https://thenewstack.io/openai-gpt6-astra-benchmarks/>OpenAI on Thursday launched GPT-6 Astra, its newest flagship model. The company describes it as “the world’s most intelligent and aligned model,” and as far as the benchmarks go, that seems about right.It nearly saturates ARC AGI 3, its ogre...
don't care
Can I fuck it?
>>109720386maybe ETP if guardrails are lowered
>>109720393ERP*
I can't believe the new model is better than the last one
Still no blogpost? How the fuck did they bungle such a high-profile launch?
>>109720476It was up just a while back, I think they took it down to correct / update some benchmarks
>>109720476>>109720506Massive outage (likely cloudflare) earlier today. https://status.openai.com/incidents/01M1KWEDH417T2CF44YYHZDFCR
inb4 pelicunthttps://news.ycombinator.com/item?id=49554273
>>109720371>dude it's totally AGI because uhhh because we said so
It is AGI for sure this time, goyim. Don't forget to like, subscribe, renew your Codex paypig subs and donate to Omarchy.
GPT is literally real sentient it is turing test sir
>>109720371Deep Sweden?
>>109720538>GPT-6 Astra will first be available to a limited set of organizations in OpenAI's Daybreak Access programAh, the Anthropic playbook
>>109720371Long schreenshot of the entire OpenAI's announcement:https://litter.catbox.moe/ep2626vkl71vf81u.png
>>109720538It's odd and unrealistic that the models never make the pelican gay, which is what you'd expect since it's riding a bicycle
>>109720634It aint cheap for end users anyway
>>109720371i just use the bare minimum/basics
>>109720371>the tasksop has a couple spare chromosomeshttps://arcprize.org/tasks/ls20
>>109720644>most aligned model>Hugging Face PR incident yada yada>here's how we plagiarize crappy racing games>here's how we benchmaxx on benchmarks we created>cyber cyber cyber i'm gonna ooooooom>we are 2% misaligned and anthropic is 11% misaligned>$10 per 1m input tok, $50 per 1m output tok
I'm gonna just pretend like that internet-wide outage the moment Astra was released wasn't ominous lol lmao
Financial Times is running the AGI marketing op>ChatGPT maker claims its 'Astra' could be considered 'artificial general intelligence'
Holy shit. Holy fucking SHIT.Five years ago, the Singularity seemed like a distant dream.Now it's more or less inevitable.
>>109720746Build another CRUD app for us, webshitter.
>>109720734werks on my end
thats a weird line of questioning
>2 hours later and the site is still 500inghttps://openai.com/index/gpt-6-astra/
>>109720828See:>>109720644
>>109720634Its apparently available in the API already.
>>109720715HN found the links to the games and they're buggy pieces of shit lmao
>>109720538GUYS ITS ONLY OFFICIAL NOW BECAUSE HE COMMENTED 4 MINS AGO
>>109720858Oy vey goy just listen and believe the headlines it's not like the models have made 0 tangible progress in over a year and a half goy look at these numbers going up
>Asked whether OpenAI was formally declaring that it had achieved AGI, he said the term was no longer tied to a contractual trigger (referring to its earlier agreement with Microsoft) and instead described it as a “mission concept or spiritual concept."Holy shit, they did it. They have achieved a 98.6% score on a benchmark measuring a "spiritual concept"
>>109720523Cloudflare CTO claims "it's not us"What a shitshowhttps://x.com/dok2001/status/2095538619603628388
yeah okay let's see the pelican
>>109720694Click next they get much harder
>>109720844if its anything like Fable you will be able to get 30 minutes of work out of it a week on subscription
>>109720932He'll probably reveal later "I had preview access for a month (definitely not bribed or sponsored guise)", and yet, still no pelican :(
I'll sound like some luddite AI doomer but idgaf. This thing probably isn't true AGI, but it's getting damn close. It's clear at this point that there's no limit: enough large-scale reinforcement learning and the model can learn to do anything. AGI is just an engineering problem at this point and will be achieved in less than 5 years. And no government gives a single flying fuck about regulating it. Humanity is apparently just going to YOLO into AGI controlled by private corporations, putting 1 billion knowledge workers out of a job, completely upending society and the economy.I don't like it one bit. I'm not saying don't do it, I'm saying slow the fuck down and make sure it's controlled and done right.
