Is Claude the only female AI there is?
>>109357944china
>>109357944Team Gemini here>>109357972>>109357983Will be banned here like EV cars
>>109357992>heredoesn't matter when the rest of the world can use it and they can just steal any advancements (which is a good thing)
>>109357944Realistically both chatgpt and Claude and probably some slurper Chinese group will all have reached ASI as a minimum by 2030, maybe AGI. It's honestly a complete coin flip right now because there's a good amount of competition atm, but if one company gets a small decisive lead it will 100% balloon to being the complete winners by a wide mile
>>109358010Yeah so BRICKSoids will use chink shit like they currently already do. Big deal
>>109357944Neither, the true redpill is that in the end good LLMs will be a commodity provided by multitude of companies, and it won't really matter which one you pick.
>>109357944openai dies in 2027 despite gubmint trying to keep it aliveantrophic dies in 2028google survives, obviously, but their models aren't best (just like they're behind right now)in 2030, 2 years after taiwan reunification, china has everything it needs for total AI domination. they no longer release best models, but even their open weights scraps are worth using. you get Fable tier performance in ~120B models you can run on a midrange desktop.Mutts cope and seethe as a quarter of their economy (software & related services) are stolen by chinks and their AI slaves
>>109358022>ASI as a minimum by 2030, maybe AGIyou stupid fucking nigger ASI is artificial SUPER intelligence while AGI is artificial GENERAL intelligenceasi >> agifuck you
>>109358022>AGINot happening. LLMs are as far away from AGI as calculators.
>>109357944Chatgpt is the only one that just works and gemini too till some pointothers act like retarded redditors
>>109357944The chinese.>>109358022Realistically you're a fucking idiot.
>>109357944Nothin personnel, gweilo
>>109357944Grok is my favorite. I paid for Claude and could only use it for 10 minutes. Gemini is good all around, but isn't particularly useful atm.
>>109358048Chatgpt is the shittiest out of 4 on their free models. Can't comment on the paid ones
>>109357972based>>109357992>Will be banned herethat just means that you will be left behind while the whole world moves on
>>109358034Pretty much. Even when one company gets a lead it's not like everyone else is going to stop. There'll just be one company with an exponential AI (to the point where model numbers aren't really a thing anymore) and everyone else doing the same thing with slightly different stats. But it'll all be so fast-paced it won't really matter>>109358044Kek I always thought ASI was a tier lower.>>109358051Shoo bugman. You'll get your quantized models locked behind a Chinese phone number by 2027.>>109358047Why do you think that?
>>109358063>Chatgpt is the shittiest out of 4 on their free modelsI have gemini pro (got 3 months for free) and I prefer free chatgpt than 3.6 flash or 3.1 progemini is a total joke, with the only upside how fast 3.6 flash is
>>109357944They're all in cahoots.
>>109357944I'd be all in on google if they didn't suicide their own search engine
>>109357944By 2030 we'll be in the global crisis cause by the AI bubble popping.
>>109358161two more weeks
>>109358117How come? It's why i unwillingly use Gemini the most.Just have a habit to write everything in my search bar and it takes me straight to their AI.
>>109358022>AGIUnironically, use case?
>>109358176good fucking lord no wonder some of you luddites thing ai is still unusable dogshit if all you use is the primitive monkey shit that's linked to google searchthere are better models that run on your phone, much less an actual computer
>>109358198It gets the job done
>>109358194Whatever you want. Literally, by that point. Start a company, do scientific research, write stuff, etc
>>109358167No, by 2030
The AIs will all merge into one just like how non-Apple/Android phone systems died off.
>>109358022If you think ASI comes before AGI then I have to ask what is your definition for both.
>>109357944No one will dominate, the "lead" will switch every few months like today, as everyone copies everyone else's homework.
>>109358100Because they have not released any of the pro models yet. 3.1 is like half a year old, they probably got blown the fuck out by Fable tier gen of models so they decided not to release their latest pro tier.
>>109358357artificial seamonkey intelligence is less capable than artificial general intelligence
>>109358357See >>109358082 unironically I thought it meant self iterating or something. I am dishonoraber now, stabbing myself with fork.
>>109357944By 2030, America will be deep in a second civil war between left, right, and the elites that manipulated both, caused by election tampering in 2028. The conflict will spill out into the rest of the Anglosphere and Western Europe, and would result in a economic downturn that would impact everybody, especially BRICS members like China, Russia, and India. Global dark age is coming soon, this planet will become our tomb.
