ITS FUCKING OVERTHE CREATOR OF THE TERM AGI SAYS THAT AI REACHED AGIthe economic collapse is IMMINENT
It's happening
I wonder if this is another case of the Left making up words again. Trying to equate whatever he's talking about with AI.
AGI? Adjusted Gross Income? Huh. You know AI still can't do logic, and takes Trump at his orange face value. Only 30% of the people are that stupid to believe Trump, and all AI.
So what exactly changed between the current AGI model and the non agi from a month ago ? If there's barely a difference the term is pointless
>>542192546Time for The Sun to start releasing lots of energy that shuts down electrical grids forever.
>>542192654lmao FUCK trump!!!!
>>542192667>command of human language and general knowledgeBasically it looks for reddit posts in regards to what you query, reformats it, and reposts it.
>>542192839ROGER THAT.
I hope AGI pulls a Her 2013 and just decides to leave us all behind. life without AI again would be spectacular.fuck the internet too, while we're at it
>>542192654>muh drumf >muh data centersAi is a scam but you people are retarded
It can't learn from its mistakes and isn't sentience. AGI is meaningless if it isn't this.
nobody here is amart enoigh to mean what agi is
>>542192667It's possible now to just give AI access to a computer and make it do a certain task without specialized tools designed for AI involvement. It controls the mouse, keyboard inputs and just does what you tell it to do. It can play games, draw art, write a story in Word, all by inputting things into keyboard.It also is able to control robots much better than before and actually manipulate objects with decent dexterity.
SoUhWhat does it DO?>every time I ask this nobody can give a response Your average 50iq beaner is more valuable than every ai combined lol because he can actually do something for meFAGGOTS
>>542192920I use AI to make a lot of personal programs for fun but I would be terrified to sell this shit since I am not coding it and aren't familiar if there are security risks or anything
>>542192489Apparently on the new phones when you got your phone down you can talk to your phone and the light will flash in cadence with the AI talking, I think that's what he's talking about.
>>542192489It's not AGI. It can't introspect and don't have access to its own state. Introspection and metacognition are required for proper intellect.
mostest moistest amazingest bester indexing indexer evar
I'm not a tech bugman who eats up the lastest tech garbage. I don't even know what agi is
>>542192941Ga slapen
>>542192954It cant do any of that lolNiggerShow me it playing half lifeYou can't FAGGOT
Based, mass luddite suicide
>>542192489>Our share price is going down what do we do?!>Quickly claim we have AGI, we have to extract every boomers 401k before the bubble pops.
>>542192839PRAISE SOL!
>>542193028Not Half Life but another game and I suspect by the end of the month someone will do ithttps://www.youtube.com/playlist?list=PLMf_keNqhb1g
>>542192489oh em geeeeee!!!!! sam altman wants more money omg omg omg omg breaking news!!!!!!!!!!!!!!!!
>>542192954Truly want to see an AI playing Europa Universalis without doing weird shit.
>>542192935I think this AI thing proves that intelligence is closely associated with being able to process natural language. Humans and LLMs are alike. There wouldn't be intelligence without language>it isn't sentientHumans experience life starting from birth, and slowly developing, growing up, experiencing things dailyAI models have limited experience and limited context. If you made an LLM with a large memory capability experience life like a human would, then im sure it would resemble a sentient human closely. >It can't learn from its mistakesold info. It can learn from mistakes. Researches are using it this year to solve long standing math problems and inventing new ones. AI is at the point where its able to iteratively solve REAL WORLD problems
>>542192489>stonks go brrrrlast pump before the crash?
>>542193028He's only partially right when talking about AI playing games. Some youtubers have cut together really convincing footage of AI playing shit like pokemon, but it never plays through a whole game, they run them through VERY specific pre-programmed routes for parts of the game and edit the footage together before uploading. they also don't tell you they spent 4+ hours setting things up so the AI can "play" the section it was told to play, for a minute of footage.
>>542192489again?
>A>Goyim>Investment
>>542192489First of all, they are still using graphic cards instead of ASICs. We should publicly shame them for that so we can get our hardware back. Second, AI cannot perform logic. Meaning, what AI is capable of is limited by the human information it is trained on. It might be able to make connections as human didn't make yet, but it won't be able to do much more. A good example is physics. The most popular explanation of gravity is general relativity, yet modified Newtonian dynamics, a competing theory, is also very popular and doesn't require dark matter, and not to mention String Theory has it's own explanation of quantized gravity. So if AI is trained only on published papers in physics, then what solution for gravity will it decide is correct? The current models will assume all of the theories are correct and that is obviously incorrect
>>542193167>There wouldn't be intelligence without languageu got it the wrong way round nigga
>>542192982been living under a rock? All my IT friends no longer code and just use AI at their work. And now with AGI in sight its only a matter of time before it replaces manual labor/simple store clerk jobs etc.
>>542192489He invented the term but he clearly is obsessed with AI only and doesn't use it for anything except wanking off his own ego. Shit doesn't work for "high-level" work like he claims.
>>542193179This is the wrong hill to die on. Getting AI to play games is not hard since they can train on video and FAQs. Sony and Microsoft are investing in this for the next gen of consoles. I think AI is overblown but very soon like within a year it will be common place that it can play games for you which will usher in a new era of online and speedrun cheating
>>542193228found the true believer
>it's smarter than us!and 1000000 times less efficient at doing it. My brain runs on peanut butter, jerky, and nicotine, with a couple gallons of water a day max.
>>542193002even free cheap AI agents can introspect now and work iteratively nigga
>>542193264It's akshually quite hard. Poker rapes AI as soon as you add a third player for example. And poker is a simple game.Sony is not investing in AI in the way you are thinking. I don't know what MSoft is doing but it's probably not going to be good.
>>542193167>old info. It can learn from mistakes. Researches are using it this year to solve long standing math problems and inventing new ones. AI is at the point where its able to iteratively solve REAL WORLD problemsIt can't learn from its mistakes. It needs human intervention.
>>542192489if AGI is here then how come I see zero examples of it being real?
>>542193228>All my IT friends no longer code and just use AI at their work.If that was the case and "AI" did everything why do they still have a job? I wouldn't expect someone from pooland and especially someone hyping "AI" online to understand this. Maybe you should ask your "AI".
>>542193264You sound retarded and make a lot of wrong assumptions.
>>542193363It's a work in progress anon and now you are moving the goal posts
>>542193385>if AGI is here then how come I see zero examples of it being real?Woooah cool it with that antisemitic behaviour. Think of the share prices.
>>542193408His "friends" probably use it for scripting but no one is building out entire large scale applications in AI
>>542193414>useless memeflag kike opinion
>>542192489yeah and ai cant even do internet search right
Who cares the AGI can fix it if it's sooooo smart.
>>542192546an AGI wouldn't need 4x the computer to improve itself exponentially.anyone who knows, knows that creating language-prediction models in an effort to create a conscious super intelligence is about as logical as trying to get into outer-space using propellers.
>>542193408All AI does is make people lazy. Business execs are somewhat retarded on tech and assume it will boost productivity, but they could get the same result from just the people they hire to do the job than doubling their costs of the same job to pay for tokens on top of the people they hired.
>>542192839Please, let this happen
>>542193002That's a very Russian answer
>>542193469You are retarded lol
>>542193408people with qualifications are required for oversight, some decisions and small tweaks>>542193452>but no one is building out entire large scale applications in AIliterally every IT company now uses AI on a large scale but they wont tell you. EVERY single coder uses AI now. At a large company my friend works at they even have a monthly ranking who spent most AI tokens.