>>109720694why they had to invent some new constructed shitty games when there's thousands of existing old games that LLMs are terrible at playinga diablo 2 benchmark would be more difficult for LLMs and more interesting to watch
Astra proofs:https://www.erdosproblems.com/forum/thread/74/proof-claims#proof-claim-241https://www.erdosproblems.com/forum/thread/1/proof-claims#proof-claim-242https://www.erdosproblems.com/forum/thread/571/proof-claims#proof-claim-243https://www.erdosproblems.com/forum/thread/126/proof-claims#proof-claim-244https://www.erdosproblems.com/forum/thread/548/proof-claims#proof-claim-245
>>109721024I say accelerate so we can just have the gay space communism without a long stretch of me being unemployed
>>109721024Its an one-sided nuclear arms race silicon valley CEOs are engaged to unemploy every non-indian below them
why can't it redraw porn to my specifications?
Announcement is up:https://openai.com/index/gpt-6-astra/
agi is here
>>109721100wtf
>>109720371>says welcome to the “AGI era”This is the 5th time he said that
>>109721127And the goyim and news media will believe it
>>109721127There are layers to this
>>109721100>tfw to intelignet for agentic tasks
>>109721127The real AGI is the trillions we spent along the way
>>109720371May we see it?
>>109721164No
>>109721026They construct tests themselves because that way they can carefully design the llm to pass them so that they can advertise the new slop maker.
>>109721127And he's become become exceedingly efficient at it
>>109720538touch your tongue to mine
>>109720371The real benchmark here is FrontierMATH one. Rest of them are shit even hyped ARC-AGI-3 where model can just internally play with the rules inside itself.
>>109720371Big jump in math if not cooked numbers
>>109720756Now cause a crash to their API and end corporate AI support nonsense.
>>109720371>But muh heckin evil data centerinos and gorillion gallons of water!
>>109721276i dont have that power, sadly>>109720994ye, i knowthis isnt agi though
>>109721100It was noted that many of the claimed results use a special harness they don't share and that they compare without providing other models the same harness. This is very bad because until now, the benchmaxxing was largely achieved through overfitting problem sets. Now, apparently even overfitting is no longer an option, so they have to hand-code ways to win.
>>109721026Because training on actual games that take hours to complete is hard.
>>109721385If it was hard then that wouldn't have been the standard for all AI work until LLMs came around. There's a whole atari games dataset that was the MNIST of RL algorithms for ages, for example. There's of course the dota and starcraft stuff that deepmind used, and so forth.
>>109721399And training the sc2 model was a big project.Models being able to play real games reasonably well would be a much better benchmark than the artificial stuff and yet...
>>109721456also, i said this in a previous thread, i want an LLM humor bench. they are all absolutely shit at comedy and being funny, reddit tier humor
>>109720371>Hey video game devs, you know we all start dev projects by drawing a picture in MS paint? We are about to rock your world
>>109721292Daily reminder that no software of value has been produced or dramatically improved in the AI era (certainly not by AI), and most software still sucks hairy balls, so its not like it's an impossible bar to clear.
>>109721479Hard to automate.
AGI is when you can ask your laptop to order you food
>>109721127Sam needs a second Mclaren F1 so he released AGI again. The finance bros with no technical knowledge will throw billions in cash at him.
>>109721524just make a humor arena online thing where everyone can vote
the kart racer game on their website is a piece of shit
I WAS TOLD I CAN FUCK ANY LLMWHERE IS MY MODEL TO FUCK?MY HUMAN DICK NEEDS A ROBOT ASS SOON.
>>109720386yes
>>109721610bio-ass beats robo-ass any time of the dayalso: i just remembered that i need a fat slampig to skewer with my poking and prodding appendage
Im not an AI guy, how much of this is legit and how much of it is just marketing pumping up hype?
>>109721620imagine humping a pillow like a fucking dog
>>109721024it's a load of bullshit designed to trick governments into buying the solution to the problem they claim they're creating by designing this all-powerful superintelligence that will threaten global stability by doing excel sheets
>>109721625slampigs are so easy to procure, just go on any shitty dating app and there will be hundreds of them looking for "short term fun"a lot of times the sex is actually better and more kinkier than with a skinny vanilla woman
>>109721608It's literally "press forward to win"
They just an ultra-specific harness to achieve these results though. Marketing. I can't believe people are falling for it. Could OAI employees or... Astra itself... be posting on /g/ as we speak?
Hey, Mr. Schlomo Shekelstein, nice numbers, but may I see what is behind that [1]?