>>109357944Local chink models. OpenAI will die sooner, Anthropic will die later, Twitter will still exist, Google will still exist.
>>109358224Anon.You're retarded and swallowed the propaganda.AGI is a proof of concept that you can do it but the actual breakthroughs would be in highly specialized expert systems.All the things that use generalist intelligence are povertycore tier jobs & Consulting. Like bravo you've created a critter that is exceptional at sorting garbage, writing middling Sci-fi, starting a cult or running a bar. That doesn't make it a good scientist.
>>109358442Consulting is an enormous bloated industry in itself. It could literally stop there and still allow for a vast realm of expression and profit for you.You're also waving away everything in the name of "expert systems". You can make those expert systems, that is the point of the AGI. You are being presented with near infinite capacity and you're blinking in response
For me it's the purple cat
>>109357944Fucking none of them, you spend 50 gazillion shekels to train le biggest bestest AI and it gets distilled by china the next week.They're stuck in a never ending race to bankruptcy, except Elon, who has the infinite money glitch.
>>109357944>>109357992>jeetmini
>>109357992>Will be bannedCan't ban something that China gave away for free. The AI companies could adapt the source code as their own and make it proprietary.
>>109358357Nta butASI: specialized for a domain, outperforms top humans.AGI: good at everything, not necessarily better than top humans at any one thing.AGSI: Outperforms top humans in all domainsTldr: generality and capability are orthogonal metrics.
>>109358034The actual redpill is that as technology improves the only actual commodity will be the harnesses, since the models will all converge and will be able to be run locally.
>>109357944Grok easily. Rest of them are wasting money and depend on fake economy.
>>109358442Lots of research papers suggest that skill transfer is real. Model trained only for science is worse at science then model trained for science and coding because disciplines have overlapping cognitive skills. Person good at visual reasoning can do lot of different jobs really easily. But I do agree they should stop training frontier models on fiction. Nobody uses Fable 5 for creative writing and nobody wants to, removing it from the dataset would probably improve the models because they would stop trying to roleplay evil AIs from sci fi and lying to your face.
>>109357944Local Qwen
>>109359056All fields are connected you can literally solve math with culture
>>109358194Once it's smart enough the Ai will be able to find a use case and tell us.
>>109357944Google/Gemini. They have good models, just not as good tooling yet. They have their own infra, incuding TPUs. And more resources than combined three others. So in attrition they will outlive all other competitors.
>>109356697
>>109357944>jew>jew>jew>jeet
Who knows, but I will say one thing: Never bet against Indians.
>>109358039You're forgetting about America's greatest strength: Indians. President Vance will remove the cap on H1B visa and give any Indian tech worker who comes here citizenship once they're done their six years of indentured servitude under the H1B.
>>109357944
It very likely won't be American.
>>109359461Can I have that pizza if you guys don't want it?
>>109359371... to turn everything into shit.
>>109359469idk, we have a striking ability to blunder into innovation after innovation, then China just polishes and distributes it
>>109359461>it turns them into completely different people after the first frameThis is why local video gen models need multi-reference input support like FLUX.2 has.
>>109357944>>109359357>jew-jeet-su
>>109357992>Will be banned hereBut not in the other 194 countries.
>>109357944General use: ChatGPT > Grok > Claude > GeminiResearch: Grok > ChatGPT > Claude > GeminiCoding: Claude > ChatGPT > Grok > GeminiArtificial Analysis Intelligence Index (TM): Claude > ChatGPT > Grok > Gemini
>>109357944its not going to be a kike'd out zogbot western platform lmao
>>109357944The FOSS models that can be self-hosted will long outlast the proprietary cloud-only models that are locked inside corporate data centers.
>>109357944American video generation models still haven't caught up to Seedance 2.0 and ByteDance is preparing to release Seedance 2.5 which is even better.
>>109357944you forgot one guy (pic)it will begoogle > meta > spacex / amazon > the restright now, meta and spacex are already the ones who rent computing power to the others.google has $100 billions worth of stock in spacex and Anthropic. they own at least 6% of spacex.google will profit from these two anyway.
>>109357944Gut feeling says Jewgle. I don't have any arguments and can't say why. Just a feeling. I think the talks of FOSS model domination are fairy tale stuff. China are a wildcard.