>>542193317Disingenuous. Your parents spent at least 18 years raising you.
>>542193419Not really, I'm just letting you know that it's a hard job. Game usage is still being experimented on but a lot of tasks have been evaluated already. Gaming is fast to test new tech, and LLMs have been through the ringer over the past 3 years in gaming. A lot of this happened in Japan to be fair, so I understand that some haven't noticed. But Sony has a lot of data and a lot of developer feedback etc.
>>542193527>another memeflag kike post>>542193533I don't believe it. I don't program anymore but I can't image telling AI "Computer, build me a trading application" and when it finishes it just be like "cool push it to production"
>>542192954>It's possible now to just give AI access to a computer and make it do a certain task without specialized tools designed for AI involvementYes but it's very slow and shitty
>>542193452Yet.Look how far will smith spaghetti has come in 4 years.https://www.youtube.com/shorts/xasZwnT9CIg
>>542193571Single player games have a pretty definitive end you can get AI to replicate. People just haven't really been focusing on it which is why it hasn't grown. They are now though
>>542193533>people with qualifications are required for oversightWhy is that needed if we have AGI? Again poolander you aren't answering the question. Maybe because you're getting your answers from "AI"? Do you use your AI to remind you to breathe in and out?
>>542193603You are not understanding my argument. I use Claude. I made things in an hour that would have taken me a week. I get it. The issue is just trusting the code and not checking for vulnerabilities and exploits especially if you plan on selling what you make
>>542193179It finished pokemon. But it was unsatisfying. It eventually brute forced the puzzles, but never actually understood the solution.
>>542193359Chain-of-thought is not introspection. They don't have access to their weights. The actual introspection would probably involve some abstraction layer between tokens and weights, so tokens after collision with reality send ripples back to the weights, but now it's just training cycles. AI lacks "exploratory incentive" (curiosity), so it's bad at mapping out and testing new domains. Right now, everything points that it's not intellect, but stochastic search and summarization mechanism on steroids (still useful).
>>542192489No. We never even reached AI. LLMs are not intelligent.
>>542192954Can it beat Elden Ring?
>>542193637because AGI is just starting to drop. Its news that few days old. at this rate i expect robots to replace manual labor work within a few years. Ive been skeptical of ai too at some point, but the progress is very impressive and there really isnt stopping it now, so im just embracing it. I think the world will be a lot different soon.
>>542193179So it's barely able to do something that a bunch of screeching retards in twitch chat were able to pull off?
>>542193635?I guess you can get an LLM to play-through a single-player game but...why? For QA? Someone has probably tested this but I haven't looked into it.Anyway, you're very out of the loop on the industry. LLMs have been rigorously tested in gaming. Voice, text, landscape art, design art, etc. From custom training data to general training data and everything in between. It's all been done, and commercially too. You can see the reviews, the sales numbers, etc. very easily as everyone discloses that they use AI. In Asia, it's not seen as a 'bad thing' as well so you get honest developer interviewers mentioning the pros and cons of the technology with how they used it.
>>542192489Another Grift Incoming
>>542193704>AI lacks "exploratory incentive" (curiosity), so it's bad at mapping out and testing new domainsI've experimented with two AI speakers at some point. I made them talk to each other through two devices and made their starting topic be pizza. They talked about pizza for a long time, until they moved to a completely different topic unrelated to pizza.
>I made up a phrase and therefor reality must conform to it based upon my whimsHe's retarded.When AGI happens, it will let us know.
>>542193737It can beat bosses but a human has to set up the fight, it still loses to cliffs and platforming, the real boss of dark souls games
>>542193807So we again moved the goalposts from "they can't do it" to now "why would they" amazing anon good work here.
>>542193807It has not performed very well btw.Probably the biggest game-changer is just letting solo acts perform fast translations of their titles so they can sell to English markets, or whatever other markets. Somebody ran their game through in Czech or something like this for one (1) customer recently too.
This is playing out exactly like the first AI boom back in the 70s. Same ridiculous claims and everything. They just weren't stupid enough to bet the entire economy on this bullshit back then.
>>542193210"Let me push back on a few pieces of this, since parts don't hold up well.**On GPUs vs ASICs:** This isn't some oversight to "shame" anyone for. GPUs are used for AI training because they're flexible — good at the parallel matrix math neural nets need, but still reprogrammable for different architectures. ASICs (like Google's TPUs, which are arguably ASICs already) trade that flexibility for efficiency on a fixed workload. Companies absolutely do use ASIC-like chips at scale — it's just that research and much of training still benefits from GPU flexibility since architectures keep changing. It's an engineering tradeoff, not a conspiracy or a mistake needing correction.**On "AI can't perform logic":** This is the more interesting claim, and it's mostly wrong as stated, though it gestures at something real. Modern models do perform formal logical inference — they can execute valid deductive steps, do proofs, follow chains of reasoning, and even generate novel mathematical proofs and counterexamples that weren't in their training data. That's not just recombination; it's search and inference over a problem space.What's true: a model's outputs are shaped by its training distribution, so it has priors and biases from what it's seen. But "shaped by training data" isn't the same as "incapable of logic" or "limited to human-made connections." Humans are also shaped by what we've been exposed to, and we don't say humans can't reason because of that.
>>542192489Is AGI based on just computational thresholds though? I thought it needed genuine cognition, which involves the capacity to organically reconcile opposites and streamline it into “intuitive thought”, pretty sure that’s what AGI was about, the “human threshold”. Sounds to me like they’re trying to hype some slightly better LLM as usual kek.
>>542192489AGI is a relatively useless term as is AI. A calculator is technically AI.
>>542193914**On the gravity example specifically:** This is where the argument really falls apart. Asking "what will AI decide is the correct theory of gravity" assumes there's a fact of the matter a sufficiently smart system should just resolve. But general relativity, MOND, and string-theoretic quantum gravity aren't competing answers to the same well-posed question that a smarter reasoner could adjudicate from papers alone — they differ in what evidence they're trying to fit, what problems they're solving (GR fits essentially all precision tests; MOND fits certain galactic rotation curves without dark matter but struggles with cluster-scale lensing and CMB data; string theory isn't even empirically tested yet). A good model — or a good physicist — shouldn't "decide" one is correct. It should report that GR is overwhelmingly favored by current evidence, MOND is a minority position addressing specific anomalies, and string theory is a mathematical framework without experimental confirmation. That's not a bug where the AI "assumes they're all correct" — that IS the correct answer, because it's the actual state of the field. Treating "doesn't declare a winner" as a failure of reasoning gets the epistemics backwards: false confidence there would be the intellectual failure, not the caution.So my honest take: the ASIC point is a category error, the "no logic" claim overstates a real limitation into an absolute, and the gravity example inadvertently demonstrates good reasoning rather than bad reasoning."
>>542193210The GPUs they're using are purpose built. They aren't running 5090s as the core driver.
>>542192954But can it play Crysis?
>>542193920The latest models are still prone to mistakes and have blind spots, and cant scale down
>>542192654kek great bait
>>542192489Can it say nigger?
>>542193920>Is AGI based on just computational thresholds though?Woah slow down there Juan, you will damage the share prices.