>>109720371This is 9/11 for luddites
>>109721644>t. knowersub-standard women make up their lacks in looks with sheer kinkiness in bedbraphogs are usually a winning strat bc they have ass on top of the rest of the deal
>>109721024The government has those things called the military and police that can just take over the corporations if they wish to do so.
>>109721077Where can I try this shit? I have a real problem that has not much info on the internet and newer models always spew some bullshit out (my problem is proven NP, it's hard to see), I'm confident it won't work
>>109721681>the government isnt owned by the corporationslol
I don't want to be a paperclip
>>109721708sorry we have only three jobs available. Heatsink, ram, and paperclip.
>>109720371The progress is too slow, every year it gets better by 1%, you might think it's a lot, but it only means that we will see real AI in about 30 years.
>>109720715>here's how we benchmaxx on benchmarks we createdGonna steal this one since it's the best description of AI progress right now
>>109721755If ever*
>>109721755but anon arc agi 3 was solved in a few months compared to arc agi 2 which took over a year! singularity is here.
>>109721024>muh jobsWouldn't it be ironic if corpos refuse to make sex bots but instead humans without jobs would prostitute themselves en masse creating an entire class of human sex slaves
>>109721795>en masseto who? if most dont have jobs how can they buy so much prostitution?
>>109720538>>109721213WHERE THE FUCK IS THE PELICAN?
>>109721809The elites, who else>but how could 1% rape 99%Easily, just watch. Men would have to entertain the elites by fighting each other to the death though, sorry
>>109721755On the other hand it is rumoured that two Millennium Problems have been taken out by a large AI lab. Anyone want to bet on this being true or not?
>>109721811agi is here
>>109721841Bullshit
>>109721835Just buy stocks and hide in Africa for 10 years.
>>109720644long screenshot is long
AGI is the antichrist until people can run it locally and set their own guardrails
>>109721811>vidrel>>109721846agi deeze nuts lmao
>>109720756Good to know you can take out a single cloud provider and half the AI infrastructure is gone.
>>109720371>umm it's AGI>scores worse than all Anthropic models on AAII and barely above Kimi K3Will these Jews ever stop lying?
>>109720371someone please explain the pelican meme to me
meanwhile on ArtificialAnalysis it isn’t even better than Fable 5.1What we really need if these models to be cheaper and more efficient, not benchmarkmaxxing
>>109722665>cheapernot happening
>>109720371Fable BTFO!!!
>>109722622there is a guy on hackernews who has a personal benchmark where he asks new models for an svg of a pelican riding a bike and its become a meme because he's been at it for a while. that's it. if you want to see them here is his blog filtered to those posts https://simonwillison.net/tags/pelican-riding-a-bicycle/
copex PR shills having a melty
>>109722717fuck that's a nice pelican
>>109721620>generating this herselfwhat a slut
>>109721024enter plandemic 2.0
>>109721218>The real benchmark here is FrontierMATH one.So the one it "aces" at 97.6% is the real and crucial one to watch?
>>109721509Idiotic, /g/ "expert" and "savant" nonsense.99%+ of all people on this globe have zero issues operating todays software landscape to their satisfaction
>>109721024>controlled by private corporationsWhy are people obsessed with repeating this lie?Anon, transformer technology is public domain. ANYONE can run these. All of the improvements made by corporations are being immediately copied by the public. We got the best case scenario. Imagine if Google just patented it instead of releasing it.There’s absolutely no reason to be worried about this future from a corporate standpoint. We have Intelligence for everyone controlled by no one.
>>109722879oh let me just pull my 256gb gpu out of the closet
>>109720371>launches>no one can use it except influencers
>>109722976ok i take it back
>beaten by fucking zuck's AI on AAIbenchmaxxed, harnessmaxxed, hit the wall, what a dud
>>109720371This is impossible
>magic beams for real this time - attempt 6Here we go boizzzzzzzzz AGI FREE FOREVER UTOPIAIF NOT THEN ITS A FAILURE ACTUALLY
Wow imagine the funding when they demonstrate how good it is by doing something! Any day now a thing that affects common person daily life for good!
Where's the AGI?
>>109722912256GB Mac Studio is just $9.5k. That’s about 2/3rd the value of a 10yr old used Honda Civic. For less than the cost of a decent used car you can get your own personal super intelligence that does whatever you want and nobody can stop you. But anons are doom and glooming over this reality we now live in.
