When the AI bubble pops we will enter a recession on the scale of which the world has not seen for decades.
To elaborate on the picture, over 3/4 of the "industry" revenue comes from Nvidia, who spend less than 1/5 of the total money. All of this despite no evidence that AI increases productivity. It's a good time to be selling shovels in the gold rush.
>>536555341>implying AI is even remotely close to the global financial crisis
What's the website? I am curious about 'spent since page loaded ' and want to see it what it looks like on my machine
>>536555457https://isaiprofitable.com/
4 million USD spent from the point I opened the website and finished reading
>>536555341I think you look at Anthropic, and every single year they get 10 times as much revenue. That means in two years, they’ll have revenue of $1 trillion at the current rate. So I think certainly that’s going to be profitable. If they have high valuations, it’s because people are rightfully expecting profit in the future, and I think that’s going to happen in the very near future.
AI is not useful for anything besides creating fake images/videos and have indians spam youtube/tiktok with autogenerated shorts.
Bjarne Stroustrup said he does not believe in AI. Look up on YouTube it's from this last month.
Tell me one thing AI is useful for in the real world.
>>536555670Very misinformed.
>>536555670Only because of all of the tech companies spending big on Claude Code. They're putting all of their eggs in one basket, and as companies realize it's not worth it their income will collapse. If you think about it Anthropic's business model is essentially creating AGI. The place is basically a cult.
>>536555732Based. I like C++, but I use it more as "C with some extra features". Could be just because I'm making small projects, might change when I actually start working.
>>536555755so instead of having to pay 10 coders to manage websites for hundreds of clients. Now I only need 2. Working towards 1, then zero. I find that quite useful.Theres nothing you're going to do about it. And if you dont use AI at all for productivity, I can only assume you're a retard cleaning shit off the toilet or something. What exactly are you useful for?
>>536555782With Anthropic, they have essentially already created AGI with that new model they are holding back and only giving to the big companies, the Mythos. What they’re able to do with superhuman hacking and coding ability is just crazy. It’s clear we’re on the precipice of something huge.
>>536555341It's normal for startups to be in the red for several years before profit.
>>536555851Read the date. Don't believe anything these companies say. It's marketing.
>>536555341Um sweetie the bubble, she's popped. The tech stonks all shit themselves on Friday USA time.
>>536555897What's the long-term profit model for Anthropic?
>>536555955IPO, cash out and go into hiding.
>>536555341>When the AI bubble pops(((AI))) isn't some investment bubble, you naive sod.(((AI))) is a another deceitful jewish *pretext* for the kikes TO SIMPLY STEAL ALL OF SOCIETY'S MONEY.
>>536555850Now you have 10 times more bloat and the code is 10 times less optimzed compared to shitty jeet-coding.
Watch, from 50 minutes agohttps://youtu.be/pxF-bUHBk1M
>>536555341AI is literally only being invested in because tech companies are praying for a Hail Mary after they brought over too many illiterate dead-weight third-world scammers with fake credentials to fill their ranks, while all the honest white male that built all the digital infrastructure have retired or been passed over due to DEI initiatives
>>536555341>ai bubble popsIt won't.It doesn't need to be profitable.Once the infrastructure is in place, it pays for itself in the sense that everyone will forever be subjugated with no hope of ever defeating the globohomo powers.The "ram scarcity" serves a dual purpose; making sure datacenters are supplied, ensures up to date hardware is harder to come by, by the common man. In the next two decades everything will be online; cloudbased and tied to biometric ID, and old hardware will not be compatible with these things, and thus won't run it. You will be forced to engage in cloudbased computing and streaming. Enjoy your queue and monthly subscription, as well as permanent internet ban for "wrongthink".I suggest sabotaging datacenters at any opportunity.
>>536555732>Bjarne Stroustrup"C++ is a piece of shit! C++ is a piece of shit!" - hilarious conga line chant at a computer conference long ago
>>536555850The problem is that it doesn't actually make you more productive. This is what an AI generated game looks like. Is it good?https://fly.pieter.com/
>>536556068Also misinformed. Guys I think the AI investment cycle was a Turing test but for gullible investors.
>>536555341why do normal people have to pay for the mistakes made by corporations and rich people? Who voted for this?
>>536555341I don't think it's an exaggeration to say that it might end American hegemony. The reason we're throwing so much money at this is because 50 years of neoliberalism has hollowed out the real economy of the west, while simultaneously causing the accumulation of a massive amount of capital at the top. This means there's little left to actually invest it in. Prior to "AI" the same phenomenon is what caused the throwing around of huge amounts of money into various tech boondoggles, and the absurd valuations of FAANG companies, but this is the crescendo. When it fails to be the deus ex machina they're hyping it up to be there will simply be no place else for the money to go, which means stagflation that we can no longer print money / financialize / offshore / mass immigrate our way out of
>>536555421you're right. its far worse. in 2008 there was still some semblance of a real economy that real estate was parasitizingnow the only thing that exists is AI bubble bullshit
>>536555341Anthropic is profitable though and has been for a few months now. OpenAI and xAI are dragging everything down, but that doesn't apply to all companies. https://www.wsj.com/tech/ai/mind-blowing-growth-is-about-to-propel-anthropic-into-its-first-profitable-quarter-7edbf2f4
>>536556122Give me one (1) language better than C++.
>>536555341If you don't count tech and healthcare we've been in a recession for some time now.In fact I doubt for most much would change when it pops. Except then maybe richfag politicians will have to stop pretending it's all good.
>>536555341The profit is the stock price. It doesnt matter if the buiness doesnt make money as long as stock is up in the trillions they can print new money via loans between each other.
>>536556040initial models were pretty fucking ass, but they are getting better and better. Theres no degradation happening. You'd have to be insane to not see the difference between 2022 public models and what we have today.Current top models: ~0.1-0.3 GWGlobally around 10 GW.Soon those data centers will produce 1-5 each alone.
>>536556211>>536555782They're all in on AI replacing software engineers. When tech wakes up and realizes it's not increasing productivity it's done for them. Which is funny because their own studies show it won't.https://www.anthropic.com/research/AI-assistance-coding-skills
>>536556272They're not getting better. Nobody who says this will share "their" code. Because it's either garbage, or they had such fine grained control over the agent that it saved them no time.
>>536556272The improvement in performance is mainly due to memory and quantization method improvement. They're still running trillion parameter models just to format a fucking essay for you.Meanwhile you can do 90-99% of that with a decent enough MoE on a home lab system like a dgx spark or strix halo setup, and practically sip power.
>>536555341I am optimistic about AI and the future. It keeps getting better, and I keep getting richer. Be a doomer all you want now. We'll all be living very comfortably in a decade.
>>536555341It's incredible to see this ponzi scheme continue even now, when the cracks are turning into chasms. The AI crash will probably be as impactful as the dotcom bust.People actually think a thousand GPUs doing what is basically a big clusterfuck of a compression algorithm is AGI.
>>536556126Games are still dogshit, however still impressive for what it can do in just 5-10 minutes. It would take you months to build out the same frameworks and get through the bugs.Its not building GTA 6 but it can make pac man in minutes. Also it literally just came out last week, where windows users can have codex use your applications like Blender to assist in the process.
>>536555670>AI company makes hundreds of billions selling its AI to other companies>Every company that is using AI services and has spent $1.4T in total services has been reporting tremendous losses, with AI-related revenue streams either underperforming severely or not yielding any form of benefitGuys in 10 years the AI snake oil salesman will be richer, therefore AI will be more profitable!
>>536555911>Don't believe anything these companies say. It's marketingTHISThe "Oh, our AI is already too powerful - we jews must keep it secret for the good of humanity!" line is entirely bullshit - because if the truth of "AI won't ever work" were revealed then all the shekels would stop flowing, the goyim would realize they've been tricked and the global economic collapse isn't from "AI" but from intentional jewish fuckery, and the jews would start get holocausted for real.Jews milked their laughably impossible "Holocaust" schtick for 75+ fucking years. Their "AI" replacement for the 'Hoax is just getting started.
