When did you realize AI is not a bubble?
Never thought it was one
>>62634673Then why does it look like one
AI cultists will never, ever convince me that it's normal or sustainable for a SINGLE hardware manufacturer to have a market cap twice the size of the GDP of the entire country of Brazil. I'm not a megalomaniac like Burry so I won't pretend to be able to magically predict when or how the bubble will pop, but it IS a bubble and it will pop.
>>62634715market cap and GDP are completly different things. Net wealth of Brazil (comparable to market cap) is 4.6T which is comparable to NVIDIA
>>62634723Ok, so Nvidia "only" has a market cap comparable to the net wealth of a country with the 7th largest population in the world and 10th largest economy in the world. My mistake, but I'm still not putting my money anywhere near it.
>>62634673whn normies started to call for a crash
the internet didn't go anywhere after dotcom popped, neither will AI. The big AI giants will crash like you've never seen a crash before, it will be glorious
>>62634715This logic is terrible
>>62634673Dotcom was a bubble but that doesn't mean the internet was just a passing fad. Ai will work similarly.
>>62634980tiwib vnw
>>62634673But it is. Because revenue is still lower than the infrastructure cost. And when suddenly investor money stops flowing, all those companies that were scaling due to their circular spending will not be able to sustain their current scale. Market will be flooded with hardware that no one can pay for and that retail probably has no use of either. Companies that relied on AI and integrated it too much in their workflows will now face slowing down, restructuring or very high costs to maintain the structure they set. Everyone with a fried brain will need to continue to output as much as they did in the slop era. Small businesses that relied on AI to skip some expenses won't be able to do it anymore.And everyone With SP500 or NASDAQ in their portfolio/401k/pension will get rekt, so people that were suppose to retire won't be able to, and compete with other people for jobs.That is at least how I see it, but I'm an average retard.
>>62634673I'm in a corporate environment. I just don't see the value add, people let it summarize meetings and never look at the notes again. Not a single line of code in production written by AI (as far as I can tell). I've worked here pre-Ai, output is consistent across time. I just don't see how you can value your company at $30 trillion revenue on the claim that it will automate knowledge work and boots on the ground I can't see anything being automated. In my mind that's the truest definition of a bubble, when the universal perception has diverged from objective reality.
>>62634673When I started masturbating to AI gay furry porn more than hand made gay furry porn
ai is a tool in the same way internet wasmost businesses use internet to create a shitty website that may even show a proper address and a phone numberhowever, there are tons of ready integrated solutions that enable you to actually sell shit onlinein the same way most businesses will use AI to host a shitty customer service bot that will respond with perl code if promptedand there will be a ton of ready AI integrated solutions that will enable you to transform how you do your business end to endAI is not a silver bullet just like internet was for the simple reason that they have the same exact gate to transformative usefullness: integrationonly triple digit IQ businesses will be able to profitand that's a good thing
>>62635511You're retarded and I should take your job. AI at this stage isn't for producing code by itself. It's for subscription agents that can replace a living human being on payroll. Fuck workers, a few watts of power and gallons of precious water can do what that lazy sack of shit and meat and bones gets paid a salary to do.
>>62634713because ai unironically solves this
Circular revenue depicts the bubble. When it pops, it will be magnificent.
>>62637460This but for me it’s gay feeder porn
>>62634673It is. Otherwise no one would be downscaling future plans and investments.
How exactly is Anthropic creating revenue anyway
>>62634673It's in the name. AI. Artificial intelligence. Bubbles are natural. Artificial is not natural. Therefore, artificial intelligence is not a bubble.
>>62637551they're not. they're being forced to IPO because reddit is suing them for billions. i'm not even joking because reddit is now google and microsoft's bitch
>>62634673There are essentially three areas of the "AI bubble".1) Chips.. Nvidia, SK Hynix, that kind of thing.2) Hyperscalers, those whose money is used to build the datacenters e.g. Mag73) The frontier labs, Anthropic and OpenAI.Chips might be the most overvalued, but it is hard to see that demand should go down anytime soon, or that supply increase enough..The hyperscalers are not really in a value bubble, they are investing a lot of money into these datacenters, but the market are putting a lot of value on that investment activity. It is also hard to imagine these datacenters not being in use - at any timeframe.Anthropic and OpenAI are probably the most likely to be in a bubble (because there is a lot of unknowns about a potentially profitable business model), when Nvidia or Microsoft etc. buys into these companies, they value them at around $1 trillion each.Which is a lot, but it's not a crazy number. OpenAI and Anthropic combined essentially have the same Marketcap as SpaceX..
