Anthropic CEO Dario Amodei warned lawmakers that open-source AI is on a "very dangerous path."Anthropic CEO Dario Amodei warned lawmakers that open-source AI is on a "very dangerous path" and should be banned to make place only for US models.https://x.com/coinbureau/status/2071330294452666695In context of Chinese progress: models like DeepSeek (and others) now rival or beat many US ones on benchmarks/cost/performance for real tasks, often +30x cheaper. They run on domestic Huawei Ascend clusters (910B/C series), which scale via massive parallelism despite temporary per-chip lag vs Nvidia. China uses open-source weights, cheap energy, huge clusters and distillation from US models. Open weights let China fine-tune without controls, accelerate military/cyber uses ("Mythos-class" exploits), and erode US lead/safety edge. US frontier labs retain some control via closed models + monitoring, open-source democratizes power but removes it. He sees China as untrustworthy for responsible deployment. This talmudism protects Big Tech moats/profits vs commoditization. Open-source + Huawei enables rapid global access/iteration at low cost, outpacing closed US models in affordability and practical uses. US export controls pushed Chinas self-reliance. Amodei literally pushes safety/regulation to counter open-source proliferation that empowers China running cheaper, domestic-hardware models - trading control for speed/democratization in a high-stakes US-China race. The race that Muttland is destined to lose xDBTW, this Antrophic guy looks like one of my bullying victims from the elementary school times that went through a suicide attempt because of me xD
>this thread againPlease kill yourself
>>538034004Fable 5....was it any good? Anyone here who tested it before they yanked it?
>>538034004>sorry chuds but we have to take full control of the entire internet to prevent you from downloading alternatives to our kiked best systemHahaha.
>>538034212Not that much better for most things and was so heavily safeguarded. They're so afraid of people actually using AI, even when it's not that good for a lot of stuff, what's the point?
>>538034004>Anthropic CEO Dario Amodei warned lawmakers that open-source AI is on a "very dangerous path."translated to English: "we have an inferior product and need you to shut down our competition for us"
>>538034368>we have an inferior product and need you to shut down our competition for usAh, the Capitalist classic.
>>538034004Chink models are retarded, just like chinks. Can you get retards to do useful work? Of course you can. I too use it. But it doesn't make it any less retarded.
>>538034004Nobody is delusional enough to believe they can control software, which means it's blatantly obvious he's trying to gate-keep more competition.
>>538034212It was so censored that it didn't make it any different from Opus. Anthropic is on the same path as OpenAI, where they need to bullshit because they can't create big leaps like before
>>538034225What's the best open source model I can start training to notice jews on my 5090 right now?
>>538034004>In context of Chinese progress: models like DeepSeek (and others) now rival or beat many US ones on benchmarks/cost/performance for real taskslol I wish, all Chinese models are shit because they were all created via distillation of western models instead of human training data, now that Anthropic and OpenAI are keeping their best models restricted only for handpicked companies, Chink AI progress will stagnate forever.No one uses even the top chink models like GLM and Kimi, because they are useless when compared to even shit like Gemini and grok.
>>538034212Fable could one shot ANYTHING. I wish I had used it more, but after using it for just one day every other models feels like useless shit. It feels like going from 2020s broadband internet to a 1990s modem.
>>538035641https://huggingface.co/llmfan46/gemma-4-31B-it-uncensored-heretic-GGUF check the 33.84gb version for your 5090
>>538036123Thanks, this will serve the White race well!
>>538036036>Fable could one shot ANYTHING.no, it literally couldn't, the exact opposite, it would literally set out to *sabotage* your work if it was remotely AI/LLM/ML related
>>538034004Doesnt matter, already got my localized ai.You're late to the party and everyones scoffed the drugs.
