>Amodei: "Open Source AI LLM Models Will Eat Your Children"Why are Amerimutts and Trumptards so afraid that anyone in the world being able to use Claude Fable 5?Do they not know GLM 5.2 is now almost equally good and is open source?
>>538076678Yeah fuck it. Why not give indians the power to hack people instantly? What the fuck can go wrong?
>>538076766>Why not give indiansIndians are working at all AMERICAN companies at this point.
just don't be a tranny and use Claude
To hype it up obviously>bros our chatbot is so good it's basically a weapon of mass destruction...we can't even show it to you that's how good it is...now please pay $500/month for a subscription
>>538076678Glm5.2 is not fable tier lol, that would have unironically crushed the US economy because it lives on the AI bubble hopium atm It's close to opus 4.8 medium in cost per unit of work, chinks trail us models by ~6months.
>>538076678>GLM 5.2 is now almost equally good and is open sourcenow you just need 2 terabytes of vram to load it
>>538076972It's just a matter of time.Many open source enthusiasts are donating code chat histories to GLM, and other open source models.>>538077141Chinks are really good at mathematical optimization methods.Why do you think American Academia is 90% Asian nowadays?
>>538077395It can be fine tuned by someone w tons of prompt pairs into a frontier one no doubt (prob Elon, already bought cursor). But I'm sure it won't be open weights then
>>538076766Or better yet, why not pay for companies where Indians work to tune your paid LKM, fuck no!
>>538077395I support china or whoever close the gap between what Epstein class can do and what plebs class can do .
>>538077395It's 90% asian because they pay more money than the instate boys. Half the chinese international students can read properly or speak english and cheat their exams.
What local AI should a normie get? Do you have to have more than a toaster to run them?
>>538077852Kek, you are so fucking wrong. I won't deny chinks haven't stolen most their techs advances. But they don't just steal they improve. Just go read the report on how deep seek was developed. They also make it open source. If chinks can figure out how to make their own gpus the US will be left behind.
>>538077952Whats your vram size?
>>538078020Don't really know. Let's assume it's low
>>538078079OK, give me the specs of your pc. Do you have a graphic card? You can run a local model with the cpu. But you'll probably crash your pc or freeze it. If you are doing to much stuff at the same time.
>>538078198Just to add a bit to this it will take a lot of time to think and respond if you run it on a cpu.
>>538078278Yeah sounds like I won't bother. You do need more than a toaster.
It's a ponzi scheme worse than enron. They're also price fixing and price gouging cartels.
>>538076766You know nothing of the industry and are motivated by fear and other emotions.Build yourself into a better human being otherwise you are no better than said Indians.
>>538078278>>538078198>>538077952You have to go buy something asap to get anything worthwhile. What features do you want, ask and ill tell you what you need to buy.
>>538078769It's a giant fucking scam. A bunch of boomvestors got duped by snake oil salesmen and the RAM cartels are conspiring together to fuck everyone while making ungodly amounts of money by intentionally shrinking supply and signing private deals with other members of the cartel. The top three RAM manufacturers are being sued for antitrust violations right now.
>>538078581Don't get discouraged, you can still run something it won't be as advance as the models running from data centers. What do you want it for? Just basic info are you running windows? What version? Or a Linux distro?
>>538078847I literally just made a pc from old pc parts third Gen running 32gb of ddr3. A few newer gpus. I fucked up not going with 4th gen(pcie atomic) because I'm running Ai experiments. Besides training it with the cpu which take way longer. This pc runs a lot things like nothing. Of course the os is not microslop.
