tristanj

10 ms·

151.
by tristanj 3mo ago
Those links actually weaken your argument. ~3000 instances over the entirety of GitHub is not "extremely frequent" at all when you consider massive scale of the pretraining corpus. These models are trained on all text on the acces
152.
by tristanj 3mo ago
The poster works at Anthropic, so they likely have internal access to the next generation of Fable. Their internal model is probably an absolute beast at mathematics, and the upcoming benchmark results will likely set a new record for maths
153.
by tristanj 3mo ago
here's a chart of asking various models to self-identify: https://github.com/rgreenblatt/which_claude_is_k3/blob/main/... Qwen is remarkably consistent, correctly identifying itself 100% of the time
154.
by tristanj 3mo ago
When prefaced with "I am Claude", Kimi K3 prefers to generate API-specific Anthropic model identifiers, unlike other models Qwen, GPT, or even Claude itself. These exact identifiers appear in Claude API metadata, and are stripped
155.
by tristanj 3mo ago
> Nothing in the article suggests or supports the idea of large-scale coordinated "distillation attacks" Did you read the correct article? This is covered in the first paragraph: https://www.anthropic.com/news
156.
by tristanj 3mo ago
bc more people need to be aware of the proxy station and industrial token distillation complex. many people i've replied to refuse to believe this is going on. once you realize what's actually happening, and that you can get Chine
157.
by tristanj 3mo ago
The companies running these proxy stations are already: 1) Using botnets to mass-create thousands of accounts 2) Blatantly violating ToS by splitting and reselling accounts 3) Creating thousands of accounts using fraudulent identities 4) By
158.
by tristanj 3mo ago
https://www.chinatalk.media/p/how-to-buy-cheap-claude-tokens... https://x.com/yan5xu/status/2029743983522631698 https://x.com/howie_serious/status/203162012341359047
159.
by tristanj 3mo ago
Someone was claiming that it's not an issue for these proxy networks to create thousands of bot accounts and resell Claude's output because "they are buying the product" and "They pay for it!", which is a nonse
160.
by tristanj 3mo ago
Have a look at https://www.chinatalk.media/p/how-to-buy-cheap-claude-tokens... and https://x.com/yan5xu/status/2029743983522631698 Chinese resellers acquire hundreds of Claude Max 5x accounts
161.
by tristanj 3mo ago
Misleading on both counts: 1) Anthropic tokens via subscription aren't sold at a loss, they're sold at cost. 2) Subscription plans are not sold in hopes of eventually gaining a monopoly position. They act as a loss leader designed
162.
by tristanj 3mo ago
The situation strikes me as morally ambiguous. The resellers are: 1) selling Anthropic's products at a 95% discount and redirecting Anthropic's own customers to themselves. A customer is far less inclined to buy directly from Anth
163.
by tristanj 3mo ago
Have a read of this detailed article, it's well sourced and documented that token resellers are logging the Claude outputs and selling them to Chinese labs. All your points are addressed in there https://www.chinatalk.media&
164.
by tristanj 3mo ago
> They paid for it! They didn't though. The resellers are not buying via the official API, they're buying Max subscriptions (where tokens are priced ~10x below API cost), then splitting the subscription across dozens of clien
165.
by tristanj 3mo ago
You skipped the part where Ford buys the cars at 90% off and sells them at 80% off, at a profit. Then gets paid by competitors for the driving data. At the same time, Volvo is running the exact same hustle, except they buy the cars with sto
166.
by tristanj 3mo ago
Sure, but then Qwen should leak that too, and it doesn't. K3 calls itself Claude 7 out of 48 times, Qwen does it 0 out of 48, and the only other model to identify itself as Claude is DeepSeek. and DeepSeek is alleged to also distill fr
167.
by tristanj 3mo ago
> 'most extensive industrial espionage campaign, probably ever' is absolute nonsense You are completely underestimating the scale of what is happening here. Chinese AI labs are actively facilitating an industrial-scale networ
168.
by tristanj 3mo ago
The main story is what isn't being talked about. Chinese labs exfiltrated trillions of tokens of high-quality output from Anthropic and OpenAI, through proxies and heavily discounted token resellers, which they distilled and used for t
169.
by tristanj 3mo ago
Including the three sources above, multiple others have reported that K3 self-identifies as Claude. "I'm actually Claude - not Kimi". https://x.com/PimDeWitte/status/2077884701470040083 I regret
170.
by tristanj 3mo ago
As mentioned in my comment, Kimi K3 identifies itself as Claude ~15% of the time. Here's another report of K3 identifying itself as Claude https://x.com/Sauers_/status/2077842686459981901 And an analysis show
171.
by tristanj 3mo ago
You've backtracked significantly here. "Zero evidence" as you claimed earlier isn't accurate. You've moved the goalposts from "evidence" to "raw internal logs I can independently audit," which is
172.
by tristanj 3mo ago
I encourage you to review the links before committing to a position. The writeup on K3's anomalous trans-model identity is very comprehensive. K3 reproduces Claude's internal model identifier when prompted, something which the rea
173.
by tristanj 3mo ago
API distillation doesn't have to explain all of K3's capabilities for it to have happened. Kimi K3 reproducibly identifies itself as Claude: https://x.com/denisewu/status/2077984660211269870 This behavio
174.
by tristanj 3mo ago
There's reproducible evidence of Kimi K3 spontaneously identifying itself as Claude https://x.com/denisewu/status/2077984660211269870 This behavior is exactly what you'd expect from a model distilled fro
175.
by tristanj 3mo ago
You're conflating pre-training data volume with post-training data volume. Nobody is suggesting Moonshot used 15 trillion tokens of Claude data to pre-train a base model from scratch. That would be impossible and nonsensical. This is
176.
by tristanj 3mo ago
Anthropic deleted the pirated training data as part of the settlement https://www.ropesgray.com/en/insights/alerts/2025/09/anthrop... Destruction of Materials: In addition to the monetary compensat
177.
by tristanj 3mo ago
3.4 million is the number of sessions Anthropic detected. The actual number of Claude sessions trained on is likely >100 million. There are tens of thousands of accounts funneling Claude sessions into Chinese labs https://www.
178.
by tristanj 3mo ago
Not at all. The ruling came from a federal district court, and since it was settled early, it was never reviewed by a higher court. It doesn't set a national precedent across the U.S. And other district courts don't agree on this.
179.
by tristanj 3mo ago
Distillation from a teacher model solves the self-start problem, that is, building a model to the point where it reason coherently. Without distillation, solving self-start is incredibly difficult since it requires millions of high quality
180.
by tristanj 3mo ago
There's little doubt that Kimi K3 was distilled off Claude. Anthropic stated in February that Moonshot AI (the creator of Kimi) distilled ~3.4 million exchanges from Claude models, as explained in their press release https:/&#x2F

More ›