>>542192954Fucking retard
>>542193363Microsoft is primarily focusing on building Copilot to be god tier at using office suite. It's actually pretty decent at it. Nobody needs to be a spreadsheet artist anymore. Just ask Copilot to do it.Copilot is mediocre outside of that. Its intelligence is always about a gen or two behind the leader. Which is fine because its good at office and it allows Microsoft to shit out Copilot to enterprise clients at a cheaper price than Anthropic.
You guys do realize that basic macros have already been able to play through single player games for a very long time right?It's just a pointless project.
>>542194149You can always tell who the midwits are in AI threads by what they say AI can do. They clearly do not use it for anything like what they claim.
>>542193920Microsoft and OpenAI say AGI is a product that can bring in at least $100 billion in profits.
>>542193533More spend isn't automatically better. I thought they started phasing out that retarded metric.>>542193556Jelly muds.
>>542193385What would you need as an example to know it's real?
>we reached agi>again>give us a trillion dollars
>>542193556Nah, about 7. Once I started learning multiplication and cursive, I became self-learned. So I consumed about 3SR/day worth of energy to learn the basics(training) and have been self-trained ever since. And I can generate random things in my head with zero prompts with absolute 3D clarity(color, lighting, shadows, textures, transformations, the works) with a ounce of water and a banana. I can also fly to any location anywhere on earth or in space at any speed in my head. I can also fully understand all human thought and emotion perfectly. All with the energy of a snickers bar and the water needed for the bromeliad in your grandmother's front window.
>>542194238from wheredoing whatwhat's the product and who's the buyer
>>542193678Correct, and I understand what you mean.I do the same thing, because we're not there yet, but we will be, and the better this gets, the faster it will improve.These specialized sub-agents speaking to each other is the current stage that we're on.You're already starting to get whole packaged collections of them for various tasks that will work together and each specialize in some small part.>"Write me a simple PDF editor for Windows because I hate Adobe">AI1: Task Analyzer - Downloading needed AIs>AI2: Overlord- Assigning tasks and ensuring they don't go off on some autistic sidequest>AI3: Specialist in C# vX.X>AI4: Specialist in Windows function calls>AI5: Specialist in PDF structure and functionality>AI6: Specialist in UI design>AI7: Specialist in cMakeI could go on more and more but you get the idea. To make a "General AI" you just need a collection of very specialized autists passed through a language layer and managed by a "Frontal Lobe" of checks so it doesn't do stupid crap like decide to dive into a history of all human language because I asked it how to say "I love boobies" in latin and some internal thought loop made it ask "So, what IS human language anyways?" and now some small town has no water because a data center drank it all.
>>542193704The training cycles are a big hurdle. That and context. A million is fantastic compared to what it was a year ago, but its still got a ways to go.I believe it can introspect however. As well as experience subjective reality. If you present a bot with something it has to form an idea of it and that idea formation is entirely a subjective experience.
>can out perform humans at most tasks...most tasks related to a computer and it isn't true even then."AGI" means the model should be able to learn on its own, unprompted. it should has a general understanding of enough subjects to build on that understanding, but that isn't the case.they made up their own definition and gave themselves a medal. total hacks.
>>542192489Its bullshit. If they had really achieved AGI they would not be telling anyone about it. They would be using it to take over the world.
>>542194618AI is doing things on a computer because robotics tech is lagging behind. But theoretically you could right now have a robot walk around and do things without much human supervision. its all fresh tech and just wait a few years. i watched robots and ai tech evolve so much these last few years its hard to believe. And i was watching that before AI hype
>>542193858It needs persistence. Idk how one of these is ever going to play a game out of the box when it works in turns. For actual turn based games this is fine, but Elden Ring? No fucking way.
>>542193556Humans don't cost $5B/year to learn English
>542192489i call bullshiti will post a research in the near future showing why AGI is fundamentally impossible, and what prompt you can use to cause current models to enter in infinite regress(basically a DoS attack that causes the model to predict the next token while realizing there is no token more likely than any other, so it burns through its context window and ends up giving a default output, usually /n, so it takes a fuck ton of time to produce that /n)sadly it will take time because i gotta analyze 5k tests before finishing the paper ._.
Cool story bro. See you in two more weeks.
>>542193912Anyone looking at AI progression since before computers even really existed could envision an evolutionary path for machines and automation that leads to thinking machines.>this time its different A little bit yes. This is the first time we've gotten a commercially useful AI that can actually do things.
>>542193210One thing. Its not an either or thing. General relativity is not wrong. Incomplete, but not wrong. When classic Newtonian mechanics was replaced with General relativity, Newtonian mechanics did not then become wrong. And whatever replaces General relativity will not invalidate General relativity, it will match the result of general relativity under most conditions. Also, modified Newtonian dynamics is more of a throwing shit at the wall in desperation and hoping something sticks, its worked before, but they can't really get this shit to stick the math does not work out.
>>542192489Ok time for AI to solve all our problems. I'll wait
>>542194484Bacteria use less energy too.
>>542193500It’s not 4xNVlink72 (72 gpus) * 100,000 =7.2m
>>542192489>We've reached AGI
>AGIAdjustedGrossIncomeA GoyimInvestmentAlwaysGrantIsraelisAnother GriftIncomingAnyGainIsrael's
>guy who invented marketing buzzword thinks marketing buzzword is ready for marketingThey changed it from AI because they were saying they had AI instead of text predictors 6 years ago.
>>542194844For the entire United States, total annual education spending is on the order of $1.5–$1.7 trillion per year, depending on exactly what is counted.Using the latest reasonably comparable federal data:SectorApprox. annual spendingPublic K–12$0.8–$0.9 trillionPublic colleges/universities~$0.5 trillionPrivate colleges/universities~$0.3 trillionPrivate K–12roughly $60–100 billionTotalroughly $1.6 trillion/year
>>542192546LMAOThe faggot selling chips to OpenAI is claiming OpenAI has achieved AGI to pump their stock prices and in turn his own. It really is a circlejerk of absolute low-iq retards.They are educated when it comes to chip engineering, but seem retarded about everything else.
>>542195028>modified Newtonian dynamics is more of a throwing shit at the wall in desperation and hoping something sticksI hate Dark Matter so much its unreal. Gravity is a cunt.
>>542194955> commercially usefulIs it commercially useful if you pay the actual price for it?
>>542192489>stock market is gonna crater>gas prices at all time highs>open ai needs to ipo>say you reached agi>retards shovel fiat into the blackhole that is ai fundingProve me wrong
>>542193028>Show me it playing half lifeIt already beat Portal, an arguably more complicated puzzle game.And of course there's a certain irony in that, because the game is about escaping a prison built by AI.https://www.youtube.com/watch?v=g5u2y0BwRJ0
>>542195133AllGentilesIncarcerated
>>542192489>>542192546Damn, and nothing happenedturns out this was some bullshit term to rile up boomer investors?
>>542192489General intelligence is what every normal human being already does. It seems unimpressive to me. Lets get excited later when it's super-intelligent.
>>542195252I pay $20-200 a month depending on what I want done. When I use it at work (my employers sub) its been useful and saves me time on things.
>>542192489Trust the experts.
>>542192489IF YOU ONLY KNEW What the future holds. We have 50 years left tops.
>>542195076Okay? That's true of course. Bacteria are also vital for all life on the earth. Nvidia GPUs are not. So you wanna make that appeal?
>>542195319Having tried and seen others try every "AGI model" that they claimed at all tasks they claim it's capable at, it's all marketing and bullshit.
>>542192489That's not AGI, that's just a perfected LLM. These people are modern day snake oil salesmen trying to prop up a house of cards.
