#distillation — Public Fediverse posts

El lado del mal - Pensamientos Robados: Cómo descifrar y weaponizar trazas de razonamiento en LLMs https://www.elladodelmal.com/2026/08/pensamientos-robados-como-descifrar-y.html #IA #AI #Gemini #ChatGPT #Claude #Privacidad #Jailbreak #PromptInjection #Criptografía #Distillation #Hacking

讀者投書 2026-08-14 12:00:00 CST

Zuckerberg Is Right About Open, Decentralized AI. He’s Also The Last Person You Should Trust To Deliver It

https://fed.brid.gy/r/https://www.techdirt.com/2026/08/12/zuckerberg-is-right-about-open-decentralized-ai-hes-also-the-last-person-you-should-trust-to-deliver-it/

"Distilling Runes" song
https://youtu.be/PI5rdC7tPpI

#AI #model #distillation #AmazonBedrock
1. Ask Teacher model questions
2. Use question + answer pairs to fine-tune Srudent model
3. Distilled model keeps Student speed+cost but with use-case-specific accuracy/knowledge from Teacher

“The cost of land (for data centres), some equipment (to fill these with ai servers) and labour (from construction workers to #AIWhizzKids) is lower in China than in America.

Chinese firms are also accused of training their models using the outputs of expensive American ones. Such #distillation, as the process is known, lets China’s ai #developers free-ride on some American spending. Some Chinese ai spending may also not be captured in the headline capital-expenditure figures.

Part of the sums that Chinese ai firms, especially smaller ones, might have spent on #Nvidia’s best chips may instead have gone towards squeezing results from inferior ones.

#DeepSeek, which gave American ai firms and investors a jump scare by launching a surprisingly powerful and efficient model in early 2025, is probably investing a fair bit in its techniques that reduce demand for #computing power.”

When you steal from the original thieves 🤣

#AI / #ArtificialIntelligence / <https://www.economist.com/finance-and-economics/2026/08/03/how-china-gets-better-bang-for-its-buck-than-america-in-ai> / <https://archive.md/aUOOT>

Ban the thieves, not the technique: the AI fight turns on a legal distinction

Repost to help others discover this.

https://1ban.news/china-ai-freeriding-sanctions-distillation/

#1ban #china #freeriding #sanctions #distillation #geopolitics

TIL that Mezcal owes its existence to Filipinos.

https://www.mezcalistas.com/filipino-still/

#mezcal #filipinos #distillation #mexico #alcohol

Anthropic Says It’s Against A Ban On Open Weight Models. It Just Wants To Ban Everything That Makes Them Good.

https://fed.brid.gy/r/https://www.techdirt.com/2026/07/29/anthropic-says-its-against-a-ban-on-open-weight-models-it-just-wants-to-ban-everything-that-makes-them-good/

The claim that China's open-source AI models pose risks related to cultural and political indoctrination via pushing out censored perspectives - can EASILY be defeated - researchers have shown this to be true.

Not only can distillation be used to build "new" models by leveraging existing models, distillation can be applied to China's open-source models to create customized models that defeat censorship aspects embedded in the original version of the model(s).

Researchers who dug into the China models nailed it . .. "These are raw materials. It’s software.”

The hard fact is that most organizations will want to run models in-house, customized for their needs, rather than run AI off-the-shelf models. https://www.semafor.com/article/07/29/2026/censorship-in-chinese-ai-models-can-be-undone-new-research-shows #AI #Distillation #FrontierModels #LLMs #DeepSeek #MoonShot #Open-Source-AI #OpenSource #Censorship #ChineseModels #Security #Risk #Chinese_AI_Models

#Anthropic does not advocate for a ban on #OpenWeightAI, recognising their value as a #publicgood. Instead, the focus should be on preventing #authoritariangovernments from accessing powerful #chips and cracking down on industrial-scale #distillation operations. Additionally, mandatory #safetytesting for all sufficiently capable models, regardless of origin or openness, is crucial to address concerns about misuse and alignment problems. https://www.anthropic.com/news/position-open-weights-models?AIagents.at #AIagent #AI #ML #NLP #LLM #GenAI

Anthropic CEO in a letter urging ban on model distillation. Same old story
- Still thinks "Distillation" is something different to scraping the Internet.
- Still thinks Chinese "authoritarian regime" is different to their own "authoritarian regime".
- Still technically powerless from being scraped, just like any other server on the internet, so they need laws to protect their specific server.

https://www.anthropic.com/news/position-open-weights-models

#anthropic #openSource #openweight #llm #distillation #tech

#Distillation, a technique used to improve #AImodels, has become a contentious issue: While some argue it enables #innovation and #efficiency, others, including U.S. officials, are concerned it allows Chinese companies to rapidly develop competitive AI models by leveraging American #intellectualproperty. This has led to a debate about whether distillation constitutes #IPtheft and how to balance innovation with #nationalsecurity concerns. https://www.cnbc.com/2026/07/25/hat-is-distillation-and-why-is-everyone-so-obsessed-with-it-this-week.html?eicker.news #tech #media #news