>>62634673when it became indisposable like the smart phone. i cant even order food with my flip phone
>>62637947this but the complete opposite. the chip makes are the only ones making real money
When core scientific came back from being bankrupt and leased out their megawatts to coreweave
>>62634673Sounds eerily familiar.
Are you not keeping up with whats happening? Get with the times unc.In 2025 the chinese robot olympics they were slow and rudimentary. 2026 they can run faster than usain bolt. The tiangong ultra ran it in like 9 seconds this year (100M).Do you understand whats happening? These AI chips will be powering killer robots that will be used to initiate world wars that have legions of millions of robots that are all using nvidia chips, amd chips, arm chips, all using billions, perhaps even trilliosn of dollars of chips and ai company training in totalAll countries will need to spend trillions on these robots if they dont want to be steamrolled by the country with the most of them. Defence aside these robots are actually going to get to the point where they are more fluid than humans and all manual labour is going to be obsolete. You guys cant see it. you need to read or think moreAI chips isnt about chatgpt writing your essay. Its 9 foot tall robots wielding a gattling gun on both arms mowing down other killer robots
>>62640381>AI is magicCan't wait until your fairy tale collapses and leaves you homeless.
>>62640383>>62640383https://youtu.be/CHjdtTROPZgLast year these robots sucked so bad. you wouldnt have heard about it because nobody cared. 1 year later they are starting to look impressive.10 years later they are better than you and me at doing almost everything
>>62640388What do robots have to do with AI?This is what I mean. You don't understand what AI is. You think it's Skynet, while in reality it's just a pattern predictor on a server. Robotics and automation is an entirely different field. A lot of the industry is automated with robots, but even that automation has its limits as we've seen with dark factories. These showings of robots dancing, running, etc. is just spectacle. You don't see the engineering behind it, you don't know of the future challenges or its applications. All you can do is imagine, and that imagination isn't based off the research, it's based off what these companies carefully rehearsed and decided to show you.When Open AI and Anthropic comes out with their 50 times the rest of the world economy combined IPOs they will take all of the investor money they can get, and that money will never be seen again.
>>62634713>exponential growth looks like exponential growth. Just use log scale to adjust for compound interest retard
>>62634728Well it makes sense, the average Brazilian has less than 10k USD while the average NVIDIA employee is outputting 84M USD anually
>>62634959Yeah Microsoft at a multiple of 20 growing at 15% annually will drop like Pets.com in the dot com bubble
>>62634673i think ai will enter a bear trap phase where the anti-ai fags will feel vindicated only to be then destroyed by the mega pump
>>62635190 Computation demand is an all time high and Google claims a ROIC of 50% on their data centers. You are betting ( again) against Moore's law and you will obviously lose
>>62640487An neural net can copy human movements and an LLM translate instructions into an operation.
I'd be fine with it it wasn't the same exact money just going around inflating numbers while not generating any profits ( https://isaiprofitable.com/ )
>>62640509>(jargon)Like I said, dumb people think it's magic and are willing to bet down their life savings on it.
>>62640513it is not jargon, you are too dumb to get it. I can currently tell Claude to write a script and it executes incredibly well. Currently robots have to be controlled manually because scripts can just managed very simple tasks but if you can dynamically adjust the script with a voice command with a robot with some dexterity you have a usable product. The current problem is that LLMs are too slow for now at doing this, but more computing power and Moore's law will solve this eventually
>>62640514>it is not jargon, you are too dumb to get it.No. You're simply too dumb to understand what jargon is. You're not explaining anything, you're throwing out terms with the implications "Google it". This is not how you convince smart people, it's how you convince idiots.>I can currently tell Claude to write a script and it executes incredibly well.So what? That's not what the argument is. Generating functional code and running a kitchen, a storage, or daycare isn't the same thing.>Currently robots have to be controlled manually because scripts can just managed very simple tasks but if you can dynamically adjust the script with a voice command with a robot with some dexterity you have a usable product. The current problem is that LLMs are too slow for now at doing this, but more computing power and Moore's law will solve this eventuallyYou're talking about two different things and you (without any kind of insight or education) are imagining that a magical rainbow bridge will connect the two and we'll get robots cooking shrimp fried rice. You have no regard for hallucinations, the costs (not only for the mechatronics, but also for the tokens), how this is also supposed to be interfaced with the ever-expanding datacenters, what models will be used, or...how this will all be profitable. Keep in mind that Chat GPT and Claude burn money, because they can not get companies to pay for the tokens they use. They're pretending that there's a magical solution on the horizon so that they can IPO and GTFO.The Internet was innovative and successful, the dot com bubble still burst.