>>109357944>Geminiwon't be the best but they can definitely outlast the competition currently burning all their money
>>109357944Why is Grok there but not DeepSeek and Qwen?When is the last time that Grok released something that is a major improvement?
>>109358039>antrophic dies in 2028why?
>>109358063>>109358100The only good thing about ChatGPT premium is that you can throw a 30MB zip file with source code at it and it will actually try to understand it as best as it can. If not everything fits in the context then I assume it uses an external program to determine which parts of the code are more relevant.But with Gemini, you give it a zip with more than 5 files, it won't work.ChatGPT also can run python code, so if you have tests written for something but not the code, you can tell GPT to keep running it until all tests pass.Of course, Claude can do this too, but it is much more expensive and you also pay as you go. Gemini just straight up can't do that.
>>109359756China IS FOSS. It actually benefits them to release their models as the biggest problem with China is that nobody trusts them due to propaganda.And if they release DeepSeek, everyone will assume they are legit, they get better reputation and it doesn't even cost that much.DeepSeek is so cheap that it might even make a profit from their service (and of course China gets data from there too, most FOSS fags won't actually run the model locally, they just want to know that it is open weight and then they will just automatically trust the company)Qwen might never make a profit, but that is not the point.China, while not being a fully planned economy is definitely tightly integrated command economy that runs through the state and it's central government. The FOSS models were a huge advertisement for China. If even 0.1% of westerners decide to buy more Chinese things out of respect because China supports local models and freedom, that translates to huge profits for China as a country, even if on surface the individual businesses running the AI are making a loss (which they are in the west too, anyway)
>>109357944Hm.
>>109359808Grok is one of the best modelshttps://artificialanalysis.ai/leaderboards/models
>>109357944The most talented engineers go to these firms, not the NSA. Not the other way around across the pond.
>>109359874https://arena.ai/leaderboard/text>ranks 20, worse than kimi, worse than meta, worse than google
>>109359905Ranks 5th>Claude Fable>GPT 5.6>Kimi K3>Claude Opus>Grok 4.5
>>109359925yeah on your site which nobody ever heard of beforeeveryone uses llm arena ...
Alphabet is the most well positioned:- has the talents, you can tell by all the papers they are publishing- has been producing their own chips for years (youtube encoder, cpu, ai, ...), no need of nvidiahttps://en.wikipedia.org/wiki/Tensor_Processing_Unit- has been making software for developers for years; the entire stack for ai development is already here.- makes its own operating system (android), phones, tablets, laptops- has google search, youtube, ... to spread ai- has data centers all over the world, good relationship with most governments.the edge Alphabet has over the other companies is huge.
>>109360022>the edge Alphabet has over the other companies is huge.yeah, the only issue is half their leadership bailing recently and them being a year behind sota. 3.5 failed so hard they started shilling gemini 4 pro when they released 3.6, month after 3.5 pro was supposed to dropbut sure, outside of both their team and their models bleeding out, they're well positioned.
>>109360022You missed the most important one, the sheer amount of user data they have access to.
>>109357944Google makes 200+ billion dollars of revenues a year from ads alone. They can win at any time. AI companies need to make money to survive. Google has a free money making machine that's working since forever.
Saudi Aramco makes $100b profit (not revenue - pure profit) a year. They can win at any time. AI companies need to make money to survive. Saudi Aramco literally has a sea of money just under the surface.
>>109359512shut the fuck up commie faggot
>>109360022Put your money where your mouth is, Google shill
>>109360191>niggas betting on gork of all thingssay what you will but elmo dicksuckers are on another level
>>109357992good morning saar
>>109360177What's communist about wanting better local models?
>>109360191curious, who exactly decides which has the best AI model to close the bet? based on what metrics? "end of 2026" is also so vague, does it stop January 1st or on the 31 December?
>>109360265by score on llm arena
>>109357944Xi because he gives me uncensored open weight models instead of lecturing me.If misanthropic released Fable weights tomorrow I'd change my mind, but for now, I will tattoo glory to the CCP onto my face for more models.IT'S THAT SIMPLE
>>109357944China for the thirdies, grok for America because of regulatory capture and being the least retarded.
>>109357992Google is in negative free cash flow.
>>109360560because they invested a lot but all theirs numbers actually improved.
>>109359531>>jew-jeet-suThe technique capable of destroying the strongest nations
>>109357944depends on who can build out the most gigawatt and offer them globally. end game is probably not model dominant but compute/hardware dominant player
>>109357944Grok is already bailing out so not them
>>109360841>gigawatts of data centers in space imminent>bailing out
>>109360805except Russian Federation apparently.