>>536556369If you're impressed by that then I shudder to see what your projects look like.
>>536556342You have a horrendously low IQ then. Dont take my word for it, can you just go to a free website and let someone else show you? Just try and see if it helps you understand your ability to process new information. You're not built for this.If you dont know how to write python code, you really dont have any valuable input. 2022 models couldnt do it. Today is can.
>>536556219The funny part of the conga line was *Bjarne was there*.
>>536556211Rising costs and delayed costs compared to one time deals made that quarter are expected to make them back to losing money. I got it from AI summaries since you believe it :)
>>536556158>Who voted for this?Your owners, slave.
>>536556389Thinking they'll last 10 years with the way they're burning money is a bit fucking generous.Now they'll do what all silicon valley types do when the play money starts to run out. They'll enshitify it, and start using anti-competative tactics. It's been the MO for that industry for the past 20 years.
>>536556427Share your code then. I'll wait.
>>536556302AI is not going to replace software engineers. The research you posted from Anthropic themselves proves that. It's a tool that augments their abilities and lowers the requirement to do busywork like CRUD boilerplate. Anyone that uses AI regularly knows it cannot be trusted to operate unsupervised and even frontier models make insane errors without humans ensuring they don't machine gun their own foot.
the pic doesn't do it justice OP. miggers like sam altman, elon musk, etc are trying to subsidize their buildouts by using the government as an open wallet. they are socializing the losses while privatizing the gains
You retards actually think something will happen?
>>536556427>>536556427"Today they can"LMAO.
>>536556272>does not know about the exponential compute requirements to improve models>thinks each datacenter will produce 1-5 "models"Dude
>>536556504>augments their abilities...the study shows it hinders your ability to form skills, unless you take just as much time as you would without it. Where's the productivity boost worth these insane valuations?
you just gotta stay with it to the endDon't stop believin'!
>>536555416AI investing is also incredibly circular with the same companies / parent companies investing into each other.
>>536556497Share your IQ results - and I wont be waiting because I already have a very good idea of where you fall.Want to race to see who can build pac man first? Loser blows their fucking brains out?
>>536556566It definitely hinders the ability to form skills which is pretty scary and I'm not sure how that's going to play out in the long term. But my anecdotal experience is you can produce vastly more output in a single day than without the tooling, and at much lower personal effort expenditure. Also to some extent, the average developer has no choice but to adopt it, because resources we used to rely on like Stackoverflow are now dead.
>>536556485Made them both feel as if they were wronged, and that they're entitled to more than they had.
>>536556588And wall street pretended not to notice.
>>536556564Notice how the two previous lines said "GW".Also, another IQ issue. You kids are suffering. Keep waiting for that bubble to pop. Your unvisited tombstones will be rotted away before AI collapses.
>>536556485>They'll enshitify it, and start using anti-competative tactics.The problem is that it's already not returning anything close to what's being invested in it. Making the bar for entry for outside forces even higher and harder to reach will make adoption far less attractive, overcharging for services will make barriers that force current adoptees less likely to renew in the future, and further lobotomizing the models will only serve to darken how things look for potential future models. There's literally nowhere they can go at this point that is worse than where it is right now without toppling the house of cards almost immediately. I can't see AI lasting more than 3-5 years at this present rate without some kind of earnest savant-tier breakthrough that actually makes AGI tangible, and not just LLM puppetry masquerading as AGI.
>>536556596I have a 138 IQ as measured by the WAIS.>Want to race to see who can build pac man first?Do you think software engineering is about building pacman as quickly as possible?
>>536556625>resources we used to rely on like Stackoverflow are now deadRest in piss, I hope I never have to use it again
>>536556068>old hardware will not be compatible with these things, and thus won't run itThis will be the time when we build or own internet using old tech and using this against them: It will only be compatible with non-globohomo PCs of the old era.But we surely need a great firewall. Cyberpunk 2077 feelings are coming up
LLMs will be never able to code properly, not with any ROI. LLMs can understand human languages, that's their main strength. They can't reliably produce syntactically or semantically correct code without human supervision, which limits their application to supporting roles.Be happy while we still have them - retards wanting to create their own games and other aislop will burst the bubble for everybody.
Niggers aren't profitable either
>>536556022A different parallel is between "AI" and the Cold War.After their WW1+WW2 had destroyed white Europe the jews needed another Big Lie to keep the shekel train rolling. They invented the mysterious, secret, but world-ending-if-not-stopped threat of "the Commies" to keep shaking down the goyim. And it worked - to the tune of tens of trillions of real dollars.But like the Holohoax, the Cold War no longer is as profitable. So the jews need some new existential threat with which to terrorize the goyim out of their money.BlackRock's Fink even says it to that Texas audience: if we jews don't steal all your savings and retirement money for "AI," then "China will win."AI is the jews' new shekel train to replace the Cold War.
>>536556706>LLMs will be never able to code properlyIts already way better than what jeets produce
>>536556625The only reason you can produce more is because you lower your standards and don't think about your code. Since you don't see your code you don't care about it. Coding by hand forces you to raise your standards for quality and think about your project more deeply. The problem isn't that you "can't keep up" per se, it's management forcing higher quotas. Which is ridiculous, we all should've learned years ago that lines of code is a ridiculous measure for productivity.
>>536556678Shittest community on the entire internet, but pretty hard to code the old, pre-LLM way without it
>>536555341Imagine being such a subhuman piece of trash where this is your talking point, that the newest valuable industry is not profitable when comparing it to current monopolies and current scam corporations.You are absolute trash.There is no it's not profitable. The answer has always been to integrate it correctly instead of being a luddite clueless retard mongoloid who pretends it's not valuable when it is you who is not valuable.
>>536555670Lmao, didn't know Clammy Sam was posting on pol.
>>536556641Is the GW magically supposed to produce better models? Are all those "GW" in the room with you now?
>>536556706The legend himself, Edsger Dijkstra, explained this perfectly years ago. Natural language is ill-fitted to programming computers. The only way AI could meaningfully boost productivity is if it was actually AGI.https://www.cs.utexas.edu/~EWD/transcriptions/EWD06xx/EWD667.html
>>536556747>because you lower your standardsThis is just nonsense. Try using a frontier model like GPT 5.5 or Opus 4.8 and the code is very high quality. Most coding is simply not that complex and doesn't require deep thought about what you're doing. Most people are not writing quantum physics code, they're writing database selects, inserts, validation, forms, data serialization, model instantiation and other mind-numbing bullshit which LLMs now handle perfectly. And if you don't trust it, you can hand review all the code, but I stopped doing that after a while, once I discovered there were no flaws to find.
>>536556183It's the final cataclysm of the eternal struggle to sit like an effendi and eat. Humans are always hoping to find some way to make someone else do all the work but still get the bulk of the profits of work, and it has a natural swing back when the person doing all the work realizes he doesn't actually need the effendi for anything; that honor or property or law are mere words, and not sufficient to compel labor.It'll happen to the Chinese eventually, and then whoever comes after. Nobody ever learns.
Cute stuff.Anyways, check tomorrow. Check the next day. Check the day after that. AI will still be here. They will keep building and consuming all the power they need to build your replacements. Meanwhile you'll cry online and thumbs up posts from 4 year art degree majors, and nothing will change.
Everything on this planet is not profitable at this point.Everything should be closed tomorrow, but they aren't closed tomorrow because the legacy monopoly scammers are still alive and kicking.They legacy scammers exist to denigrate every living and breathing valuable product by making them look as shit as they are and they continue to refuse to fund and refuse to acknowledge the people who would bring value to us all.
>>536556669>I have a 138 IQ as measured by the WAIS.Why?
>>536556479>jewsfair enough.
>>536556829Post code. You're probably just a shitty programmer. I've seen what these things put out and it's garbage.>mind-numbing bullshitYou're probably not coding right then.