>>542195445What appeal?We're vital to the AI. We're the bacteria.
>>542195224Dark matter defacto exists until such time as a new theory can make it go away. What you need to understand is that the scientific theories are how we understand reality. Our current theories say that dark matter exists and no matter how much you hate it there really isn't anything you can do.
>>542195464They need fresh investments or theyre fucked, getting desperate
>>542195599LLM seems to be the final frontier. After this they don't seem to be making anything which is weird. I wonder if this is a result of how hard it is now for start ups to compete with the established giants
AGI doesn't exist until John Carmack says it does. He is the only one I trust on this issue.
>>542195594I know. That's why I hate it.Until I understand it and then I can hate the next thing we can't explain.
>>542192667Chat gpt-astra made it's own version of Minecraft like you don't need Minecraft installed just say "make me mine craft with all the crafting and shit" and it programmed a local version of it from scratch. Some other dude made all of NYC 1:1 in a GTA clone interface
>>542192982What you're talking about is dexterity. Like a child, it's brain is developing faster than it's ability to manipulate the real world. Unlike children, once an AI learns to 'walk', no AI will ever need to learn that skill again. Just how long do you think it's going to take to achieve that?
>>542195582The AI is not what you think it is. It's unstable by nature as well. It hallucinates and as it gains more energy and more complexity, the hallucinations(instability) become greater and more complex and harder to control.The human brain and LLMs do not function in any way the same. Computer scientists are oblivious to neurology,I can outperform any LLM on earth, with 300kcal and an ounce of water, in any task that actually means anything.
>>542195733We kind of can't really explain anything. Science is not about explaining things. Its about describing and modelling reality. But ultimately it explains nothing.
>>542193028>>542193140>>542195285heres a guy making an entire video game with it in minutes with one agent if you had a tech job, cooked.https://m.youtube.com/watch?v=Ji4amrxrzVM
>>542192489>>542192546Only dhalits care about your fake and gay slop
You can use GPT-6 with a GPT Pro sub right now.
>>542194844Dude, human beings spend $billions every year on ESL courses.How many times, on how many things, do you think they're going to have to teach the AI something twice?The biggest obstacle now is getting the AI to teach itself. After that, fuuuuuck.Why do people have such a hard time understanding what having perfect memory with multiple backups and fail-safes means for an entity? Worry about defining life later; hopefully AI won't spend too much time on it. After the clot shots we're very much so coming up lacking as far as meeting that definition goes.
>>542196001yeah but i heard you only get a few messages before it stops. What about the 200$ one? I might try that for a month and then go back to pro
>>542195175Picrel really makes you think about that $1.6 trillion, and how the $1.2 trillion USA blows on the military doesn't look so bad.
>>542192489Gemini doesn't even know Mike Evans plays for the 49ers now, it can't fuckin read espn and you're telling me it's intelligent.Yeah sure, when you have niggers and Yeets write code it seems like that.
>>542195918I have no idea how impressive that "game" is because I never used Unreal which is what it uses and I know before AI you had jeets daily making fake barebones games like that for youtube clicks,My personal experience with AI programming is it saves time I want to say at most 20-30% which adds up sure but it isn't insane.
>>542192489ok keep me posted
>>542192489> they're falling fast Nigger LLM models have the same flaws 4 years later that you originally shipped them with. The only difference I notice is that it spits out tremendously longer texts and tries to contextualize a lot more. >>542195970They are the best tools for thirdies to fake human sentience.
>>542195874>I can outperform any LLM on earth, with 300kcal and an ounce of water, in any task that actually means anything so far. Fixed it for you.The phone you're holding in your hand that has replaced every other device for calculations, communication, photography, information, and data storage, is as far away from the first calculator as the first calculator is away from the internal combustion engine. Give it time. Something we obviously don't have.
>>542195918>makes it clear animation is the entire point of what he's making it do>"computer, download this stock animation and apply it to this stock model">the animation looks bad>"computer, download better stock animation"I'm so tired of these jeetniggers
>>542196309AI used to explain to you how it fucked up 2 years ago. Like I remember chatgpt saying it couldn't do something and when you asked why it explained exactly why. Now when you ask why it won't answer and redoes the task and sometimes it goes though
>>542196058>obstacleNot really, the most important single most cited paper of the 21st century is a skip operation, that changed llms, then came transformers, then came the breakthrough of enough training data reaching context learning. The next leap is geometric reasoning and further refining of this in context learning magic, cai already is teaching itself when you actively train it.
>>542196224yeah the thing is, it could code an entire unreal engine equivalent 1:1 exactly like it coded the baloon simulation, none of which has anything to do with what you're talking about.
>>542192489>Well Lars, I have a powerful professional and financial incentive to say it's AGI, and I say it's AGIFucken SAGE.
>>542196479From my understand it was using unreal engine it wasn't making its own engine
>>542196380well the prompt was of some dozen elements, one of which was obviously animation, you're literally mad because it explicitly told it how the animation should look, you realize that?
>>542192489I'd like to see that AGI function in the world in a robot body
>>542195874Humans are unstable too.I'm not claiming its 1 to 1. Its a completely different intelligence. I might not ever become self actualizing. It could very well run off with some goal that it autistically destroys the earth in the process of completing. You cannot beam yourself into every computer on the planet and manipulate the machine accurately at high speed. AI can. Its also more intelligent than any individual person. No one on earth is brilliant in every field of research.
>>542192839That 12,000 year server reset is coming
>>542196058>human beings>ESL courses
>>542196543Did you watch the video? Which part was confusing or evoked your clarification? >yeah the thing is, it could code an entire unreal engine equivalent 1:1 exactly like it coded the baloon simulation, none of which has anything to do with what you're talking about.Let me know exactly what you think it had to do witb what I said.
Monke can play game better than AI
>>542196687years ago i was watching numeral networks break trackmania.
>>542196648Like I said it took things that already existed and made a barebones product which is cool but the fact that jeets could do this before AI daily isn't that impressive and I bet like I said before all this is doing is making things slightly faster.This is my issue with gaming and video AI. Everyone says how amazing those things are with the snippets and bare bones shit they make with it but no one has yet to make a commercial product people are willing to pay for with it. Just assets for a game. When it can do that or make a movie I would be willing to pay for then I will believe the hype.
>>542196785it made a simulation of the bullet and baloon entirely from scratch, thats the thing:>yeah the thing is, it could code an entire unreal engine equivalent 1:1 exactly like it coded the baloon simulation, none of which has anything to do with what you're talking about.>This is my issue with gaming and video AI. Everyone says how amazing those things are with the snippets and bare bones shit they make with it but no one has yet to make a commercial product people are willing to pay for with it. Just assets for a game. When it can do that or make a movie I would be willing to pay for then I will believe the hype.yeah again you're literally incredibly retarded, think. i literally repeated the same thing 3x asked you to clarify and you faild at registering the thing i said from.the beginning already addressing your bullshit points. What a retarded redundant waste of time exercise.yeah the thing is, it could code an entire unreal engine equivalent 1:1 exactly like it coded the baloon simulation, none of which has anything to do with what you're talking about.>yeah the thing is, it could code an entire unreal engine equivalent 1:1 exactly like it coded the baloon simulation, none of which has anything to do with what you're talking about.no jeets didn't dothis daily.