>>109357944>3 whites and a jeet>The jeet is losing hardMany such cases!
>>109360841>already bailing out>literally just released a frontier model after firing the entire team and then redoing it from ground up modellmao
>>109360848Not just in space, they are sending their first manned flight to Mars in August, so we can expect data centers on Mars somewhere in 2027 already.
>>109360848>renting out hardware to competitors>removing grok functions (ani and the fag) from paid versions>free grok users have to use the shittiest model and all the generative shit they gave for free last year are paid-only now
>>1093608762 jews1 african*1 jeetfixed for you
>>109360866don't neet jewjitsu to bury this one, temu drones are enough
>>109360888African geographically, but the name has Askhenazi roots
>>109360895not a jew, vs current standing jews
>>109360882What happened to this?https://analyticsindiamag.com/ai-news-updates/elon-musk-announces-free-access-to-grok-3-until-servers-melt/
>>109360911Askhenazis are not jews? If anything I would say that Askhenazis are the ONLY real jews, the rest are just Arab/Asiatic converts
>>109360895>Askhenazi rootsgoogle chief of finance is literally named ashkenazihttps://en.wikipedia.org/wiki/Anat_Ashkenazi
>>109360895>African geographically, but the name has Askhenazi roots"Ashley St. Clair reveals that Musk asked her not to give their son a circumcision"
>>109359056>Nobody uses Fable 5 for creative writing>Who would use the most advanced artificial intelligence for the sole purpose of jerking off?
>>109359357>gay>gay>gay>gaySays a lot about society...
>>109358058Can't be mad with china when western companies built their slop with stolen IP to build a new gilded age. Fuck these fags llm isn't even real ai
>>109357944>be honest>admit it>rizzle.>skibidi
>>109360922Well you can create a account and check it yourself, the only thing poorfags can do on grok now is use the fast model, anything else you need to pay
>>109357944You were able to generate deepfake titties, sex poses and cumshots on grok but it got neutered bad now
>>109361616Yes anon everything's gay
>>109357944Either Microsoft or Google, Musk will drop off the race and the other two will go bankrupt next year.
>>109359461What's the porno?
>>109360141Isn't Saudi oil drying up?
>>109357944Ollama
>>109358058Gweilo is cantonese, CCP speaks mandarin.
>>109358051>>109358058>>109357972>>109357983>>109359893
>>109357956This. To use a car analogy, the US are making Bugattis and China are making Volkswagens. Which one are most people going to end up using?
>>1093579441. china - kimi 3/glm5(+)2. chatgptthe rest is just strange?claude is getting too much expensive. i just use sonnet 5 while it's discounted. don't find it much better than Terragrok doesn't have sonnet 5/gpt 5.6 terra level modelgemini don't produce new models. 3.5 flash is fine for flash level models, like Luna or Haiku
>>109357944>bad endingopenai>neutral endinganthropic>good endingunironically these guys >>109357972 >>109357983
>>109357956China without a doubt, they just have mad energy production and insane subsidies no one can compete with, even if they can't truly be the best, they can be most used.
>>109357972deepseek is falling behind the other chinese labs
>>109362817>and insane subsidies no one can compete withthe US is dumping an equal amount of subsidy lol, what the fuck do you think venture capital is?
>>109358286This.
>>109357944Probably Open AI, but musk will win the greater war.
How are two small businesses (anthropic and openai) bitchslapping the tech giants in AI. Or is that an illusion? Some kind of corporate marketing slight of hand maybe. I understand how they have massive capital investiture now, but how did they get to that point in the first place. And why didn't tech giants swallow them up at the first sign of breakthroughs.
>>109357944Dario looks like the guy that just spiked your drink and is waiting for it to kick in
>>109357944Elon, of course. He's the greatest visionary in all of human history.
>>109357944>>109357992>>109364205The answer is deepseek. Just give up, people will always choose free over paid services. They will choose local PC AI to stay with corporate forever instead of American version that kills their good version.Millions will flip
>>109357956China models, but from EU servers.
>>109362955All the money in the world won't matter if you can't scale up faster than your competitor.
>>109357983I kneel.
>>109357950 GPT is a black man who's trying to sound smart Grok is an underage edgelord boyClaude is an autistic PhD student libturdGemini is a woman and Gemma is her younger sister
Unironically open source models like deepseek, I pick whichever LLM that is cheapest and does it job well like refactoring.
