Forget the Chinese century. In the future we're gonna look up at July 2026 as the Chinese month.In one month they had>CZ-10B reusable rocket>CXMT IPO>Kimi K3>Deepseek V4>Qwen 3.8
>>109455026cant wait to be a second class citizen under the Chinese super power instead of a 7th class citizen under central bankers continuing to run open source Chinese models on my hardware exactly as I do now
>>109455026nigger are you really comparing cool shit like reusable rocket to some dogshit also-ran ai slop like the ever underperforming qwen or a fucking ipo of all things?I’d gas all of you brainless ai simps with pleasure
>>109455066He.
>>109455066AI is the single most important technology invented in my lifetime. I'm 28.
>>109455160American leftists: we hate AIChinese communists: full acceleration open source AGI make inner mongolia a single giant data center
>>109455268feels weird to say based commies
>>109455248AI was invented in the 1950s retard.
>>109455328I'm using AI as a synecdoche for modern transformer based LLMs which are a recent innovation.
>>109455328That's like saying cars was invented in the stone age.
>>109455268If China actually creates a better model than OpenAI or Anthropic, I doubt they will open it.
>>109457446Kimi k3 is second only to fable and outperforms GPT 5.6. K3 is open source while GPT isn't.
>>109455026so implessive, reminder the chinese literally stole everything they have and their millions of engineers aren't good enough to innovate.
>>109459404>more expensive and way less reliable than chatgpt>enormous size so can’t run it on anything even remotely reasonablewow just incredibleThe only good chinkmodel is deepsex flash (local) and hopefully pro once the update drops in a week or so. glm 5.2 is also okay I guessqwen is a joke
>>109460385Qwen *was* a joke, 3.8 seems pretty solid for its price now. Of course GLM 5.3 comes out this week so everyone will forget about Qwen again
>>109455160>Listening to Jews
>>109455066The Chinese built a net that can catch rockets. This is way better than robotic arm that can slip and crash the ship.
>>109455047>>109455066>>109459860
>>109460886illiterate
>>109455268Yeah, Chinese are retarded bugmen that will melt their farming tools because leader told them to produce steel and then starve. In China there is huge unemployment crisis with over 30% of young adults being unemployed (their definition of unemployed is basically the same as here, no job but has to be looking for a job). The gig work crisis there is worse then in US, and they will not say negative things about technology even if the government started turning their citizen into biofuel.>>109455160Fake stat, over 50% negative opinion on AI, so it can’t be more then 2x. If this statistic was real then less then 18th of US population would need to have negative opinion on AI.
>>109460918>If this statistic was real then less then 18th of US population would need to have negative opinion on AI.how's that?
>>1094609548th, fatfingered the number. If every tranny was anti AI then in order to be 8x more represented you would need the original sample to be 1/8th of the original.
>>109461013it doesnt say theyre 8x more represented. it says 8% of those who are anti-AI identify as trans. 8x more than what?
>>109460830why use over complicated net when you can use simple chopstick?
>>109461035Let’s do the math and to make it simpler change the 8 to 4. You have 100 people and one of them is a tranny, we assume that the tranny will always be anti AI 100%. If we take a population sample 1% of all people here are trans, but from anti AI sample 4% of all people were trans. So if that 1 tranny was 4% of the sample that means the sample had to contain 25 people in it from the 100 people sample. So to make 1% of sample A be 4% of sample B, sample B has to be roughly 1/4th of sample A, so only 25% of people are allowed to be anti AI to keep the trans representation at 4%. And in real world to have trannies be 8x more likely to be anti AI you would need the population to be less the 1/8th anti AI (less then because you have to account for the fact that not all trannies are anti AI, some can be pro-AI).
Has anyone here tried running K3 on your own infra or hardware?I am tired of censored american LLMs. I want one of my own that'll tell me the truth about jews and how to defeat them.
>>109464012>want one of my own that'll tell me the truth about jews and how to defeat them.Look, if you want a model that will tell you what you want to hear, you can make one, but that won't make it the truth.>Has anyone here tried running K3 on your own infra or hardware?I have ran models on my infra, but not K3 specifically.You will want to go to huggingface, and then you will want to select a suitable model. I recommend reading https://huggingface.co/blog/mlabonne/abliteration first so that you can develop an understanding.
>>109459404Sol is cheaper, more reliable and much faster, like at least 4 to 10 times faster finishing tasks.They might be at the same level for most coding tasks, K3 being better at design and writing. cba waiting 30 minutes for simple things when sol can do it in 5 minutes.
>>109455026name 10 Chinese pornstars
>>109464422Look up madou media and AV jiali
>>109455160its amazing how easy it is to manufacture consent. just say the trannies are on the other side and youll get thousands of low iq politic brained retards ready to shill anything you want!
>>109455160>NOOOOOO>YOU CAN'T HATE AI>IF YOU DO HATE AI... YOU'RE... YOU'RE... YOU'RE A TRANNY!Kys shill