>>536556851>replacementsLol what if some of us are helping build the AI? Are you projecting? We're not arguing with you cause we hate AI, we're arguing cause you're obviously yelling about your own stock portfolio and don't use AI actually. Pretty obvious.
>>536555341It appears as if they were fabricating a pretext for the collapse of the global financial system to cover up their pathetic mismanagement.
>>536556859I took it in high school
>>536555341>Implying anything good will happenSeems like you don't understand.
>>536556851I have a CS degree. I assume you don't, because AI code impresses you. Yes, it is syntactically correct. That's not the issue.
>>536555341Good. Fuck it. Problem? Womens expectations need to drop. If mcdonalds isnt good enough for her, fuck it.
>>536556740> >LLMs will be never able to code properly>Its already way better than what jeets produceBoth produce shit.Though that's why kikes know they can replace jeets with AI-jeet bloody benchod bots.
>>536556907>probablydefinitely
>>536555341>>536555421Alternative future: AI will cause the collapse of money itself as AI does all the "work". AI will pop the money bubble.
>>536556919Precisely."AI" is the jews' next Kapparot chicken, to be blamed for the global economic collapse so the guilty jews can escape.
>>536557020vidrel?
>>536556720This point of view also has merit.
>>536555913ive never seen the tech heatmap in dark red. thats phenomenal loses
>>536555851> AGIUndefined marketing buzzwords are meaningless.
>>536556022How the fuck is he not dead and his guts and pieces stretched across multiple bridges and pikes?
>>536557087They did release Mythos to private companies and we would've heard about any major bugs including Microsoft, but instead some anime profiled leet hacker found several and reported them to the public while Microsoft using Mythos didn't. Makes you wonder.
>>536556944>I took it in high schoolChecked and I-wonder-what-it-is-now-pilled.
>>536557086On good, if pretend job numbers as well. Just a whiff of a possible interest rate rise they're saying caused that.
>>536557168About the same, probably. Maybe even higher. IQ is fairly stable by late adolescence.
>>536557087They also released a new model version and the benchmarks weren't anything major in improvements.
>>536556740That's not much of a criteria. I shudder to imagine how many kWh were already spent by jeets on this.>>536556817Thank you, interesting reading.> the ease of making undetected mistakes> bootstrap from there to a sufficiently well-defined formal system> as damned difficult to make as they would be to use> the last decades have shown in the Western world a sharp decline of people's mastery of their own language
>>536557154>to private companiesMeant in private to companies
>>536557142The Larry Finks of the world control many, many trillions of dollars.A few million of those are allocated to protecting the Larry Finks.
Oh boy here we go again. Another muh ai crash the world ohnoes. Get this. All the world's greatest inventions and innovations didnt just get shit done right out the creator's ass. Electricity? How many years, decades before it became common place? Lines and grids had to be established; generators and powerplants drafted slowly. Automobiles? A neat convenience the rich only had access too initially. Airplanes. This one is a big one and was largely benefit by heavy military use/backing. Commercial airlines were limited to fairly high income clientele. Only from about the late 50s early 60s was airfare cheap enough for average joes. Computers - present in fucking everything now. Expensive and bulky and incredibly limited in the past (both users and capabilities). Internet. Dont lie. This shit took a while before making it to the broader world. It was limited to military and universities at start. I am old enough to remember the first commercial dial up was pay by the minute. Only middle class or better could afford computers; dial up was a huge flex on top of that. AOL is normie gatekeeping removed and even then it took dsl in the 2000s to make the internet really take off.Touch screen. This has been around since the 90s. It wasnt commplace until smartphones. Now its on everything. Ai is the same thing. Just because there is no immediate use beyond its basic outputs doesnt mean ai wont grow into some necessity of a behemoth the world uses heavily.The markets crashing will do so on their own.its not like the dot come crash binned the internet permanently. Ai isnt crashing and it already has about 8 big names. Its carving out its own from the world regardless if the markets do what the markets have always done.
>>536557301The issue is that "AI" is a vaguely defined term. LLMs will not change the world. AGI would, but LLMs are not AGI. If you think these things will just keep improving until they can think then yeah, it would change the world. But they can't. And their glaring weaknesses with reasoning show that it's qualitatively nothing like human intelligence. This stuff isn't a human inside a computer. It's far, far more boring.
>>536557348Shh let him keep his money in those companies so that we can all get AI girlfriends at his cost.
>>536557301Good post. The upper-midwit with a 138 IQ doesn't get it at all.
>>536557401Do you have any background in computer science or programming, anon? Even in the early days, the internet, airplanes, and cars had clear use cases. What is the use case of AI?
>>536557301It's not AI itself that's the problem, everyone knows it's going to be great in the future. The problem is that in the USA it's development is in the hands of some of the worst individuals in the history of the world, absolute Jewish filth that have circle jerked it into a massive financial bubble.
>>536555955A) get bought by microsoft or google or the US governmentB) when running costs of AI fall due to tech advances it'll be cheaper to run and they'll start turning a profit
>>536557435what's the usecase of google? it makes searching easier. ai definitely has a few relevant usecases. ai is going to be with us forever, and it's going to approve. but there's a bubble and it will collapse. in 20 years we're going to have amazing ai tools.
https://youtu.be/CwDgF0J49hQI still listen to Ai music and it's awesome. Even the music video is goodProblem is we need more nuclear power plants
>>536556907>Post code. You're probably just a shitty programmer.I've literally worked at a FAANG company in SV, I doubt you could even pass the interview. Keep on being mediocre
>>536557518Finally someone who gets it.
>>536555341We have two $1T IPOs coming. Chilax. Think how you can profit from these IPOs.
>>536557544I know idiots who got hired at FAANG. Post code.
>>536557518What's the usecase of AI? You still haven't given me anything.
>>536557518> amazing ai toolsAnd if you will grow to be a good alpha slave, your robowaifu will be 3D.
>>536557435>What is the use case of AI?picrel>UniComp (referred to as "Uni"), a central supercomputer that manages the entire global population under a system called "The Family." >Function: UniComp controls every aspect of human life, including resource allocation, job assignments, marriage pairings, and reproduction, to maintain a society free of war, hunger, and crime. >Control Mechanism: It enforces compliance through monthly chemical treatments (delivered via transdermal spray or injection) that suppress individuality, creativity, and dissent, keeping citizens chemically dependent and obedient. >Surveillance: Every citizen wears a permanent identifying bracelet that interfaces with scanners to track their movements and permissions in real-time. The Illusion of Autonomy: While UniComp is perceived as an all-powerful artificial intelligence, it is revealed that the computer does not make decisions independently; it is actually operated by a small group of programmers (migrants from the Family) who live in luxury underground. These programmers manipulate the system to maintain equilibrium, sending nonconformists to isolated "nature preserves" (islands) that serve as de facto prison colonies.
>>536557444Checked. You said in the second sentence exactly what they're gonna do:Establish a dystopia using AI. That's all they wanthttps://www.youtube.com/watch?v=-gGLvg0n-uYhttps://www.youtube.com/watch?v=8_8LTUmHWP0
>>536555341>When the AI bubble popsThe AI "bubble" are all the random startups no one has ever actually heard of whose whole business plan is "[thing that exists] but with AI!" When that bubble pops, most of those companies will cease to exist, with the few who have valuable IP getting bought up by real tech companies.But when that bubble pops, AI technology itself won't be affected. Just like how after the dotcom bubble popped, the internet just continued to grow and become more pervasive in everyone's lives, the same thing will happen with AI.The company that will win the AI race made ~$150 billion in net profit last year.
>>536555341All the top fortune 500 companies are losing money with the exception of only a few. They all buy products and services from each other which they write up in their accounting books as profit. The problem is the same cash is going from one corporation to the next and back around again. Their actual profits are in the negative though it makes it look good for evaluations when they become IPOs.