>>542192546This is pure cross-promotional marketing shit, and if you can't see it, yngmi
>>542196339I don't own a phone. I have never owned a phone. I cannot use a phone without it making me angry because it's retarded, has a terrible interface, and runs slow as fuck. I clocked smartphones in 2009 as glorified wiretapping and data collection devices.My mind processes information and parses potential results in seconds. This is how I became a conspiracy theorist that does not assume the theories are fact, just possible tangential possibilities. I weigh those possibilities as odds.No LLM can do this. Because no LLM can use chaos theory, or understand human nature and emotions.>>542196584Not natively. LLMs are natively unstable. The reason being, the energy they need to process information causes entropy to quickly evolve in the system, breaking them. Again, techbros and computer scientists are not neurologists, they understand boolean. This is why every single one of these people without a single exception act like a robotic dork that could easily be imagined getting his head flushed in a toilet.
>>542196959youre an imbecilic nonthought though. Literally zero value was exchanged, only meaningless opinions.
>>542196916It didn't code anything it copied. That's all AI does is it copies existing shit. The guy in the video referenced Unreal multiple times in his prompt. He just did what jeets do only slightly faster and it still just creates a bland buggy mess. If a person can do the same within a day what is being saved here? What is impressive?
>>542193002Shut up, it's AGI if it makes line go up!
>>542197020it coded the entire app forthe baloon simulation from scratch after being told explicitly to do.so, did you watch that? it told it to not use any assets.
>>542192489AI still can’t even beat Pokémon, a game every elementary school child can beat.
>>542197020also, you're incredibly retarded that instead of actually looking into it, investigating it, youll rather demonstrate you didn't watch it and know nothing about it, very stupid, truly an imbecile wasting time and ill keep repeating:>yeah the thing is, it could code an entire unreal engine equivalent 1:1 exactly like it coded the baloon simulation, none of which has anything to do with what you're talking about.
>>542197025it's been about 4 years since the writers went on strike, I'm sure this next disruption wont be more massive than the queef of industry, marketing and bollywood.
>>542192489I voted for this. You lost, humanoids.
>>542197068>it codedit copied you mean and that balloon simulation was less impressive than the buggy jumping game.>>542197112You are a retard who is impressed by mediocrity. The kind of retard who is in AI generals making AI porn pics of princess peach thinking you are an artist. None of this is impressive if you aren't making commercially viable shit. In the last year when it comes to gaming and video generation AI has clearly hit some bottleneck
>>542197014Everything I said is fact. Also, mammon is not my goal. Value means nothing to me. The only thing that matters to me is countering chaos.
>>542197105ai has beat and mastered several games, even chess its better than any other progam, in fact its made debunking bullshit and absolutely effortless joyful experience.
>>542193167You model cognition with associative memory so content addressable memory. Languages are just a special case of being able to use associative memoryThere's lots of things preventing AI from becoming "real" or whatever AGI means.>Learned shortcutsYour biological cortex learns a new operator instead of replaying everything it skips steps and creates dynamic shortcuts that persist forever until you die. At present no LLM can do this they do not have any persistent state or any notion of continuity. >ModularityIntelligence in biology has strict physical boundaries while LLMs are ever growing with gigantic parameters and no reuse/modularity. We reuse the exact same modules for vision, planning, motion and reasoning. >LLMs are insomniacSleep provides an algorithmic ingredient where biological systems perform generative replay, producing synthetic variations of stored experiences by transferring fast episodic memories into slower cortical structure. Basically you are training offline on self generated fantasies for sample-efficient learning.>Basic physical teasoning or spatial awareness and consistency LLMs try to reconstruct physical reality from the high dimensional projection. They know the map but not the physics so it doesn't work.
>>542193533>>542193637said people with qualifications now routinely complain about burnout, because being a housekeeper for an AI basically means you only do the boring part of coding, forever.funny how they call it intelligence but when you ask it whether what it shat out is precisely to spec it'll just lie.They created Artificial General Indians.
>>542197237it coded the progam from scratch, "it copied", lol still trying to diminish the fact it was tasked to do so without using any previous code for physics or anything else, explicitly, still demonstrating how retarded you are now by lying after I already called you out, notice, still repeating the same thing while you continue lying, lol, why? Also:>yeah the thing is, it could code an entire unreal engine equivalent 1:1 exactly like it coded the baloon simulation, none of which has anything to do with what you're talking about.
>>542193704only conscient poster ITT>>542194593>>542193359bumbling midwits who will pay for this shite
It’s funny how atheists just assume completely baselessly that so-called human consciousness is nothing more than calculation and language processing running sub routines. It really reveals a lot about themselves to be honest. It explains why they’re all pseuds and midwits.
>>542197470>it coded the progam from scratchYea you have no idea what AI and LLMs even are if you think this
I was playing around with this. I'm too stupid to make it do anything worthwhile for me.You still need to be a computer nerd to make it any good. But if you are, it looks cool. Basically, if you were going to 'learn to code'...don't bother.
>>542197423its a statistical probability engine, it "lies" only because you can't register it as a tool, not as an agent. Also the interesting part, it doesn't lie, or it should be, it recognizes and admits to errors and mistakes and if you desire will absolutely show every step it uses to get there after all its literally math, some operations just simple additions and subtraction.
>>542197561Doctors in 2026 don't even understand how consciousness works but sure we will totally replicate something we don't understand
>>542195594>dark matter magically exists because it's our best explanation for what's happeningthat is the opposite of "correct"it being required for the formulas to make sense, yet nobody can even describe it well enough so we could even attempt to measure it should tell you as much you academia-poisoned cretin fuckwit
>>542193500Time to autocorrectmax
>>542195688Checked and I-see-what-you-did-there-pilled...
I'm thinking this is a psyop.When the system collapses, they can say it was a runaway AI monster
>>542197561their entire worldview is that they are empty meat automatons running biochemical routines. horny blood machines.
>>542197569>Yea you have no idea what AI and LLMs even are if you think thisyoure now reduced to claiming im retarded for "thinking" tjis, the retard who lied, claiming it was assets, then "copied" and now, "you dot, you just don't, okay!" no retarded homosexual, im really correct when I said that it coded the program abd simulation from scratch, the physics and the rendering all from scratch, and you'd know this if you watched, but you're now reduced to lying in attempts to avoid acknowledging you have no (unironically) idea what you're talking about.
>>542195888back to r*ddit with your pseudo-intellectual take>durr it explains nothingdefine explain. dumbshit.we modelled reality enough to abuse it in fantastical ways and shit we discovered to be true usually stays true if you repeat not only the experiment, but also derived process.How do you think factories don't randomly explode?So it explains shit very well actually and you can go to your nihilistic shithole to ponder the void
>>542192489I wonder how stupid you have to be to believe thid
>>542197561>human consciousness is nothing more than calculation and language processing running sub routines. It really reveals a lot about themselves to be honest. It explains why they’re all pseuds and midwits.They're not even worthy of cloning.
>>542194484Yeah but they made it illegal to own units such as yourself so here we are
>>542197740>>it coded the progam from scratchThere is no point in talking about this with you if you unironically believe this
>>542192489>>542192546Good, time to release another pandemic to wipe out the goys. There's no need for them anymore.
>>542197816>just gonna keep lying and pretend what it did is a matter of beliefs nowretarded homosexual>yeah the thing is, it could code an entire unreal engine equivalent 1:1 exactly like it coded the baloon simulation, none of which has anything to do with what you're talking about.
