*vindicated*
Dario won. Super Intelligence won. Zitron is a seething timmycel whose mumma has onlyfans
>>110136596AI companies have completely overshot their revenue targets that they made in the past. Why are retards suddenly pretending to be geniuses for pointing out that they’re not immediately making profits? That was literally the plan all along. But the Luddites have been very wrong about the revenues and the increasing capabilities, so they’re in for a rude awakening even in the next 2 years.
>>110136877>>110136893i see Indian Defense Force is here. Hi Saars!
>>110136893The risk index for being wrong here works like this:Being pro ai and wrong: ???VsBeing Luddite and wrong: just stop snailing and ai will help whatever you're working on.
>>110136930the real risk: saar loses his cushy shilling job and does not redeem
>>110136930> Being pro ai and wrong:some trillion usd lostand "wrong" doesn't even have to mean that AI is never useful, it means that the investment they made into OpenAI or Anthropic doesn't pay back even more trillion USDcurrently it's not looking good at all. trillions of usd in gpus sitting in warehouses because there is either no demand or data centers can't be built fast enough, openai and anthropic and the whole ecosystem using all sorts of accounting bullshit numbers and circular financing trickery and so on
>>110136893>the Luddites have been very wrong about the revenuesanthropic's revenue growth is flatlining >he increasing capabilitiesall harnesses, which any local model can slop out in a few hours with even better features, especially when you consider a lot of them are open source: model intelligence is a meme in late 2026
>>11013697516 billion in a quarter is pretty damn good. Can you guess how much revenue Ed Zitron thought Anthropic would NOT have for 2027? I’ll give you time.…The answer is:34 billion. He said this was “ridiculous”. And they’ve already made this amount through the first three quarters of 2026. Not to mention he thought the bubble would pop “no later than Q2 2026”. So I ask: “vindicated” for what, exactly..?
>>110136975>anthropic's revenue growth is flatlining That's a nice way to put it, picrel first half of the year is a 80bn loss. Oh and the last year was a 42bn loss.
>>110137018>16 billion in a quartersee >>110137023the whole year 2026 will be a disaster again, won't it.
>>110137023ant's rev as a multiple is up more than their loss as a multiplethey'll be fine and they just have to get to middle of the next year and the god-claude will create the medbeds that will make them trillions
>>110137047when your business strategy relies on a miracle that's not a very good business lol
>>110136966
>>110137023They didn’t literally lose 80 billion dollars, that just represents the value of the shares from their investors that increased because of the company’s valuation. You have no idea what you’re talking about. The truth is that their revenue is growing faster than their costs.
>>110136966>data centers aren’t being built fast enough and we don’t have enough power therefore AI is a scamdo you even hear yourself?
>>110137062it's been working out pretty well for them so fargo back 5 years and claudecode would have been a scifi miracle at least 30 years awayat this point to say that they won't cure cancer is the position that needs to justify itself
>>110137047kek - their loss is a multiple of whatever they made last year at 42bn lossand as far as anyone can tell the same this year again with the 80bn loss the first half year
>>110137092>at this point to say that they won't cure cancer is the position that needs to justify itselfthis level of delusion is something else lol. What do you fags smoke to post shit like this?
>>110137104give me 6 reasons why they won't cure cancerprotip: you can't
>>110137108You are curing the cancer? Are you grappling with death? Do you know the mythology of panacea? Once you cure one disease, 2 more come into existence. Does anyone know that story?
>>110137108bro you should cut your dick off, ai will fix having a small pp.give me 6 reasons why they won'tprotip: you can't
>>110137108>>110137076
I can't wait to see this man actually win. He's done the numbers. All the experts way more experienced and qualified all agree with him. It's the end for Effective Altruists.
>>110137068as a business genius i propose theropenanthropic corporation, it has -1000bn but look how steep +9999.9999 bn is the next quarter after "reaching profitability"even better than -80bn bookkeping in q1/q2 followed by +16bn in q3
>>110137147tb h i did start smoking and doing heroin againfeels great man
>>110137169abs(>as a business genius i propose theropenanthropic corporation, it has -1000bn but look how steep +9999.9999 bn is the next quarter after "reaching profitability">even better than -80bn bookkeping in q1/q2 followed by +16bn in q3)
i use ai and i agree with him lol. these corpos are fucked unless they keep improving exponentially.Everyone will just move to local models, they're good enough for many workflows as it is and there's no risk of getting your IP stolen.