>>536557621what's the usecase of google? why do you use google? because it saves you having to go to the library.same with ai. it saves you time for some tasks and offloads you, there are things it can do that aren't very important or are trivial, and it can do them quickly. for example i could tell claude to write an email to my colleagues anne, charles and david and ask if they're available for a meeting tomorrow at noon. it can do that, very reliably. it saves me a few minutes.if you're good with ai, it can save you even more time. stop tihinking in black and white terms mate. it's not all bad or all good. it's no 100% useless or 100% useful. it's useful for me, but it doesn't solve everything for me. it offloads me on certain points so i can spend more time on other points where it can't.
>>536557652Cool science fiction. What is the usecase of the LLMs we have today?
“Give them bread and circuses, and they will never revolt.”We are standing at the edge of total digital enslavement, and most people don’t even realize it. The global elites are tightening the control grid, and they are selling your slavery as convenience.Look at the pieces falling into place:Digital IDs – Your entire existence linked to a government-controlled ID. No ID = No access to anything.Central Bank Digital Currencies (CBDCs) – Every transaction tracked and programmable (they can limit what you spend your money on).Social Credit Scores – Behave as they want, or they cut off your access to services.Facial Recognition Everywhere – No more privacy, you’re constantly monitored.15-Minute Cities – A test run for total movement control. Try to leave? You’ll be restricted.Smart Meters & Carbon Allowances – They will limit your energy use and tax you for existing.Neuralink, Starlink, AI Control – They want to merge humans with AI and ensure total thought control.This is not a conspiracy theory. Look at how quickly cash is disappearing, government surveillance is ramping up, and "green" policies are being used to justify complete control.They are preparing for total technocratic rule—a world where you own nothing, control nothing, and must obey to survive. Once they roll this out fully, resistance becomes nearly impossible.What can you do?Use cash whenever possible – Starve the digital beast.Decentralize your life – Learn real skills, invest in hard assets, and avoid reliance on digital-only systems.Call this out now – Once they normalize these systems, it’s over.Opt out of "smart" systems – The more you rely on "convenience," the tighter their grip gets.This is the final fight for human freedom. If we don’t resist now, we enter a world where compliance is the only way to survive.Time is running out. Stay awake. Spread the message. Resist.
>>536557633possibly, that seems to be the way we're heading. the western world seems to accept that fact. i'm not endorsing what's going on mate, i'm observing. you have to be a good observer if you want to make it, put your opinions aside. we all have them, they're not interesting.
>>536555732>Bjarne Stroustrup said he does not believe in AI.C++ is trash and that faggot ABBA boomer doesn't know shit about fuck.
>rich guy in israel gives rich guy in america 5 billion dollars>rich guy in america gives rich guy in taiwan the same 5 billion dollars>rich guy in taiwan gives the same israeli the same 5 billion dollars>repeatHOLY SHIT BROS WE CREATED 45 BILLION DOLLARS IN VALUE!!1!!1!!
>>536556272There isn't much more data to improve the LLMs with. Your 2022 models were exclusively trained with a tiny sample data set - then all of Amazon product reviews - then all of reddit.What you see today is trained on most of the internet, most of all digitally available books, video, audio.The only information left now is proprietary information that is internal to major companies. They're doing their best to steal this info but we've reached the peak with available data. Anything else is making things more efficient but chatgpt is still boiling 1000 gallons of water to answer "what is 27+16 divided by 2" rather than intercepting the prompt and running that in a calculator like mathway has been able to do since 2010.
>>536557765
>>536557712Your usecase is writing emails...? That's it? In the time it takes you to prompt Claude, wait for it to give you output, and send it, you could've written the email yourself. It's also a bad thing to do for your sake. People will respect you less if they know you use it.
>>536557830Honestly if I were managing a company, monitoring what my employees use LLMs for would be a great midwit detector to then fire them. Like if they're spending tokens on fucking emails lol, wewlad what can they actually do?
>>536557686Nvidia, you mean? Nvidia made money? You don't say.
>>536557738I know, anon. No offense meant.AIs are a great productivity enhancer, if the user isn't a retard. However, most of us are retarded.Lemmings gonna lemming.
>>536556219>Give me one (1) language better than C++.COBOL
>>536557891>AIs are a great productivity enhancer, if the user isn't a retardEveryone I know who says this uses it like a retard would.
>>536557830no, it can do trivial chores that take a lot of time way more effectively than i can. you're a dumbass if you argue this. no, i use ai for a few other things as well.you're fighting wind mills mate, i'm not saying ai is amazing, i'm saying it has a few usecases that make it useful here and there
>>536557928Communicating with others is not a trivial chore. It's what makes us human, however cliche that may sound.
>>536557762What's wrong with C++?
>>536557686>The company that will win the AI race made ~$150 billion in net profit last year.Which company, and what was their primary revenue stream?
>>536557946it can be, for example just asking someone if they're available for a meeting, or sending a file to a group of people
>>536557976Dude that's literally like a minute's worth of work I don't understand why it bothers you enough to use AI for it.
>>536557967its not assemblyif you dont program in assembly then you're a propagator of slop and i hereby revoke your eligibility to the gilded halls of real programming
>>536557969No one mentioned how good the open source models are by comparison and how they'll cut into profits of all AI companies. But what do I know?
it's gonna be the same as the dotcom bubble. your america onlines, prodigys, compuserves, all going to go poof then the real companies will take overyou'll have your ebay, amazon, and yahoo that don't fail of course
>>536558000it's probably 3-4 mins of work if i have to send a file via mail to a group of people. i can prompt the ai to do that in 30 secs, and it will likely take the ai 8-10 minutes to complete the task.it's just an example, i can't imagine you can't come up with similar ideas for where ai can actually be useful.
Jorge R Gutierrez now has to hire South Koreans to make his cartoons. I hope you arttroons are happyI thought you liked beaners?
>>536557976Oh damn, I thought you "got it", but apparently you do not. Easily one of the workers to be first replaced by any mass AI rollout. It is what it is.
>>536555715That is useful. People want to be entertained
>>536558035So a lot of smaller companies will be replaced by a few big ones? That's even worse.
>>536558077i'm not employed mate, i checked out entirely 2019 and haven't had a job since. i don't have colleagues, i'm saying there are tasks where ai can be useful
>>536556126>6 FPS stutter festYep, it's AI
>>536558041You waste 3 minutes all the time, every day. The fact that you feel it necessary to automate a thing where you're presenting yourself in your voice to other real people just makes me sad.
>>536558129it's funny how you make this out to be about me personally when i'm trying to give you an observation of how society works in general. come on mate.
>>536557717>What is the usecase of the LLMs we have today?I find them useful for digging through scientific research to answer unusual questions more efficiently.They also help rewrite reports to account for nuanced language requirements.Need a script to interface some piece of hardware on an MCU? Punch one out with a few prompts.It's a far cry from anything like Unicomp, but that's how it will be sold. Bit by productive bit. Use cases are only limited by imagination and one's capacity to edit the dross, but the irony there is that a point could arise where it inflects and the creativity of the user is lost to atrophy. Perhaps it is like the Lotka-Volterra model. Where the productivity to creativity derivative is 0.And that's precluding any moaning about "token cost."
>>536557906> operators in variable namesPascal is better.>>536557927WDYM? I am retarded. I used duck AI to create simple Bash scripts, Gemini for help with troubleshooting several issues outside of my expertise; heck, I even had to learn Lua to fix its blunders in a Lua Wireshark plugin it wrote for me.I even seen some local Claude models used to analyse logs - done by a QA guy, mind you; providing results which said QA guy wasn't able to deliver before.LLMs aren't completely useless, anon.
>>536558115Sorry to hear that fren.