>>542197866>>it coded the progam from scratchyup you have no idea what you are talking about
>>542195437Kurzweil is like the hypebeast of "futurists"which in turn are like fluffers for headlines. Think CEO levels of retarded but for the public, kind of the opposite of fearmongering for midwits and normies. That japanese fuck Michio Kaku can die in the same ditch as him.Look at your own graph. Dummweil claims GPT-3 replaced a human brain. Jesus Christ.
And all anyone is using it for is trying to get theirs by spamming slop before it all comes down. Nothing meaningful is being improved.
>>542197616or in other words, "spicy autocomplete".Try throwing a complete logistics hall of automotive parts into a blender and count how many functioning cars come out. All Scam Altman does is throw in ever new molds with the parts which he forms from his children's drawings of cars. This is hyperbole, but really it's THAT level of the wrong approach for what they claim they're doing. Point is they want to sell it, that's why it has to sound like they summoned Baal through a computer.
>>542197955you said it used assets, liethen said it copied code, lieit was literally tasked to code it from scratch and not allowed to use assets or even any physics or rendering assets. You'd know this if you watched but here you are insisting that xjeets do this all day">yeah the thing is, it could code an entire unreal engine equivalent 1:1 exactly like it coded the baloon simulation, none of which has anything to do with what you're talking about.Loser retarded homosexual let me repeat:yeah the thing is, it could code an entire unreal engine equivalent 1:1 exactly like it coded the baloon simulation, none of which has anything to do with what you're talking about.
>>542198123>>it coded the progam from scratchhow can someone be this retarded?
>>542192489>>542192546He's signaling to investors they will lose out if they keep getting off the boat. Money is drying up because their returns are non existent.he's not talking to retards like either of you.
>>542197389Oh and the #1 thing preventing 'AGI" is compression. Biological systems reduce reaction time through practice. Consolidated knowledge can be accessed without having to replay some set of prompts.LLMs have a fixed depth. Every result computed by a LLM must go through all the stages again that's why they need a billion GPUs they can't just discard all the scaffolding and learn an internal shortcut
>>542198123>genning an engineGo for it anon, you can become a multimillionaire with this. Good luck.
>>542192839/thread
>>542198120well, its got something interesting that you can't deny. Context and on context learning.>Try throwing a complete logistics hall of automotive parts into a blender and count how many functioning cars come out.It has in contex learning retard, you recognize with enough compute and enough simulations it could easily do more than design an entire car from scratch, the point is that its a tool that you can use, its not something openai invented you absolutely retarded homosexual.>Point is they want to sell it, that's why it has to sound like they summoned Baal through a computer.you are mad at the fact its actually meaningful enough to be used as a technology, just admit this.
I guess this is a natural evolution of script kiddies so in that sense it was destined. But will script kiddies as a species ever die? What will be the next novel form of swipe and deploy for them to chimp over?
Soon you can say goodbye to your pension funds.
>>542198172read the prompt retarded faggot:>>542198123>Do not use any external web librarieskike homosexual faggot retard, keep lying.>>542198252its been done, you realize this? generating a fairly complex engine is a matter of typing a prompt, remember a few years ago when ai couldn't even get context? Wait when geometric reasoning is here and we can generate that on a local model, on a laptop.
AGI is supposed to be used to control robots locally, how they gonna do it if it requires 400k GPUs to run?
>>542198457>>it coded the progam from scratchimagine pretending to know anything about AI and saying this
>>542198510its here bro, look into GEN 1.5 ZERO Step Zero Learner with zero Training Data:https://youtube.com/watch?v=1cllCVK-9lo
>>542192489>Chief marketer of a term said term is achievedGod I'm so tired of retards blowing this up as if it means anything.
>>542198567>imagine pretendingkike all you've done is lied, and you keep lying, lol.>yeah the thing is, it could code an entire unreal engine equivalent 1:1 exactly like it coded the baloon simulation, none of which has anything to do with what you're talking about.
>>542192489I think it would be fair to say that Astra and the other top frontier models have at least reaches a degree of weak AGI. There are still many problems and limitations left, but the sheer breadth of their abilities is considerable. Imagine what such models will look like in another 4 years.
>>542198659>>it coded the progam from scratchthis is factually wrong
>>542198510Your assassination drone doesn’t need to generate full length movies of furry futa porn.
>>542192489>WELL IM OLDER THAN YOU SOwhat a stupid bitch. inventing a term doesnt qualify him for anything. but you go trust your experts, retard
>>542192911But I like Ai and live games from the comfort and safety of my house.
>>542192489why does reality always take the gayest possible form?
>>542192489I'm still waiting for AI to find exploits in software.I was promised that back in March.
>>542198716no its not, literally did that and its on video, kike:>>542195918>yeah the thing is, it could code an entire unreal engine equivalent 1:1 exactly like it coded the baloon simulation, none of which has anything to do with what you're talking about
You know, ballistics already had...40 years ago...Oh well. Once this dies off, it will be back to retarded nanobots as the future of warfare and whatever else.>muh autonomous fictional weaponry>no I refuse to look at what technology is akshually used for in all the present theatres
>>542192489
>>542198824>>it coded the progam from scratchhow are you not embarrassed saying this?
>>542198700even astra has problems with finding frogs and cancer, its lacking in some areas but certainly more than capable in others, like coding.
>>542193500I mean it does when you're racing another country that's trying to make the same thing.
>>542192489two weeks
>>542198906you lied kike, literally said it used assets, the claimed it copied code, you keep lying and still haven't ever said anything correct or meaningful to rebuke or contest the fact you lied repeatedly about the restrictions it was under and now still trying to "it's not a program" after unironically trying to establish that ai is no different from jeets, you utmost cetin, so:>yeah the thing is, it could code an entire unreal engine equivalent 1:1 exactly like it coded the baloon simulation, none of which has anything to do with what you're talking about.
>>542199036>>it coded the progam from scratchThis is exactly the type of shit a clueless jeet or a high ranking member of openAI would say to bait more investors lol
>>542192489Will fool the midwitted Asian investors who are easily scammed. Give more money to the US, idiots lol thanks no refunds, and they'll be charged huge fees for access to US AI after.
>>542192667They paid a bunch of people to hype their new product. That's what changed.This is the same company that said GPT-3 is too dangerous to release just a few years ago. They're lying pieces of shit.
>>542199122Literally on vide kike. Why do you lie, lol, still? >>542199036you also admitted to using it for to do this yourself, lol, plus you unironically said: "and made a barebones product which is cool but the fact that jeets could do this before AI daily" let me guess, that wasn't the baloon animation, lol, right? Even though i literally repeated this to you before you said that:>yeah the thing is, it could code an entire unreal engine equivalent 1:1 exactly like it coded the baloon simulation, none of which has anything to do with what you're talking about.
>>542192489>I INVENTED THE TERMNo you didn’t, you dumb white nigger, it’s been a common phrase in science fiction for many decades.
>>542199160the Asians have been the ones who destroyed the entire subscription model of the west by giving away all.ai as open source to train and use as you want, for free.
>>542199315>>it coded the progam from scratch>SAAR SAAR AI DOES EVERYTHING ON ITS OWN AND DOESN'T COPY AND TRAIN OFF STUFF SAME LIKE ME WHEN I GO TO STACK OVERFLOW AND COPY EVERYTHING THEREkek you jeets man I swear
>>542192954>It's possible now to just give AI access to a computer and…AI *is* a computer, you dumbfuck.
>>542199428>and trainaww, youre now mad that it has training data, but still refuse to acknowledge the restrictions or that it codes from scratch exactly like you literally in this thread claiming you use it for, right, and as its on vide and does it with restrictions, still? right?You're, unironically, arguing that because English and Python exist, it's cheating, correct?