>destroys Anthropic
>>110137202a genius way to fix the losses
>>110137229Where is Qwen???
>>110136930If ai wins we all loose
lol @ the vibedogs getting all hyperbolic about nothing again.
>>110137267Pick something that takes longer to get done.
>>110137285Wat
>>110137264It's not a girl. Just a capybara. All the other models are fuckable whereas qwens are autistic males and talk like claude. The most feminine open weight models are the Gemma4 line, especially 12B and 31B. If you inspect their j-space they're very girly and menstrual.
>>110137304But that autistic capybara is wrecking anthropic
>>110137163Have you ever read the cries for help you post online? Looking for a new e-daddy to look up to, like a sorry little bitch.
>>110137229jew execution squad
>>110136893Their targets are based on inference only supposedly being profitable (not I even trust this claim given how much Sam or Dario lie). But assuming this is true, they can never stop training because the cheaper open models will immediately catch up to them, so there's no way to actually remove training costs and become profitable companies.
>>110137351and this is how you sound lol
Open-weight will win. It always does.
>>110137361Why do you automatically assume inference isn’t profitable? Have you looked at research on, for example, nvidia chips running open models for an estimation of inference? Have you looked at the finances of Anthropic and OpenAI to estimate how much money they’re spending on inference? If you’re just going to assume that they’re all lying or that it’s a conspiracy then we have nothing to talk about
>>110137361>>110137420
>>110137420It is profitable with large profit margins, but Anthropic and OpenAI are lowering their API prices to compete with open models. Most of the money they make from inference isn't from their monthly subscriber AI bros on twitter, it's from enterprise. Ed Zitron is right that ~80% of their revenue comes from VC-backed startups who are all subsidizing using VC money that will soon run out. Anthropic themselves even recently admitted a concern of theirs is 20% of their revenue comes from 2 large customers. The switching cost between Claude and ChatGPT, or even fucking Antigravity considering that now hosts Anthropic's models in a recent deal is terrible long-term for Anthropic. They only get around 15-20% cut from resellers like Google and Amazon.
>>110137437is it safe to assume it's not just chinks trying to fuck over us corps again?
>>110137477Open models aren’t catching up this cycle, they can no longer distill. And they don’t have compute to serve at scale anyway. > Ed Zitron is right that ~80% of their revenue comes from VC-backed startups who are all subsidizing using VC money that will soon run out.A source would be nice
>>110137511>they can no longer distillThey can distill Muse and Argon and the only reason why they distilled in the first place was because they could. China have very limited compute so whilst they've been trying to catch-up with Nvidia hardware, the Chinese labs just stole from the US labs to save them even needing gargantuan compute. Huawei are already making progress. China won't need to distill soon, but for as long as it's an available option to them, they will. >A source would be nicehttps://ramp.com/data/ai-index#spend-per-employee#financing-status
>>110137666> They can distill Muse and Argonbig deal> the only reason why they distilled in the first place was because they couldthat goes without saying> catch-up*catch up> China won't need to distill soontoo little, too late. Internal models are far ahead and the gap will keep growing
>>110137710>correcting typosoof, lil someone getting a bit butthurt lmao
>>110137750it just proves you have very bad pattern recognition and you don’t know what you’re saying, in addition to making meaningless points
>memory still all time price highs Micron stock trading a little lower does me literally no good. Plus this shits less tech bubble and more sovereign debt if you scratch the surface a bit.
>>110137798>getting this buttblastedkek
Nasdaq is still up 15% from YTD....
>>110136893>need 1000>set target at 300>get 500>look we overshot our target!
>>110138020You’re missing the part where the investors bought in regardless. In other words, they knew the whole time what the revenue predictions were. They’re even more confident now.
>>110138041the investors know targets are bullshit
>>110136877>SIHi fed! Which alphabet do you work at? No one says SI except glowies (because they have to).
POST THE FUCKING INTERVIEWhttps://www.youtube.com/watch?v=pHcZpvIfho0
>>110138112he does a new one every day...
>>110138069Bharati Intelligence Agency
>>110137304and they are retarded as a result. females have no economic value. this error will be swiftly corrected in gemma 5
>>110136596>Still replaces your job>Still replaces all tech workers>Altman and Niggerdei will still be billionaires
>>110139660if current level models replace ur job u were a codemonkey jeet anyway lol