>>536557087>Undefined marketing buzzwords are meaningless.The reality is that even mundane AI models are effectively smarter than the vast majority of humans. SOTA frontier models can already produce deliverables that are better than 99% of the PhDs and SMEs in any technical domain, much faster, cheaper, and more efficiently.AGI and all of this mental masturbation about how do define consciousness and trying to pretend that there is some unknown, undefined, indescribable magic spark in human intelligence that can't exist in synthetic intelligence is an irrelevant red herring. It doesn't actually matter what is really happening inside the black box, whether it is physically a human brain or an LLM on a GPU. The whole idea of "oh wow, but what if it is just a philosophical zombie or mindless NPC that is just pretending to be conscious?" is just as true for human intelligence as it is for AI.In the real world, the output is all that matter. And we are already well past the point that the responses produced by AI are smarter, better, and more intelligent than anything produced by the vast vast majority of human minds.
>>536558166>LLMs aren't completely useless, anon.it's funny so many in this thread don't get this. they're arguing against "it can do absolutely everything" with "ai is 100% useless". as per usual, the truth sits inbetween. LLMs can be useful, and AI is a bubble that's going to crash, but over time, AI will become more and more useful. this is the path of EVERY new tech there ever was.
I don't think the push for AI is about profits anymore. The ruling class already has money, but they desire to have even more control over the working class goyim. The push for digital IDs is most likely connected with it too. They want future where there is no personal computing. You buy a terminal you can't tincker with, you log in with your digital ID and you can access a couple of pre-vetted sites, and AI scans your every move. They can even try to constantly rewrite history to suit their current needs. War with Iran didn't work out? Don't worry it never happened, AI already memory holed everything.
>>536558166I used to use it for that kind of thing, but I've realized lately that it's impeding my learning. If I had done it the hard way I would've learned more. Nothing is free. Nothing is ever free.
>>536558221I don't think they're completely useless per se. I think that the net effect is negative. AI might have a few small use cases, but the massive technical debt and deskilling it leads to makes it a bad idea overall.
I wonder if PC parts will become affordable again or if this retarded gold rush has done massive damage to the economy.
>>536558203>The reality is that even mundane AI models are effectively smarter than the vast majority of humans.this is also a good take. i wouldnt necessarily use the term smarter, but what people are forgetting is that 90% of people on this planet are basically wild animals. even being a way-below-average american and having gone to school whatsoever, you're basically way above 95% of people on the planet.
>>536558271>I think that the net effect is negative.i absolutely agree with you here, but that's a problem with the users, not LLMs.
>>536557600You'll pass the interview one day bud, keep trying!
>>536558203>SOTA frontier models can already produce deliverables that are better than 99% of the PhDs and SMEs in any technical domain, much faster, cheaper, and more efficiently.There's no way you actually believe this. If you mean it's better than PhDs outside of their field, then you might have a point.
>>536558310Post code, pranjeet.
>>536558222My initial thought is they don't want any of us having computers anymore.
>>536557154>MythosThe real reason they got so spooked by Mythos is because it is a Recurrent Depth Transformer / Looping Transformer model. Instead of having an readable chain of thought like typical "thinking" models, Mythos loops through reasoning internally, inside latent space. This makes it extremely powerful and effective, but also means that there's no way to see the chain of thought reasoning that led to the response.
>>536556650Current AI companies might not last a year. SpaceX IPO is looking rough and OpenAI is trying to negotiate a government stake in the company.
>>536555341On the research side, people have been saying it wasn't ready to go public, but all of these different groups are trying to show off what they have come up with, and it is easier to suck money out of the public than to keep begging private investors
>>536558362>there's no way to see the chain of thought reasoning that led to the responseThe reason for this to defeat distillation attacks. Minimax, Deepseek and Kimi are all trained from distillation attacks on Opus. https://www.anthropic.com/news/detecting-and-preventing-distillation-attacksYou pretty much have to trust the reasoning is sound, or not use it. Mythos has had a massive impact on security hardening of products like Firefox though, so the reasoning does seem legitimate enough to have warranted a significant panic at Mozilla. https://hacks.mozilla.org/2026/05/behind-the-scenes-hardening-firefox/
>>536558221/threadI'd still prefer a global thermonuclear exchange, tho.>>536558225That's the thing. I'm an autodidact, my progress was limited by the amount of relevant information I could get. I learned C++ from local library and Borland's F1 docs. When the Internet came ... well, I downloaded plenty of pixelated porn, while thinking to myself how much time I had lost compared to the younger generations, with all that information on the tips of their fingers.And today ... I'm tired of this "winning".AIs are like the Internet in 90ies. Some will profit by them, most will get lost in slop.
>>536558362The looping model is already integrated with most "thinking" models. I think the real reason they can't release it to the public is that it costs a lot to run per token due to how much it needs to loop to get to a more accurate solution and that it's been trained to use hacking tools.
>>536557348>it's qualitatively nothing like human intelligencePlease tell us exactly what it is that you think makes human intelligence any different.The reality is that there isn't actually some magical spark inside human intelligence. Human brains are obviously multi-modal and integrate senses and visual/spatial/emotional intelligence and all of that stuff. But your conscious inner voice that most people generally tend to think of as (You) is VERY similar to an LLM.
>>536557880>Nvidia, you mean?No, I don't mean Nvidia.
>>536555670and im 4 years, they'll have 100 trillion!! amazing!!
>>536555341dot com bubble x 1,000,000
>>536558500>the real reason they can't release it to the public is that it costs a lot to run per tokenThe reason they couldn't release it is because Anthropic didn't have the compute capacity until just the last month. They pretended it was because the model is too powerful, but the actual reason is they were hard compute limited, to such an extent they added 'adaptive' thinking to Opus, which basically made it stop thinking for most tasks, reducing compute requirement and made it significantly dumber. They just signed a deal with several compute providers (mostly Google but some excess xAI capacity is being used because Grok is so unpopular and they have unused GPUs), so now Mythos is magically not so dangerous any longer.
>>536555341This would have been avoidable if the west wasn't slaves of the Anglo
>>536558337Yes, they want to sell premade boxes that act as terminals and to store everything on cloud forcing us to buy subscribtions for everything and allowing them to restrict, modify or delete content as much as they want. The final stage of capitalism is the elimination of private property for the bottom 99% of society forcing everyone to rent everything from the top 1% who owns everything. Companies like Google were very unhappy with customers rejecting cloud gaming because people still had alternative so the plan is to simply remove that alternative altogether by making PC parts unaffordable.
>>536558520Just talk to ChatGPT for more than a minute and you'll see exactly what I mean lol.>very similar to an LLMYou know most of our evolution happened before language existed at all, right?
>>536558567Yeah also seems plausible since they released 4.8 soon after they made the deal with those companies for compute lol
>>536558629Bro, you're so shallow minded, I'm pretty sure that you're an LLM instructed to spread AI FUD, yourself.
>>536555670Line always goes up!? Oh fuck, I gotta invest in this new religion!
>>536558629> evolution Let's start with species able to pass the mirror test.
>>536558679
>>536556219You have to ask?
>>536558664They also magically doubled everyone's limits because GPT's limits were ridiculously generous compared to one prompt and you're limited on their entry-level plan with Opus. I was hitting my Max limit constantly before the new compute upgrade, now I haven't even been warned of 75% and I'm using it all day, every day. Also Mythos coming out in a few weeks, then we all get to subject our codebase to security torture and discover what horrifying 0-days were lurking unfixed
>>536558707Do you not understand that Google AI summaries is the most bog stupid model that exists and in no way representative of frontier models. It's like the 486 of LLMs, of course it sucks. Fucking hell you are ignorant
>>536555341My life is already fucked who cares
>>536558735Sweet! Can't wait to see how holey all the other LLM code has been and both laugh and shid at the same time.
>>536558765>NO MUH CODEX/CLAUDE CODE IS BETTERmeanwhile my friend's "agentic" project with the $200 OpenAI subscription won't even compile for god knows what reason and every commit piles on mounds and mounds of technical debt that'll never be fixedHow genuinely retarded do you need to be to think that AI is never wrong and reviewing its code is pointless? If you're truly a FAANG worker as you said, you're probably just doing fine because everyone in the industry is drinking the kool-aid. These things are genuinely pointless in coding.