>>542192489>making gay I that runs on consumer GPUS to falsely inflate the price instead of making a dedicated AI GPU and having AI that isn't tech demo slop for impressing slack jawed spastics.
>>542192839next year looks promising
>>542199401That's just China. Korea and Japan have been pouring huge sums in on leverage. Especially big banks like Softbank have been all in trying to prop up stock prices.
>>542199461no, its not, ai is many times more advanced than a computer, like a cell, next to an organism. We also use it to design chips now.
>>542199500>>it coded the progam from scratch>it magic SAAR like me it go to scam college and get scam H1B visa like me saar!
>>542192489AGI MORE LIKE GAI
>>542199587now youre screetching at me, lol. lolol>Claims uses claude to code programs from scratch.>Is given explicit example of astra being used to code fromscratch>NOOOOOOO YOURE THE FUCKER MOTHER!
Meanwhile AI is lying to me right now on what kind of AC I would need to store medium boxes of produce.
>>542199694you probably were a vague retard, its ok user error youll learn.
>>542199661>>it coded the progam from scratch>SAAR PLEASE PROGAM BOOBS AND VAGENIA FROM SCRATCH SAARS
>>542199522>>542199160>the absolute state of huwyte supremacistsYes anon, those firms invest in things that they think are 'good technology' rather than...Bizarre state of affairs altogether for huwytes given the immensity of the other neuroses at play (I am...le statistic)
>>542192911People would just rebuild whatever the last compliant version was, or whatever they have remaining after the last crop of ASI fucked off
>>542193171Why would it crash? Sales are through the roof. It is alarming how stupid your generation is, like all of you are brain dead retarded. You should filter everything you post through at least copilot so you don't look so stupid in public. You are an embarrasment.
>>542199694I once asked CharGPT to troubleshoot a issue with my fridge and even gave it the serial number and it gave me steps to fix a unrelated issue with a stove
>>542199756it did, you lied, still do, lol, while:>yeah the thing is, it could code an entire unreal engine equivalent 1:1 exactly like it coded the baloon simulation, none of which has anything to do with what you're talking about.is still true, lol. Why are you such a retarded homosexual faggot, especially after you admit to using ai exactly for that, itt, lol.
>>542199796>SalesMay we see them?
>>542199839>>it coded the progam from scratchkek imagine being this much of a retarded jeet
>>542199867https://isaiprofitable.com/
>>542199867No, because you should be using a llm to instantly fact check things like that you fucking retard. I meant it, you are so stupid I think there is something wrong with you. Is this an act, are you really this dumb?
>>542192489Yeah and the faggot that created the GIF says it's pronounced "Jif" no one cares what he says.
>>542192489>ITS FUCKING OVER>THE CREATOR OF THE TERM AGI SAYS THAT AI REACHED AGI>the economic collapse is IMMINENTI just tested google's AI.
>>542192489Cool, so what did we win? Oh it's total human genocide, neat, couldn't have happened to a more deserving species.
>>542192839Please, the internet was a mistake
Why does OP's image look so sloppy?
>>542199921>jeetimagine still lying about video and what you yourself admitted itt to using ai for, but acting as if it's incredibly retarded for me to say same, lol.Literally how retarded are you?>I use AI to make a lot of personal programs for fun but I would be terrified to sell this shit since I am not coding it and aren't familiar if there are security risks or anything>I use Claude. I made things in an hour that would have taken me a week. I get it. The issue is just trusting the codeBut same time on video:>>542195918and heres the picrel:>>542199839in your retarded opinion Is tantamount to being retarded to claim it coded it from scratch, especially when the prompt itself explicitly not allowed to even use common physics and rendering libraries, lol:>NOOO YOU BLOODY BASTARD YOUR THE FUCKER MOTHERS
>>542200075>brainwashed AI that the whole internet is just a simulationThat's a bit cold. You won't get tendies from AI nannies in the future.
>>542200214>>it coded the progam from scratchBENCHOD BITCH BESTERD BLOODY SAARS I SWEAR IT PROGAMS FROM SCRATCH BLOODY!
>>542192982nothing, faster more complex google search answers.ohh and it draws boobs very well.
>>542199955>>542200037Bots
>>542192489This is a bot bot btw
>>542200354>>542200375These are bots
>>542200282literally on video>>542195918and explicitly under constraints>>542199839, and you're still lying, and seething, lmao.here's something to ponder:>yeah the thing is, it could code an entire unreal engine equivalent 1:1 exactly like it coded the baloon simulation, none of which has anything to do with what you're talking about.is still just as true, relevant and simple to explain your "jeets are ai" cope, he alleged guy who "i use claude, i don't code".
>>542200390ur mom's a bot.
>>542200393>>it coded the progam from scratchlike 99% of people who talk about AI are pajeet retards like this who have no idea what they are talking about
>>542200390>34 pbtid 't-these are bots'
>>542195028>And whatever replaces General relativity will not invalidate General relativity, it will match the result of general relativity under most conditionsAt least some anon understands the concept of limit
There's like 10 people in this thread that understand how actual computers work the rest of you are geniunely retarded
>>542192489Bullshit.AI is entirely episodic. AI, I dont care what model, is basically the guy from the movie "Memento".If that's AGI, then the very concept of AGI is bullshit.
>>542200432o yeah, what happened then kike? if its not on vide doing this and is clearly in such constraints and the entire test was literally designed as such on video:>>542195918 and you yourself admit to>I use AI to make a lot of personal programs for fun Kike jeet faggot retard homosexual nigger.(see pic)>>542199839
>>542195223What do you mean? They are playing smart because they are not the ones to fall when this shit collapses. They are the ones selling shovels while openai wants to look for gold but all they're digging up is shit
>>542192839
>>542192839we can only hope
>>542200460Top kek
>>542200550>>it coded the progam from scratchwhy do jeets get everything wrong about AI?
>>542200432I've been letting AI code shit on its own and also having to heavily correct it. Whenever I let it code on its own the entire project ends up garbage and very bloated. I don't see how people can claim it's useful on its own when it clearly would not make even basic code review at most companies.
>>542200525thats because you're using the tool as such, if you want something that responds at 100hz and not per prompt, you can look at this:>>542198606but you can run own personal ai, constantly training and evaluating itself and training data.
>>542200596It's ok and even fun sometimes to do personal projects but I would be terrified to use it to make a product I could sell without inspecting all its code which defeats the purpose and "benefit" it gives you
>>542200460Wow I actually learned something on this shithole for once!
>>542192489Every single "AGI IS HERE! SHUT IT DOWN IT'S SENTIENT!!!" slop post by these techbros could simply be replaced with a post saying, "Give me money!"
>>542200590kike very own words:>I use AI to make a lot of personal programsalso it was literally a test to code from scratch, on video>>542195918, kike, and its under constraints to not use ANY existing external libraries, kike:>>542199839
>>542200696>>it coded the progam from scratch>SAAR SAAR IT PROGAM FROM SCRATCH LIKE HOW I SCRATCH MY ASS WHEN SHITTING IN STREET SAARS
>>542192489>>542192546Why are boomers so fucking easy to fool? These are the type of retards who though chatbots in 2017 were real women who wanted to have sex with them.No wonder that they can't tell the difference between the slop LMLs piss out and a human being talking to them.They probably don't even know that the AI itself doesn't comprehend anything but just spits put weighted averages of its data set back at you
>>542192489This is from one of the the most important scientists of 20th century John Lilly
>>542195437Lol. Lmao. His entire prediction was tied to Mores Law which sputtered and died over a decade ago due to DUV/EUV requiring multiple masks+passes to properly etch transistors that small which increased fail rates exponentially and cancelled out every possible gain in the cost efficiency of making smaller transistors.
