it's definitely valid
zobzu•13m ago
honestly 99% of hn, reddit, etc. is the exact opposite of that statement.
refulgentis•28m ago
I didn’t participate in the discussion yesterday because I find it implausible Fable was available long enough (3-4 weeks cumulative?) to get data and train and have it fundamentally affect.

But I don’t grok training enough to know that’s silly.

If my new prior is you can…that’s a pretty thin moat that’s essentially indefensible.

credit_guy•7m ago
My guess is they used the other Anthropic models extensively for synthetic data generation. The top most similar models for K3 are (lower means more similar)

Fable 5 -> 0.42

Opus 4.8 -> 0.45

Sonnet 5 -> 0.45

Opus 4.7 -> 0.46

Grok 4.3 -> 0.52

There's an obvious jump at Grok 4.3, and it would not surprise me that the similarity there is because Grok used Anthropic models for training too (you can get the similarity list for Grok and it does look like the top most similar models for Grok are either Anthropic models or some Chinese models).

The damning evidence that K3 used Anthropic models for training is that K3 is more similar to those models than it is to K2.6. If you look at the Anthropic, OpenAI or Google models, they are most similar with their own other models. Not so with K3, where K2.6 is less similar than 15 other models.

Now, why is Fable 5 the most similar to K3 and not Opus 4.8. I think it's quite likely that K3 did some fine tuning at the end, when Fable 5 became available. They probably had all the infrastructure in place, and Mythos had been announced for months, so they were probably waiting for the second the newest Anthropic model was released to start using it for synthetic data generation.

pandoro•27m ago
All frontier models have been trained without any regards for IP protection laws. I don't see how anyone can argue in good faith that distillation is not fair game and does not ultimately "benefit humanity™"
skybrian•22m ago
Does distilling actually violate any IP protection laws? Sure, it’s against their terms of service.
nonethewiser•21m ago
Training a model is very expensive and creates something no individual rights-holder could. Distilling a model copies this value add and captures it without bearing the cost that created it.
mosura•18m ago
What of the costs for creating the data that was used to train the model being distilled?
yandie•15m ago
At least 1.5B by stealing books, per a recent ruling.

I don’t feel sorry for the model companies

mosura•12m ago
They only had to pay for storing the books on a server for later possible use. They did not have to pay anything for the training which was declared fair use.
throwa356262•10m ago
This data sort of disqualifies itself: unless Moonshot has a time machine, K3 should be more similar to Opus 4.5-4.8 than Fable 5.

Keep in mind, Anthropic started limiting access and introduced anti-distillation measures around 4.5-4.6 (?). So the majority of distillation should have happened on earlier models.

Maybe a better explanation is that they have access to the same training datasets? Which if private can again raise questions about theft, but on a very different level.

sebastianconcpt•7m ago
Honest take: different surface, same category.
josefritzishere•5m ago
AI stealing from AI? That's shocking that an industry built on stealing would engage in stealing.
causal•2m ago
So this shows distance relative to other models, but I don't have a good sense for what these numbers say in absolute terms.

K3-to-Fable is blue at 0.42. Is 0.42 meaningful, or did we set 0.4 as the lower bound because it makes 0.42 look significant?

Sol-to-Fable is 0.69. It's dark yellow, making this look VERY different from 0.42. But is it? What do these numbers mean in absolute terms?

8note•2m ago
given that github is full of vibe coded repos, reddit is full of bot comments, and most blog posts are ai slop, isnt it guaranteed that anyone training on big public data sets will be closely tracking each other in text style?
titzer•18m ago
An incredibly ironic comment.
jayd16•14m ago
I'm going to hope this was sarcasm and if it is, it's great.
8note•7m ago
the outputs of the model have no property protections, and training a model on the outputs of another model does the exact same thing - its expensive and creates new value over what was in the input - a set of documents.
pandoro•4m ago
Could you give an example of the value that only training a model can create but none of the rights-holder could? I feel like if you got a direct, instant communication channel to any of the rights-holder that created the content in the training set of those models, you'd get more value than what the LLM could ever give you on any specific subject.
shlewis•3m ago
Yes. I bet Moonshot paid for API access as opposed to pirating like Anthropic did.
applicative•2m ago
One mentions distillation to intimate that one is in fact still the leader. It has zero to do with fairness.