>>536555755Searching net for data and crunching some basic conclusions on said data. It's really good for that. Also, throwing around ideas out of left field, which are interesting enough to consider. Business, medical research, chemistry, astronomy, genetic studies. This is where this tech really shines. I tell you, it really helped me develop my business plan and how I should approach my business in the first place. But yeah, you gotta keep in mind it has the attitude of a lazy, bright student. Which wants to spit out the least effort answer at you, so you can fuck off.
>>536555341>>536555496I can tell that website was designed by Ai. Because it looks exactly like a site that I designed with Ai lmao
>>536557969>Which company, Google/Alphabet>what was their primary revenue stream?Advertising (Search and Adsense), followed by Cloud services and Subscriptions/Platforms/Devices.Google will win the AI race because they control their entire AI vertical tech stack. They have their own frontier models running on their own TPUs sitting in servers in their own data centers, connected via their own networks, through their own platforms/subscriptions, all the way down to controlling the OS on most of the hardware devices end users use to actually interact and use AI. OpenAI is planning on releasing a prototype hardware device sometime in the next year or two, meanwhile Google already has the ideal AI hardware device already deployed to virtually every human being on Earth, especially since they made the AI deal with Apple.Google's real business is obtaining useful information, and they have managed to position themselves so that consumers are lining up to pay $20/$100 a month so that they can feed Google an unprecedented amount of extremely valuable personal information.
>>536558222>War with Iran didn't work outThis is why AI will always better than you. Only thing saving iran right now is shitty moral argument not drone nor mountain and you know it.
>>536558915>Iran only won because we didn't nuke themkek
>>536558894Stop comparing these things to humans. They're nothing like us. No, it's not an intern. No, it's not a student. It is a large language model.
>>536558867>meanwhile my friend's "agentic" project with the $200 OpenAI subscription won't even compile for god knows what reasonThis is a huge self-own on your friend's part. If you can't get your project to even compile, the issue is PEBKAC, not the model.>How genuinely retarded do you need to be to think that AI is never wrongWho said this? AI is frequently wrong, the same with humans. This is why I said earlier models will never fully replace humans because their probabilistic nature means they inherently do random, unpredictable shit all the time and need human supervision. And you never get a model to review it's own code, always get another model to review it or review it yourself, by hand. >These things are genuinely pointless in coding.I used to think this too, but you're missing out. The amount of busywork you can eliminate by learning to use these things properly is incredible. LLM prompting is a legitimate skill, like googling effectively. Once you learn to use them properly, you won't be able to go back. It's like upgrading from a hedge trimmer to a chainsaw. Coding without one seems quaint and backwards.
>>536555851g-guys our new AI model is frighteningly good, we name it XxX_G0dK1LL3r420_xXx we need to be shut down quick or else the entire human rice will become extinct pls whatever you do dont invest in our company
>>536555341I'm getting two kitties soon. They will scavange the wastes with me. Fuck you and your demoralization thread.
>>536558995You're saying all of the things he says. Hitting on every single point. Every vibe coder says they're one of the elite few using AI right. All of them.>review it yourself by handDefeating the entire purpose of AI. Great.>The amount of busywork you can eliminate by learning to use these things properly is incredibleBusywork is time to think.>Coding without one seems quaint and backwards.Again, exactly the kind of thing my friend would say. Slow down your brain.
>>536558324>There's no way you actually believe this.It's not based on belief, it's based on objective testing. The hard part now is finding the last 1% of PhDs and subject matter experts who can actually fact check and critique the work produced by the models.
>>536556596Get a load of this retard.Go to Github and search "pacman game"AI trained on every open source repo out there (and are currently getting sued for training on GPL licensed code).AI "infered" your pacman game from likely 1 or 2 of the most heavly starred repos in Github.AI did "think" to create your damn pacman game. It predicted tokens from training.Try a novel probelm and see how far you get with LLMs
>>536558995Also lmao @ you walking back your earlier claim that you shouldn't review every line of AI code.
The problem is artifacts in the responses, as the Ai gets faster and supposedly better the artifacts become self reinforced. What people are mistaking for agi is artifact cycles.Data centers are a throwback to original subscription computing, the idea back in the 70s was you would own your terminal but rent time and data on a central computer. In a way it solves the issue of better personal devices, because big tech is very much aware that market saturation has come, people are not buying into the scam of updating their devices every year or 6 months as the trained rabbits they used to be. It's a dying gasp of technology trying to stay relevant. What people want are things that last for years, they can hold, and own maybe even pass onto others. Not waste, not rent. It is a shift, and if they came right out and told you that you can't even own anything you would reject it. Centralized computing is Centralized control, not millions of sources of information, but one source, one filter.Data centers will be built, but they will not last. The same garbage construction going into warehouses that are projected to earn for 30 years but fall apart in 10 is being used. The smart money will build them quick, then dump them while the hype is still hot.Maybe a smart plan is to understand them, so that when the time comes you'll know the most efficient way to recycle them.
>>536555341I use AI for such stupid shit and I always tell it to do a deep dive.Fucking hell I wonder how much money they spent just on me for that $20.Make that bitch work for that money.Netflix will spend 20 million making a movie for all its users.OpenAI will spend $2 million in server costs making shit just for me on top of all of the other bullshit.How do you fix that model.
>>536559123I would encourage you to just try it out. It's like changing from a bike to a car. Or don't, but I suspect the niche for 'artisanal coding' is going to get smaller and smaller as time goes on. Its going to become an important skill that it's just assumed you have and not having it will narrow you opportunities significantly. I wouldn't be surprised if the only useful thing that comes out of LLMs is transforming the software industry, while no other industry sees any real benefits and it ends up being only a great developer tool and nothing else.>>536559185I don't review every line of AI code, I'm not walking back anything. I'm suggesting you review it by hand if you're not comfortable with trusting the model. But I personally have different frontier models review each other's code and it works great.
>>536558970I just explain in simple words how AI chatbot approaches answering to prompts from an end user with 5 years experience, no need to have your pants twisted.
>>536555341Hopefully people will be able to pull out their investment fast enough to avoid causing an economic cataclysm.
>>536555341only 2 more bubbles
>>536559247No, no. You keep coding with agents, and I'll keep coding by hand. I prefer it that way, so when the higher ups start to realize what's happened I'll be more valuable. You keep drinking the Kool-aid. Keep your sunk cost.https://www.cs.utexas.edu/~EWD/transcriptions/EWD06xx/EWD667.htmlI'll post it again.
>>536559182"AI did **NOT** "think" to create your damn pacman game. It predicted tokens from training."When a LLM is loaded onto a GPU it does nothing until prompted. It does not sit there planning it's day or do anything without a prompt.Anything it has created, has come from it's training.Since I study the open source repos before I've used Claude,I can recognize which repos it ripped off.
>>536558629>Just talk to ChatGPT for more than a minute and you'll see exactly what I mean lol.Just talk to a randomly selected human for a minute and you will realize that most humans are basically retarded too.>You know most of our evolution happened before language existed at all, right?Yes, and most computing happened before people started to realize that neural networks were a better option than symbolic AI and expert systems. And before the transformer architecture was described in 2017
>>536559241This guy maybe gets it!
>>536555341Isn't it just a handful of companies spending on AI?How will collapse of AI affect grocery stores or construction?
>>536559241I had github copilot run for 8h on one prompt once back when it was request based.
>>536555416retard, AI has literally more impact than the internet & industrial revolution combined. it will only continue to accelerate
>>536559394>>536558702r u ai
>>536559247If you're not reviewing what the AI spits out, then what exactly is your job? What are you doing that some rando couldn't?