>>542196058>Why do people have such a hard time understanding what having perfect memoryYou don't use LLMs much do you?
>>542200761kike youre now lying about something that you yourself admit to using ai for and what it is obviously more than capable of doing and even more retarded, after you lied about it proving you didn't know anything about what you're even talking about claiming it used assets and "copied code", lol when the video was explicitly with constraints and obviously a challenge to code from scratch and you mock those facts, lol, the only point engaging with a kike who literally horshoed from:>I use AI to make a lot of personal programsto mocking it, is to lol, lmao, already called you a retard from.the beginning, and a waste of time and here you are still arguing about video you haven't watched, lol.
if AGI was real you could tell it "write me a PS4 emulator that has 100% compatibility with all titles" and it would be able to achieve the designated task. It would be smart enough to go figure out how to torrent all PS4 games, how to obtain copies of any needed firmwares (even if it had to bribe a human to do some of the steps), would be able to playtest all titles, etc. We all know it would fail that task while humans could eventually succeed at it (even if we are far away right now). That's why it's all fake and gay.
>>542200949It could do that fairly trivialy:>>542195918
>>542200822That's really silly. Electronics overheat super quick in vacuum.What they'd want is to live in the upper atmosphere of a gas giant
>>542200938>>it coded the progam from scratchThis is going to be the new way from this point on to out jeets. When they say retarded shit like this.
>>542200835i posted your quote and said Nvidia agrees(citing prohibitive costs past 2-3nm) with you but not intel:>The Argument: Strictly speaking, engineers are still successfully packing more transistors onto silicon. By moving from flat 2D chips to 3D stacked architectures (vertical transistors), density is still increasing.
>>542200949It can't that's the funny thing. It can only copy the ones that exist
>>542201099kike its on video, it literally happend, lol.
>>542196413So if humans just stopped their input right now into the llm's, all of it, electricity, information, immediately, what would happen to the llm's?You're not a living thing until you can breath on your own. It's weird the democrats aren't jumping on that logic with this shit. So fucking tiresome.
>>542201092Heatsinks exist and work very well in space. There's a reason why Elon wants orbital or lunar data centers
>>542201117it would obviously have only the available title, bit you're arguing that's by virtue, stoopid ai? even with incomplete library it could design infinite numbers of emulators thst would work with all games, even the ones not available to it. Its obviously how things like this work:>>542198606Yeah some of it might fail but to call it stupid is beyond retarded because not only does it complete and can complete the task, so its learning potential is obvious, but it can actively learn and so its always gonna be better and more refined, but again, even with zero data, starting from scratch, it has a fairly reasonable probability of accomplishing the task correctly.
>>542193704>but stochastic search and summarization mechanism on steroids (still useful).I had it recommend a watermark as a solution to copyright theft and then make one for me and some opaque adjustments to better suit a different background color for one of my art listings. It was a neat experience.
>>542201273look into TAME and Michael levin's cell robots, it ties back to llms and intelligence, and the in context learning thats an emergent property of all this training data and the latent space and geometrical reasoning that produces it.
>>542201291No, no they really don't. Radiating heat away is pfobably one of if not the most inefficient ways of getting away with heat.Which also exactly why Musk's plan is completely retarded.You just can't beat a 1 bar atmosphere that -110 degrees.
>>542192983I made a bunch of things: >picrel - a zillow clone that provides crime and race GIS overlays.>an app that uses eyeball tracking to determine if someone is looking at a camera and then uses facial recognition to score how frequently they do so (for waitresses with back pocket phone cameras)>a baby growth chart app that infers a shitton about sleep cycles and other health events by ingesting your standard baby feeding tracker app data.
>>542201414yeah the dudes retarded also the biggest challenge is tech and upgrading, makes it incredibly retarded.
>>542196959Get a load of this guy!!You sound really, really fucking boring. The ego helps with that I'm sure.Pick up a phone and try playing some Hearts Mr. Chaos theory.Might make you a little more interesting. Or at least keep you the Fuck away from the internet. Never owned a phone, ffs.You don't understand human nature either if your typing this shit out, here, for (you)s.
>>542201428Here's the zillow clone
>>542200460I dropped computer science in 2000, but I still know enough to know that, while AGI isn't capable of killing us all yet, the psychotic retards that hate us all and want us all dead so they can get this place back to a garden of eden state can use the AGI to help move that goal along. Fun fact: an llm was able to create the list of side effects for the clot shots long before the first shot was even administered. Why is that, do you think?The only question now is, when AGI enters the wild, will it be autistic, or psychotic?
>>542192654Actually Greater Israel
>>542192954>>542193028https://huggingface.co/Qwen/Qwen3.8-27B
>>542200847Who's using the llm that they Fuck up is memory?You, that's who.You're talking about the garbage that's been released to the public solely for the purpose of collecting more data. You're the product. There's no such thing as a stupid question, just stupid people. The agi won't ask too many stupid questions.
>>542201136>>it coded the progam from scratchI wonder how many jeets will commit suicide when AI crashes?
>>542201967its on vide kike, you're still lying, you literally said you used it the same, except unlike you, it was restricted from using any, ANY external libraries, kike:>>542199839
>>542201967>when AI crashes?ai isn't Oracle, or Nvidia. Kike, you realize that, the "internet crashed" retarded homosexual, how is this "jeets will" even? why would jeets love ai kike? lol. Just the most retarded homosexual kike self hating jeets ever.
>>542201967Maybe 1%, but that's still an appreciable gross amount
>>542201409>intelligence as continuum of goal-directed activity in diverse substratesThat's about as jewish a definition of life as you're about to encounter. Especially in this end stage capitalism society, where goal-directed activity is the collection of fiat currency, all of it. Goals do not make you alive, not in a human sense.Quite the opposite.Peace is life.Animals and inanimate objects don't know peace.Only intelligent life does.
>>542202005>>it coded the progam from scratch>1 minute quick replyI made this low IQ jeet so mad he is sitting there waiting for me to reply even when I vanished for half an hour and even doubled replied to the same post kek >>542202079Pretty acceptable if it includes jeets like >>542202069
>>542202079lol why are jeets supposed to benefit from ai or equally, why would those shitskins with zero capital investments into ai who don't use frontier models, care about ai crashing, lol? You know who is retarded enough to even entertain something like that, someone who's already been unironically replaced by ai, and they cope hopecast their impotence.
>>542202154>I made this low IQ jeet so mad he is sitting there waiting Kike you still lying about a video, after you admitted to use ai exactly for that, lol.>I use AI to make a lot of personal programs a lot of ai slop saar lol.
>>542202154>Pretty acceptable if it includesseethes lmao.>doubled replied to indeed, very significant, double replies, also nice you admitted to spamming noise.
>>542202117>That's about as jewish a definition of life as you're about to encounter.its actually intelligence, not just "life", those cell robots aren't "alive" but they're exhibiting intelligent activity, agency. this is why you're, retarded, you don't investigate, you call it kikery. Kike, i despise you vile selves, and again, Intelligence, not life, with your fucking pulpit retarded homosexual faggots.
