Which one of this two going to fall first? The chinese are pretty hellbent on destroying us ai market by flooding them with open source models. And this two AI labs has nothing to offer other than AI so they can't absorb fallout like other big tech companies.
OpenAI has a chance of surviving because Sam Altman kneels to the US govt and sucks drumpf's misaligned wiener, it has nothing to do with the profitability of either company though which is nonexistent
>>109613451>it has nothing to do with the profitability of either company though which is nonexistentAsk Anthropic
>>109613477Did your LLM malfunction? Are you an ESL or something? What are you trying to say?
>>109613440if ive learned anything from the previous tech "bubble" its that it doesn't have to make financial/logistical/economic sense to still make a ludicrous amount of money decades beyond what you thought it would.
>>109613440neither cuz open source ai is inferior and doesn't have trillion dollar jewish companies backing it to provide marketing and infrastructure that makes a successful productI think OAI at this point has a better chance of blowing up than Anthropic though, if either of them do
they can both become profitable pretty fast if they stop investing so heavily, which sounds stupid but it means neither are likely to ever disappear but you won't get a new fable or sol release every few months anymore
>>109614052>they can both become profitable pretty fast if they stop investing so heavily,Just plain inference is more expensive than what they're selling it for
>>109614078their bulk of hardware is focused on training, it becomes a lot cheaper when that gets repurposed towards running existing models not training new ones
>>109614085>implying that hardware existsWouldn't be """pretty fast""" either.
>>109613440I am trying to prop up Anthropic. Currently rewriting all the old chan sites just for fun.
>>109613440None. Both are here to stay, keep seething and coping ludder
>>109613440>And this two AI labs hasHoly low IQ
>>109614112xAi made a mistake building Colossus 1 and ended up with heterogenous cluster of H100/H200/GB200 GPUs. This is a nightmare to train huge models on because each card has different specs, features, and hardware requirements. During gradient synchronization, a heterogeneous cluster would bottleneck on the slowest GPU (H100) so the faster GPUs would end up idling. They also probably ran into unexpected compatibility issues, which are difficult to resolve.It makes more sense to use this cluster for inference, since they can segment the cluster by GPU type and avoid GPU mixing. xAI doesn't have enough inference customers so it makes sense to monetize this to companies that need inference compute such as Anthropic.
>>109613498Anthropic is already profitable mullato. Is that clear enough?
>>109613440AI economics make no sense but AI has been picked up by regular tech company CEOs now, so they're putting all their money on it. They're subsidising us for supposed 10x developer gains in hope of replacing their employees. We should enjoy while it lasts because it's about to get a whole lot more expensive.
>>109614204time to pay their debts then
>>109614201The way they network their datacenter is just a minor factorThe bulk cost is nvidia silicon and hbm ram. Both are required for training and inference to the same amount and both are in a supply shortage.So that's why inference costs almost the same as training.
>>109614245training is far more demanding, it just is from a technical standpoint
>>109613440Anthropic is the Apple of AI labs, so it will survive even though it should be buried and forgotten.
>>109614271funny, a couple years ago OAI was the apple of AI labs
>>109614271Microsoft invested in Apple and OpenAI.Google and Amazon invested in Anthropic.
>>109614298didn't apple switch to Gemini anyways
>>109614302Apple isn't acting rationally but is very pissed at OpenAI for poaching Apple employeeshttps://www.reuters.com/legal/litigation/apple-sues-openai-alleging-misappropriation-trade-secrets-court-records-show-2026-07-10/
>>109614233It's circular economics.>Google spends X billions on anthropic shares.>Anthropic spends most of that renting GPUs from google, massively boosting their revenue>Every time anthropic sells shares, it's for a higher price, so google marks up the value of their old shares>This is counted as pure profit for google>repeatIt's a tangled web between many more companies but this is a example to explain the circular funding. Now oAI will probably go bankrupt first because they are slightly behind in model development but they are the ones with the most low-end customers that just cost them tons of money.To stop burning money they have to remove free tiers, 20x the subscription price and ~3x the API cost and that will tank confidence.
>>109613440scam altman *won* thisanthropic are a bunch of nerds scamming enterprise for revenues. they already can't compete when claudes get pitted against models of same weight class. basically their training sucks, they don't have the sauce. they was winning just by being ahead of the curve in model scalingmeanwhile scam altman already went all in on hardware and infra investment for a long time, back when people had doubt on AI
>>109614381>can't compete when claudes get pitted against models of same weight classnice cope when no one else has a model comparable to mythos
>>109614381>scam altman already went all in on hardware and infra investment for a long time??oai had an exclusive agreement with microsoft till like last year. the only hardware and infra oai were allowed was microsoft's azure.
>>109614429>weight classread it againthere's no model as big as mythosmeanwhile, sonnet 5 release was a disaster, opus getting raped by gpt for months. opus 5 is much newer than the heavily delayed Sol yet isn't ahead at all. anything below mythos, claudes have been slaughtered by gpt and chinksI wouldn't be so sure anthropic could still compete once other labs release bigger models
>>109614444last year was a long time ago
>>109614052Or they could drop the censorship. They'd rather go under than do that, though.
>>109614845you can just fill out this form to drop censorshiphttps://openai.com/form/enterprise-trusted-access-for-cyber/
>>109613440Anthropic GPT has the most normie adoption and its serviceable enough for enterprise and more investmentClaude has a reputation as "the best" but its also the most expensive and Anthropic is the sketchiest of the AI companies. They seem the most cultish. China will steal market share in enterprise due to cheaper models, GPT will keep the normie chatbot crowd
>>109614926i've heard people gushing over supporting anthropic for being wholesome and standing up to trump (literally hitler)We might have the stupidest timeline where politics props both up
>>109613440>this twoSaar do not redeem the ESL