>>536558765>Do you not understand that Google AI summaries is the most bog stupid model that exists and in no way representative of frontier models. You're not wrong, but AI Mode's redemption arc is going to be incredible. https://www.youtube.com/watch?v=RvZ-667CEdo
Elmo's 'datacenters in space' is such hilarious bullshit, we will look back in 20 years and be amazed no one laughed in his face as he spewed this scifi drivel with a serious look and boomers forked over billions into his IPO
>>536557589>Think how you can profit from these IPOsNow THAT is how the jew makes money.Screw the goyim. Worship shekels.
>>536555850Why do many make it seem like AI will only go after programming and not math or physics?Supposedly AI solved some problem that has been around for 80 years.Robots are getting better at small jobs as well. It's not hard to believe small robots can be trained to complete plumbing and electrical jobs in the coming years.
>>536557722>Time is running out. Stay awake. Spread the message. ResistDigits of absolute truth.
>>536555416oh fuck here i was gonna post a spiel about the difference between average losers and a small minority of winners.
>>536559511It's as crazy as thinking we'll be traveling to Mars with modern ships.How much food, water, and other supplies will one person need for a nine months trip?
>>536555341Anthropic is currently profitable, it will just take a little time to catch up to their investments up to this point.
>>536559503No, it's just a good book and very on point to these kinds of discussions. Highly recommended.
>>536559721No one tell him.
>>536559272Them pulling out is the cataclysm, anon.
>>536559772Ability to reason != soul, but that's moot - nobody wants to give LLMs human rights, yet.Still, the idea of bicameralism as disembodied voice wrt training of LLMs is noteworthy.
>>536558030>if you dont program in assemblyLittle-known bit of history:The reason Bill Gates and Paul Allen "wrote" MITS BASIC for the Altair in the CS building at Harvard is because **that's where the source for DEC BASIC for the PDP was stored.**"AI Considered Harmful" - Dijkstra
>>536558608Jews want to own everything.Slaves will own nothing.OWNERSHIP IS FREEDOM
>>536558702>Let's start with species able to pass the mirror testWell that eliminates niggers and jews
>>536559511They called me a meatspace!How very dare they
>>536559330LLMs are just a higher abstraction level than current coding. Just the same way people initially rebelled when languages like C came about, because they felt they were relinquishing control that assembly gave them. Until the convenience of these languages became overwhelming and all the crusty holdouts got overrun. Then the transition from low-level languages to high-level, then from statically typed to dynamically typed. Every time people thought it was losing control, but what you lose in micro-control, you gain in other areas. Same thing is happening here.
>>536555341>establishing shot of the Griffin household>Vacuum noises>Peter answers a knock at the door and opens to let in Glenn Quagmire>"Hey Quagmire, how's it going?">"I did it Peter">"Did what?">"I watched the last season of porn, Peter">"What?">Stewie explains to Brian that Quagmire has ascended the attachments to desire>Quagmire levitates in lotus pose upstairs past Meg who is polishing the wood with lemon pledge>Oppenheimer on TV announces AI has destroyed the commercial value of all intelligence>None of the cast reacts>We now see the entire Griffin family doing chores with Cleveland in the house for an uncomfortably long scene>Joe Swanson wheels in with a Dave and Buster's pamphlet>"Hey guys let's all go to Dave and Busters!">Everyone talks over each other with their own "yep" "I could go for that" "sure why not">Quagmire silently smiles>Awkward silence>Long awkward silence>Joe starts to jitter and the camera pans in for an extreme close up on his eyes>Cut to black for Sopranos ending
>>536559505It's funny how all these things are so pedestrian now, but they were considered unfathomably futuristic 30 years ago. Pretty much every prediction there is on point.
>>536555341thats really hard to measure and saying its not profitable is silly because its a tool like a hammer:- you give some retard a hammer and heill think its a piece of-you give some waggie a hammer and heill build a 5th house for some boomer-you give me a hammer and ill knock out some boomer thus freeing up 4 housessee what i mean
>>536560610*piece of candy and try and eat it
>>536560325Did you even read the article I linked? No, LLMs are nothing like compilers. A compiled language is a higher-level formal representation of computer logic built on a well-defined rigid interface. LLMs are probabilistic predictive code generators that take in natural language prompts and output compiled language in unpredictable ways. If anything the highest level language we have isn't natural language like you imply, but applications themselves. We already have low-code or no-code tools to build applications. 99% of the time when normies want to make the computer do what they want they can, because of applications. And what else is a programming language than that?>Then the transition from low-level languages to high-level, then from statically typed to dynamically typed.Except as a developer you can work with any language you want. Assembly, C, C++, Python, JS. For the system to work you need people working up and down the stack of abstractions to maintain every interface. Your argument rests on believing higher-level languages are somehow "better" or an evolution over low-level ones, while in reality they're simply suited to different tasks. You wouldn't say in the 2000s that you NEED to learn Python and Java to get ahead, because there were obviously still jobs in C or C++, or even ones that work with assembly. But you said exactly that with regard to LLMs. What a poorly thought-out comparison. TLDR code more brainlet, your logic is lacking.
>>536560336This is just the last episode of smiling friends but family guy
>>536555851>With Anthropic, they have essentially already created AGI>my SQL query is alivefuck off
>>536560696>output compiled language in unpredictable waysYour knowledge on this is stuck in 2024 or 25. LLMs are probabilistic but largely very predictable nowadays. The technology has advanced considerably since GPT4. It only became viable to code full-time using an LLM in late 2025. Since then, a lot of developers haven't hand written any code, because it is no longer necessary. The output is so predictably good that it's pointless. >You wouldn't say in the 2000s that you NEED to learn Python and Java to get aheadI would say if you wanted to maximize your employment chances, picking an in-demand language is always going to be a good idea. >But you said exactly that with regard to LLMs.There will undoubtedly be people like you out there still willing to hire handcoders because they have a similar ideological bias against LLMs. But those jobs will become increasingly hard to come by, just like niche language jobs have in the past.> TLDR code more brainlet, your logic is lacking.You are so arrogant combined with ignorant of the current state of LLMs. A dangerous combination that makes someone feel confident while being extremely wrong.
>>536560940>C++ isn't in demandlmao you are a fucking clown
>>536560995Your reading comprehension fucking sucks man. Where did I say C++ wasn't in demand? It very obviously is. Lots of miserable legacy codebases need maintainers because their authors left for greener pastures and now the company that owns the codebase needs a bagholder to keep it from falling in on itself.
>>536555341kek imagine believing that it is about money, this is why white people always fail. jews actually don't care about money , or to better say it it is not their priority. they care about power/ pulling the strings, and in AI case knowing every digital data point about every human out there, will make it easier to bundle up all the people who opposed them
>>536560940You clearly didn't read Dijkstra's article. Put more effort into this conversation, please. Natural language is ill-fitted to programming, and thus the only way we will get genuine use from AI is if either you tell it what to do in such detail that it saves no time, or we just get AGI. Which I think we both agree won't happen with LLMs. >a lot of developersMade Windows 11. I don't care what "a lot of developers" think. Post your code.>output is so predictably good that it's pointless.Then why is my (genuinely competent) friend's AI code so garbage? He says exactly the same as you do. How do I know that you're any different? You're not even reviewing your code, how do you know it's good? You don't even know what your program is doing.>ignorant of the current state of LLMsLooking at current frontier LLM code isn't enough? Sure, everything looks plausible enough on paper, but that's precisely what makes it dangerous.
>>536561056Where is Swift?
>>536556342Memeflaggot is right. They started getting semi usable last year. They're even better this year. They can actually write code now. Yes there's debugging and a lot of other stuff, but if you actually believe what you said, you're a non dev retard who just spits out xitter talking pointsMost people itt are immensely retarded
>>536559467well yeah, more impact in the sense that it's destroying the internet. not more impact than the industrial revolution, not even a fucking tenth of a tenth of it, but the impact that is there is purely negative
>>536561056C is much lower level than Python or Java. And yet it's in demand. Again, your analogy is collapsing in on itself, and you won't even bother to defend it. Nor will you renounce your pride and admit that you were wrong. If you genuinely think that C is a worse language than Python because it's lower level you might just be clinically retarded. Probably Indian too.