Writing by hand is good for your brain

https://nealstephenson.substack.com/p/writing-by-hand-is-good-for-your
178•dwwoelfel•1h ago•57 comments

Astronomers may have found the first exomoon

https://www.eso.org/public/news/eso2610/
102•MarcoDewey•2h ago•40 comments

Software rendering in 500 lines of bare C++

https://haqr.eu/tinyrenderer/
71•mpweiher•1h ago•14 comments

Startup founders urge Trump not to shut off Chinese open weight AI

https://www.politico.com/news/2026/07/22/startup-founders-urge-trump-not-to-shut-off-chinese-open...
143•theanonymousone•47m ago•130 comments

Learn OpenGL, extensive tutorial resource for learning Modern OpenGL

https://learnopengl.com/
39•ibobev•1h ago•5 comments

Writing fingerprint analysis of responses reveals Kimi's similarity to Claude

https://typebulb.com/u/lab/you-re-relatively-right/full
26•maxloh•1h ago•32 comments

How much energy do data centers and artificial intelligence use?

https://ourworldindata.org/how-much-energy-do-data-centers-and-artificial-intelligence-use
29•surprisetalk•1h ago•7 comments

Encryption and Globalization 15 Years Later: E2EE and the "Going Dark" Debate

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6959699
13•iamnothere•1h ago•5 comments

AI Companies Are Trying to Hide a Staggering Amount of Debt

https://futurism.com/artificial-intelligence/ai-companies-hide-debt-off-balance-sheet
260•technewssss•2h ago•125 comments

DARPA, U.S. Air Force fly AI-controlled F-16

https://www.darpa.mil/news/2026/darpa-us-air-force-fly-ai-controlled-f-16
28•r2sk5t•2h ago•10 comments

Hybrid-Electric Aicraft Engine Targeting 30% Fuel Efficiency

https://www.rtx.com/news/news-center/2026/07/21/rtxs-pratt-whitney-canada-advances-hybrid-electri...
19•r2sk5t•1h ago•1 comments

NerdLens: Mirror your computer with cardboard VR

https://github.com/ThatXliner/NerdLens
9•thatxliner•1h ago•2 comments

Show HN: Remux – an open-source tmux workspace designed for iPhone

https://github.com/h3nock/remux
19•bitwise42•1h ago•2 comments

Learn WebGPU for C++

https://eliemichel.github.io/LearnWebGPU/
9•ibobev•1h ago•1 comments

Show HN: Whetuu – a zero-config cross-shell prompt written in Zig

https://yamafaktory.github.io/whetuu/
10•yamafaktory•1h ago•2 comments

What happens when the information runs out

https://blog.jimgrey.net/2026/06/30/what-happens-when-the-information-runs-out/
14•mooreds•2h ago•0 comments

Free way to get your data out of ChatGPT Business accounts

https://github.com/Conradqh/scrapemychats
8•conradqh•20m ago•2 comments

OpenStreetMap: Admin_level for All Countries

https://wiki.openstreetmap.org/wiki/Tag:boundary%3Dadministrative#Table_:_Admin_level_for_all_cou...
26•ivanjermakov•4d ago•1 comments

Scanning for Pangram Errors

https://veryfineprint.substack.com/p/scanning-for-pangram-errors
49•jsnell•6d ago•26 comments

Vulkan Tutorial

https://vulkan-tutorial.com/
15•ibobev•1h ago•1 comments

Show HN: macOS menu-bar manager for SSH port forwards

https://github.com/lx2026/RelayBar
14•linxy97•1h ago•1 comments

Cruller: Bun's Zig Runtime, Continued on Zig 0.16

https://ziggit.dev/t/cruller-buns-zig-runtime-continued-on-zig-0-16/16734
134•Erenay09•10h ago•93 comments

The Telegarden (1995-2004)

https://goldberg.berkeley.edu/garden/
18•zetamax•2d ago•2 comments

The Unity CLI: manage Unity from your terminal

https://unity.com/blog/meet-the-unity-cli
65•nateb2022•1d ago•19 comments

Silent Replacement of Trusted macOS App Executables

https://mysk.blog/2026/07/23/macos-overwrite-app-executables/
11•jurip•2h ago•0 comments

OpenAI’s accidental attack against Hugging Face is science fiction that happened

https://simonwillison.net/2026/Jul/22/openai-cyberattack/
157•abhisek•14h ago•118 comments

Brow6el: A full-featured web browser for the terminal using Chromium

https://tangled.org/janantos.tngl.sh/brow6el
15•nerdypepper•2h ago•0 comments

Alphabet's cash burn raises alarm for Big Tech as AI spending climbs

https://www.reuters.com/business/retail-consumer/alphabets-cash-burn-raises-alarm-big-tech-ai-spe...
224•1vuio0pswjnm7•2h ago•208 comments

UK households free to install plug-in balcony solar panels from end of August

https://www.theguardian.com/money/2026/jul/23/uk-households-free-install-plug-in-balcony-solar-pa...
11•DamonHD•30m ago•2 comments

Ask HN: Is there a website that tracks excessive writes to SSDs in OS/app betas?

7•amichail•49m ago•8 comments