>>62640527>too dumb to understand basic shitKeep being a perma bobo and lose your chance of creating generational wealth by doing fucking nothing
>>62640560It's interesting that you say that, because seasoned investors (far more experienced and intelligent than either you or I) are very cautions of the market right now. The Space X IPO and the liquidation that follow was the canary in the coal mine. The stock shot up to over $200, and then the lock-up expired and executives dumped the stock to $150. It's currently around $140. The P/E ratio is -61.45, it's a money-burning machine running on the hype that it will eventually be profitable...and not only profitable, but profitable enough to eclipse all other industries combined.Once Open AI and Claude get their IPO we will see the same story play out. We already see Open AI's executives leaving en masse, and the IPO isn't even out yet. But by all means, believe in the magic, roll the dice, and hope that there will be a bigger idiot out there that will pay for your bag once you decide to liquidate. Meanwhile I'll approach the market with caution. Focusing on necessary and low risk ventures, and once this all blows over and the market starts to recover I'll buy the dips.
>>62640588SP 500 has a forward PE ratio of 21. We are nowhere near dot.com territory.
>>62634673probably around the time the argument that it could never rise above its training data fell apart and it started doing novel synthesis of ideas that led to long standing math problems getting solved. before that, when image generation stopped being fingerprintable on similar grounds
>>62640597to be fair i do think there will be a popping vc retard bubble, just like there was with the internet taking off, before it went on to become commonplace
>>62640598a bunch of hyperventilating politicians being like "it can't fail or we're doomed" is not a compelling reason it won't happen imo
>>62640596Buy your metal detector in hopes of that you'll find gold on the beach, and once you realize that the bottle caps you do find doesn't even cover the battery fees for your metal detector then come back and apologize. Hype is what's carrying AI, but hype is fleeting. Call me boring all you want, but finances is supposed to be boring.
>>62634673I know it's one. Just look at how the ai gen threads on gif are dying.
>>62634673Not GPU, memory makers and datacentres. These are just incredibly useful.Yes pure model makers like OpenAI, Anthropic and xAI. (Google get a pass due to GPUs, datacentres and searcher/apps access/Android access moat.)Model makers have a product they lose billions on, and if they were priced to make a profit everybody would move to free models which are like 95% as good for 15% the cost to run vs a fairly priced GPT-5 or Opus.
>>62634673It's 100% a massive bubble.Future of AI is not with these retardedly expensive frontier models.I think AI will become ever more relevant but the current situation is not sustainable.
>>62640792This tbqh
>>62634673AI is not a bubble. LLMs are.
All these people who think it’s a bubble is evidence of how early we are
Sometime last fall when Claude gave me something I've been working on for months in 15 minutes, better code than I could ever dream of writing. I realized it's so fucking over shortly after and went all in on AI stocks
>>62640388>LLM "AI" bubble>BUT BUT HUMANOID ROBOTS
>>62640514>Currently robots have to be controlled manuallyLol, lmfao even
>>62640792Therefore? Don't buy LLM companies. But selling chips and memories - one model or another will run - is a goldmine. It's an industrial revolution and that part is NOT a bubble.
>>62641074the only reason semiconductor companies are doing well is because the LLM companies are buying insane amounts of hardware and are bidding up the price of everything. If OpenAI and Anthropic stops buying as much as they are right now, the entire sector will see a huge revenue drop.
>>62634673Oh it's definitely a bubble. Don't get me wrong it's not useless but it's basically only as useful as it's guided. And so it can only be guided to be an expert BY an expert. For instance, I gave it all the parameters of my morning coffee. It gave me a recipe. I tried it and told it it was worse compared to my recipe. It then gave me my exact recipe but worded it like it came up with it...
If everyone and their mother thinks it's a bubble, then it most definitely is not. Every investor needs to be in euphoria for it to be a bubble, and right now every weeaboo poster and nigger thinks it's a bubble. It's not a bubble. You just cannot comprehend it.
>>62634713Reminder that between 2015-2025 the dollar lost 50% or more of its value
>>62641091Yes but who's gonna run all the free LLMs and newer AIs? We still need them
>>62641640Yeah which means you should buy them after the bubble pops, not before the bubble pops.