🤡 In a groundbreaking revelation, it turns out you can distill #AI models like fine whiskey—except now, they just taste cheaper! 🥴 GitHub's new tool promises frontier quality at half the cost, proving once again that if you can't dazzle them with brilliance, baffle them with #buzzwords. 🙄
https://github.com/experientiallabs/world-model-optimizer #Distillation #GitHub #Innovation #CostEfficiency #TechTrends #HackerNews #ngated

Distill and serve small models with frontier quality for half the cost

https://github.com/experientiallabs/world-model-optimizer

Comments: https://news.ycombinator.com/item?id=49063454

#HackerNews #distillation #modeloptimization #smallmodels #costefficiency #frontierquality

#BigTech #AI #TechBro companies crying to mommy to get the bad companies to stop (maybe) using their models' output for training is now a standard play from the #monopolies and wannabe monopolies.

If a (Chinese) company gets too good at doing what American companies do, then force an embargo or force a sell. It was the #Meta/#Facebook with #Tiktok play, and now it's the #LLM play.

If #scraping copyrighted works for training is legal, so is #distillation.

#Capitalists sure do hate #Capitalism.

China's Moonshot AI stole IP from Anthropic? NOT so fast say industry boffins!

U.S. gov peeps go full bore claiming Moonshot AI employed distillation to transfer knowledge from Anthropic's Fable 5 AI model into the Kimi K3 AI Model. So far the claimed "evidence" is non-existent.

Industry researchers say 1) not enough time has passed for distillation to have accomplished this feat, and 2) if it was distillation then EVERYONE should be able to do this.... and that has not happened. https://siliconangle.com/2026/07/23/senior-white-house-official-accuses-moonshot-ai-copying-anthropics-leading-frontier-model/ #AI #Kimi #K3 #Distillation #IP #IPTheft #China #USGov #Fable #Anthropic #Moonshot #AIDistillation

I do find the #openweight discussions quite interesting. As a #floss developer I generally favour openness although I'm never quite clear what the training inputs are. They could be the product of #distillation attacks but that's basically hoovering up the internet with extra steps which doesn't seem so different from the big #frontier models anyway: https://www.youtube.com/watch?v=YP73B9D20V4 #youtube #fireship

The Moose has Left the Woods!

OR when your business model is crumbling before your eyes - go BEG the government to "protect" your product!

AI Executives are shitting their pants given the release of Moonshot AI’s Kimi K3. I seem to recall a similar situation re DeepSeek-1 in Jan. 2025.

All the freaking out about "winning and losing" and whining about "distillation" re China are the WRONG things to focus on.

It is a fallacy to think US Frontier Labs would be successful in locking down AI models and charging high prices for access in perpetuity. It is inevitable that open weight models would catch up, and put SERIOUS pressure on pricing.

Welcome to software - It just happens that China is leading the charge on this one.

The magic is NOT about the model. The magic is creating the guardrails and infrastructure (ecosystem) around the model, and supporting practical use cases the model excels at, and making them so good to use that no user wants to switch to something "almost as good", even if it’s a touch cheaper.

Net-net, governments cannot stop software at the border. Import/Export controls will NOT be effective at stopping Chinese firms from distilling US models, even if the USA is successful at starving China of inference, which also will NOT happen. https://futurism.com/artificial-intelligence/ai-execs-quaking-boots-chinese-models #AI #LLMs #Kimi-K3 #Kimi #FrontierModels #USA #China #ImportControls #ExportControls #Software #Distillation #Open-Weight-Models #AIPloicy #Protectionism #AIRegulation #DeepSeek #Software #OpenSource #Moonshot

The #Trump administration is investigating whether #ChineseAI models, like #MoonshotAI’s #Kimi #K3, are using American intellectual property through a process called #distillation. This concern arises as Chinese models gain traction against US leaders like #OpenAI and #Anthropic. The US is considering #sanctions if evidence of theft is found. https://www.cnbc.com/2026/07/21/bessent-china-ai-sanctions.html?AIagents.at #AIagent #AI #ML #NLP #LLM #GenAI

#AIcompanies like #Anthropic, #OpenAI, and #Google are facing challenges from competitors using #distillation to improve their models. This practice, which involves using outputs from one #AImodel to enhance another, mirrors the way these companies have utilised internet content for their own AI development. While they argue this is unfair, the reality is that information shared online is often used in unexpected ways. https://www.businessinsider.com/ai-giants-learn-hard-truth-modern-internet-anthropic-openai-google-2026-7?AIagents.at #AIagent #AI #ML #NLP #LLM #GenAI

#Opensource #AImodels face potential #regulation in the #US, with a ban on #models exceeding the capabilities of GPT 5.5, Claude Opus 4.8, or GLM-5.2, likely within six months. This is driven by concerns about #China developing advanced open models and the potential for model #distillation. Managing #frontier #openweight models requires careful consideration, as a ban could isolate the US from the global open-source community and hinder innovation. https://www.interconnects.ai/p/6-months-to-live-for-open-models?eicker.news #tech #media #news