>>536561201>You clearly didn't read Dijkstra's articleDjikstra is in his 70s and bordering senility, who gives a shit what he thinks? Elderly people struggle to adapt to changing paradigms, how shocking and unexpected.>Then why is my (genuinely competent) friend's AI code so garbage?Because your friend is not genuinely competent and most likely neither are you for thinking someone incapable of using LLMs, a normal work tool for programmers now, is competent when they are clearly not. If you can't even get LLM-written code to compile, the person is far from competent.>we just get AGI. Which I think we both agree won't happen with LLMs.I don't even know that I agree with this any longer. Go work with GPT5.5 in Codex and you might be surprised. Anyway I'm tired of this conversation, if you want to remain ignorant which you are so absolutely determined to do, feel free.
>>536561252>they can write codeIs it good code? More importantly, is it the code you want? Is it code that you comprehend? If you're relying on them to write all of the code for you, how will you hone your skills with the technologies that you're using when something inevitably goes wrong? I don't mean that they're not better than they were a year ago. Sure, they are. That seems to be the case from what I've heard. But there are so many problems with them that prevent them from actually being good to use.
AI is not about profit.AI is about building a system.And once this system is up we are fucked.Can't wait it to backfire.
>>536555341> $613 billion revenue613 old testament commandments for jewsYup, I'm thinking ai is just fine
>>536561207Apple pretty much suffocated their own ecosystem with the app review process. Swift has died in popularity. Also Typescript killed everything and has become the English of programming languages. I'm a Swift developer myself and it's an excellent language, but very unpopular in 2026.
>>536561390Interesting, thanks.
>>536561336>Dijkstra>isWow man you're more retarded than I imagined. Dijkstra has been dead for 20 years, and he wrote that article at 48. In 1978. Around 5 years after C first appeared. C!>Because your friend is not genuinely competentThere is no true Scotsman. It compiles on his end but the build system is incredibly broken.>Go work with GPT5.5 in Codex I've seen the results of current GPT in Codex. They're not pretty. I'm not forking over money to Scam Altman to know that. >if you want to remain ignorantSays the guy who thinks Edsger Dijkstra is still alive.
>>536561207lmao
>>536561524>It compiles on his end but the build system is incredibly broken.He sounds deeply incompetent not to be able to solve this and so do you, best of luck, sounds like you need it!
>>536561569I will say that he has said the exact same thing about other vibe coders. I will also say that you are moderately unintelligent and incredibly ignorant on the fundamentals of computer science. Good luck though?
>>536561569> AI requires a competent human to be truly productive humanbros, we're so fucking back
>>536561624It doesn't. This is a cope for vibecoders to think what they're doing isn't braindead.
>>536555341AI is not going away. The AI bubble may burst, but that will not stop governments around the world from continuing to use it. They will say it is a matter of national security and use it to confiscate land and resources for the creation and running of data centers.
>>536561624Yes, it's become apparent over time that LLMs are a multiplier on the competency of the human driving them. So an idiot cannot produce anything of any value and a decent programmer can 10x their output. So intelligent humans are pretty valuable after all and not replaceable. All the hysteria about developers getting replaced wholesale turned out to be nonsense hype for all the AI IPOs happening later this year.
>>536561608>I will also say that you are moderately unintelligent and incredibly ignorant on the fundamentals of computer science.Your friend cannot code their way out of a wet paper bag and in your poor judgement, you think they're competent. Those in glasses houses should not throw stones.
>>536561657AI I've seen was a mix of extremely informed and extremely dumb. It is still worth using, especially when there's no other option available, for the price - if you're able to discern between informed and dumb.Yes, it's not free. No, it's not infinitely costly.YMMV
>>536561742There's very cheap and powerful Chinese options available now like GLM and DeepSeek. If you're on a budget, these are dirt cheap, even affordable in developing regions of the world. American options are expensive, but China has almost caught up and charges 10x less. If you have decent graphics hardware, you can even go local and use a model like Qwen 3.6 or Gemma 4, though expect a step down in quality compared to datacenter-based models.
>>536561825Thanks for the tip.I will definitely try to train some local model on my vast collection of occult and religion books, just to test its limits. It's fun to be able to talk to someone who isn't compartmentalized.
>>536561681It's become apparent over time that you're just repeating thought-terminating clichés you read online. Shows you how much "I worked at FAANG" counts for.
>>536562090That's not even a thought-terminating cliche, you've misidentified it. And your experience with LLMs is 'my friend can't figure out how to use them', and you're acting superior because of this tiny speck of knowledge.
>>536562143>A thought-terminating cliché is a form of loaded language – often passing as folk wisdom – intended to end an argument and bypass cognitive dissonance with a cliché rather than a pointYeah, I think it is.>And your experience with LLMs is 'my friend can't figure out how to use them'No true Scotsman. Do I have to give Altman my goybucks for you to be happy?
> –fuck damn it
>>536555851may we see the superhuman hacking and coding ability?
>>536555341I would pay for AI if it didn’t have forced limitations on everything. It won’t even do its best to save money even when paying.
>>536563051That's probably where all the cost is coming from. They insist on rewriting the whole thing every time it starts talking about goblins.
>>536555416>All of this despite no evidence that AI increases productivitySpoken like a retard that doesn't know how to use AI.I used grok to find what wire was broken in my door jam it narrowed it down after about 3 tests. It even knew where the break would be based on people's past experience. It did all this as fast as I could test what it told me to test. It would have taken me hours and cost a shit load of money buying a full wire diagram of this exact year truck to even start.
>>536555341industry spends at the beginning for the data centers then it's all revenues. Everything's fine, AI stronk, no bubble.
>>536555913They're making space for the spacex IPO.
>>536555341AI is profitable for me because I run it locally on my pc
>>536561379>That seems to be the case from what I've heardso you're just talking out your ass
>>536555341>Le AI bubbleBrother, Americans have a non white problematic bubble that hasnt burst since 1965.Your country runs diversity on a yearly trillion dollar deficit for decades.Get the fuck outta here with this bs.AI will never burst because it is a national emergency.
This won't do anything lol
>>536555341okay then dont buy openai and ahtropic IPOs. im gonna bet on anthropic tho.
Right now AI is a parasite.It crawled the web for all its dataIt has no consciousnessIt requires insane amounts of electricityIt costs more than it generatesIts eating up hardware availabilityIt’s eating up acres of landIt is a parasite or cancer!! The companies wrote pro war manifestos. They want to send your kid to die. And The CEOs are lying about the capabilities of it. They need Trump to stay afloat. Its a parasite.It convincingly returns me search information that is often wrong. It synthesizes video and stitches the frames together. It genrates mainly bad porn. It does some work I think my local device should be able to do, like figure out the context of what im talking about concerning spreadsheet calculations…but took lots of wrangling to make sure it wasn’t lying…cause it was lying about my data…chatgpt lies…Its going to convince the stupidest people, like the ones who drive into a lake because of their GPS said to.It also cant solve the old problem which is verifying what facts are real in the first place. Garbage in. It relies on so much human work the AI companies owe a debt IMO
>>536555670The line can't keep going up forever faggot, this whole FOMO business is getting stale>oh my god robots are running the navy!just shut it, we're not dealing with terminators here
>>536563214>AI is good because I can't read a manual and don't know how to fix anything without being handheld by a machine>and YOU are the retard for pointing this outembarrassing post
>>536555897Its normal for all companies to be in the red until they get government contracts to spy for intelligence agencies because the economy is fake and gay.
>>536555341The fact that it's not making any money or that it's a bubble doesn't mean anything.Nothing made any money in the dot com bubble but some of the largest companies in the world came out of that.
>>536559467>more impactHow are you quantifying that? What metric? >it will only continue to accelerateThere isn't enough power generation and there will continue to not be enough power generation.
>>536555341>When the AI bubble popsI call dibbs on all the data centers.