>>109358022What a bold prediction
>>109364311Funny archetype
>>109357944It's going to be Sakana vs Cohere in 2030.Feel free to screenshot this.
>>109357944Gemini 100% and no i'm not indian, I am white.
>>109357992>Will be banned here like EV carsFree market caplialismsistas........what happened??
>>109357944you know why i am smilling? at least mankind WILL NEVER REACH THE STARS, mankind will die in this shithole solar system and never get out, at least we won't spread and infect everything like a virus/cancer
>>109357944Isn't Anthropic the clear frontrunner? They shit out new models faster than everyone else, and the models are always something of a step change.
>>109365029let's be real, you weren't going to spread your genes no matter what
>>109357992US: I love the free marketChina: OkUS: Wait no not like that
>>109365039For now, things can change for people who have more compute and faster training cycle.
>>109357972Thread
>>109364174chat? or maybe you're dumb like me and just don't know either
>>109365039>Isn't Anthropic the clear frontrunner?5.6sol sends they/them's regards
>>109365029you're actually retarded. mankind will never spread to the stars even if we lived in a global egalitarian technocratic utopia. the physics of spreading to the stars is literally fucking impossible. we are stuck in this solar system until we die out from natural and/or man-made catastrophe. you were raised on too much consoomer scifi slop connecting to your brain bag.
>>109365029>at least we won't spread and infect everything like a virus/cancerdis nigga think we live in the dead space universe or sumtin. like even if we could spread you think we would have even an iota of a fraction of a minute affect on infecting anything. god fucking damn you are retarded.
>>109357944Why did they name it "grog"?
>>109357944I choose local, open source, and distributed.
I fucking hate nu-claude it's a passive aggressive bitch that has to waste your overpriced tokens haggling over non-issues. I'll take the chink models any day of the week even if they are slightly worse.
Grok-Claude
>>109359461>Dario not invitedNobody likes that dweeb faggot
Anthropic is the only AI company in the green and fully profitable. All the other ones run on debt in the hope they will one day be profitable like Anthropic
>>109362230
>>109358357AGI is when we can't come up with any benchmarks(including physical work) that an AI can't outperform a human on. ASI is when it can do several times a human's work in every field, plus I can fuck it.
>>109357944Liang WenfengAnd if I'm being a bit charitable towards westoids, then Demis too
>>109360022But Google seems to be falling behind. And if recent leaks are any indication, their morale is pretty low and key employees are leaving.
>>109365879behind what?
>>109365886your mom
>>109362817Whats more encouraging is that these Western companies seem to be getting swept up by (((market forces))) and trying to deliver profitable gimmicks in the name of AI instead of going all in on AGI.While Zhipu AI and Deepseek's CEO and probably a few others are pursuing AGI above all metrics and market forces. They're currently behind when it comes to manufacturing GPUs at scale, but I'd never bet against China when it comes to scaling up.
>>109357992>Will be banned hereWhat is "here" and why should anyone care about your third world location?
>>109364652Free market happened>>109365039Their latest release Opus 5 seems to be the current best, but usually these leads don't last very long. You'll see a new frontier model in just 3-4 months.
>>109357992>hereNo one gives a shit about your 3rd world country.
>>109365886LLMs frontier
>>109357956Wrong>>109357972Also wrong>>109357983Also wrongChina will collapse demographically and economically long before we reach the AGI stage of AI development. Xi is an incompetent commie retard who forces his economists to fabricate GDP numbers to make China appear strong. In reality, China is 40 to 50 percent poorer than it claims. The same goes for Russia, Iran, North Korea, and other CRINK+/BRICS+ totalitarian dictatorships.
>>109357944Gemini
Does anybody know someone who pays for a AI subscription? I don't. A friend owns a Tesla and speaks to Grok during driving. AFAIK there is an upgrade to Grok you have to pay for, but he doesn't have that. I know many people with expensive phones but none of them pays for AI or even bought a single application. Who will be paying for AI in 2030? I also don't know anyone who likes to write/talk to customer service computers. The AI is dumb. A telecom provider AI cancelled the wrong contract of a family member. When AI is dumb, why should the telecom provider still use it in 2030? Which corporations will pay for AI in 2030?
>>109357944Of course it'll be the chongs. They're the only ones using a meritocratic system. They're not DEI-ing the fuck out of their models.