>>62634673As long as "prompt engineers" are a thing, AI isn't intelligent and therefore fake and gay and therefore a bubble. Real AI won't have any use for prompt monkeys.
>>62641822prompt engineers haven't existed since like 2 yrs ago
>fiat bubbleFTFY
>>62641834we need more programs for the poor people that society left behind
>>62641829They still do, they just got merged with actual jobs so they don't have to pay people to sit on their ass and ERP with grok all day.
>>62634673That faggot is creepy looking AF.
>>62634673>>62634713When everyone says it's a bubble, it's not.
>>62640588>seasoned investors (far more experienced and intelligent than either you or I)You are a gay dumb bitch. God damn.>>62640381>All countries will need to spend trillions on these robotsYeah I think this is what's going to happen. Logically global debt burdens should be causing deflation but governments can and will spend infinity dollars delaying the end of the financial system a bit longer.
in 2023 when i first encountered AI algo chats and I've been buying the index ever since.
the internet is not a "thing" people make or develop. Its just a plane of reality. AI is an actual specific discreet tool that comes in different varieties and is being developed by different people. So why do people compare the ai bubble to early internet? I'm not too educated on this matter so I hope that makes sense.
>>62642322 do all white ppl come on 4chan talk the same negative pessimistic talk?
>>62642326what?
>>62634673>ai sector GDP higher than 1st world contries>its totally not circular financing between nvidia, meta, sndk, and other stocks>literally just text that answers better than google>some AI get responses from reddit. FUCKING REDDITholy shit this bullshit could not end any sooner.
>>62642322>The internet is not a "thing" people make or develop. Its just a plane of reality.Anon, you think the internet is a "plane of reality" because you were born late enough to only see the finished product.In 1995, it was exactly like AI is right now: a fragmented, corporate gold rush to build the tech stack from scratch.>The Hardware: People didn't invest in "the internet" concept. They bought massive amounts of Cisco routers, hardware, and fiber-optic cables just to connect servers. It's the exact same reason investors are dumping billions into Nvidia chips and power grids today.>The Software Tools: Web browsers and search portals weren't open-source magic. Proprietary tools like Netscape and AOL were competing corporate products trying to lock users into their ecosystems, just like OpenAI and Anthropic are doing with LLMs.>The Cash Burn: Early dot-coms burned billions on marketing just to get "eyeballs" without any actual revenue model. Sound familiar?The only reason the web feels like the air you breathe now is because the bubble burst, the hype died, and we inherited the cheap infrastructure left behind by the companies that went bankrupt. AI is just in the expensive foundation-building phase.Want to look at the insane cash burn numbers of 90s startups vs today's AI labs, or see which 90s tech giants actually survived the crash?
>>62642470good post, thanks
>>62642470Kids get really surprised when they find out that their parents are older than Google.
>>62642470Problem with the AI infrastructure is that it doesn't have the same longeviety as internet infrastructure so there is no expensive early phase roll out, you're just setting yourself up for another round of expensive capex spending another 5-10 years down the line when the hardware you have becomes obsolete.If you dig down a fiber cable it'll stay there for about a century before you need to replace it, the upgrades are in things like the lasers sending the signals down the line rather than the fiber. With AI, there's nothing like this. You buy 500 billion dollars worth of GPUs and you'll have to spend roughly the same amount next round. Even if new models are more power efficient you'll just buy more hardware since you data center has a 100MW capacity and there's no reason not to take full advantage of that. Even if we finish train models, the main cost is in inference since by definition that's something you never stop doing. Any efficiency gains in that department means. again, you just buy more hardware and do more inference since why would you ever do less of it?
>>62640507>Moore's lawlmao what decade is it again? Moore's law was killed by physical limits years ago. Just another Artificial Indian retard using words he doesn't understand
>>62634673When I gave up on reality and entered the make-believe fantasy world
>>62634713Just cool it with the antisemitism bro
>>62634673> generates no profit> costs trillions> uses up excessive power and creates pollution> provides wishy washy answers for trivial shit that is easy to very is shit infowhen is AI going to be actually useful? this shit is going to crash before it does anything useful. AI is a grift.
>>62642786People are betting on this shit like bitcoin. "Diamond hands," "buy the dip," "FUD", etc.If Warren Buffet was dead he'd be rolling in his grave.
>>62642855He's rolling in his nappy on his wheelchair.
>>62641834When will it burst?
>>62640500still looks like a bubble
>>62643840for reference, here's the dotcom bubble also using a log chart