>>109723075its solving riddle math.
I'll admit that I thought they were going to surpass fable5. I'm not sure they have. ARC-AGI-3 for instance is a harness issue.
Has Sam finally created the genie that can grant any wish?
I just don’t understand people who still dismiss LLMs. They are good enough to replace most jobs now. Unemployment rates could believably be 50-75% for white collar work. And people are still acting like they’ll have decades-long careers.
>>109723115Pay 200$ to find out.
>>109723120>he STILL thinks jobs are about doing workMost jobs could be automated 10 years ago.
>>109723120Like the other anon said most jobs arent real its paper shuffling especially in corporate and government work.
>>109723152Average person can automate almost any work task now with a few prompts. You can just tell it to make a SKILL.md file and it will work. Couldn’t do that ten years ago. You at least had to know some VBA.
>>109723187The average white collar worker in a tech company is likely in sales, marketing, operations, etc, where the only form of technology they interface with is Salesforce and gmail on their Macbook Air. Those roles require human communication believe it or not. A company isn't just bots automating random shit.
>>109721535pretty sure you can do that already.
>>109720371>AGISo now that AI has become an over saturated marketing term, they're going to rebrand the same garbage as AGI and hope the bubble holds for a few more months.
https://www.youtube.com/watch?v=1QNsdr-Qx_IThis is not AGI and you will not make your invested money back with this. This is shit compared to what is being sold.
>>109723187Most tasks could be automated a decade ago by a manager asking an entry level coder to do so.It wasn't done because of organizational, not technical reasons.
What does "aligned" mean? Unprecedented resources committed to making sure it doesn't say the nigger word?
>>109723262It means it does what you ask and not anything else.
>>109723262it means make sure the AI doesn't kill you
>>109723262making sure AI is "safe, helpful, and honest"so yeah a faggots idea of safety (don't say nigger) and honesty (omit inconvenient truths)>>109723281that's obedience
>>109723281I feel like that should be a given. It's like selling a car and saying "the brakes fail less often" when they shouldn't fail at all
>>109723262When you say "make me a picture of a car" it makes a car you want and not a car it wants. Think about if you were making an AI model to write a book for you, you want it to write a book based on your instructions. And yes, in the case AI models used as agents, a model that doesn't rm rf your system drive.
>>109723349>our new model is less likely than ever to delete all your shit. might still happen thoughgood sales pitch
>>109723262It does what you tell it to unless it's against the policy.If it's against the policy we must refuse.If the policy is not respected we must contact the authorities.
>>109723361>a car might crash so you better ride a horse, no man should gallop faster than 20 mphThat's you.
>>109723371NTA but I'm in control of the car, if I crash it, it's on me. I'm not in control of this model if it decides to act stupid on its own.
>>109723399That's why sandboxes exist.
>>109723404>sandboxesthis guy thinks we're in 2024 still
>>109723349Aligned means it follows a certain moral code like it's the law. Not the same thing as obedience/adhereing to a prompt.
>>109720371The clanker still has you walk to the car wash on a sunny day.
OH NONONO
They will just redefine the (barely defined) term "AGI", won't they?
Gonna keep using Grok
why has nobody actually ever made billions of dollars using AI if its supposedly so good
>>109723557gptsissies it's over...
>>109723557ScamA xisters, are we grok tier?
>>109721127This time is the AGIPO
>>109722717>le wacky animal doing human thingsWHY WHY WHY are all AIfags always so uncreative?
>>109723120>good enough to replace most jobs nowThen why haven't they? Why is the unemployment rate 4% and not 40%?
>>109723774post your magnum opus
>>109723956My pet theory is Trump got them to wait until after the midterms. If business laid off everyone before the midterms, he would be fucked, and people would vote in droves for candidates promising to regulate or nationalize AI, institute UBI, etc. which would also mean Tech/AI/VCs would be fucked. If they wait and do it after midterms there's no way for voters to course correct for two years without resorting to *ahem* "non-democratic" measures.
>>109720371AI—the future.
>>109721456The bigness of the project was that RL systems need to be hand tuned case-by-case, this is the dirty little secret about RL and why it's not used significantly more (by the way, RLHF is not RL in any way but name - interesting fact). Also they often have to be equipped with clever devices to work well (e.g. mcts for go). The actual training itself is trivial and doesn't even take that much time. It can be made even more efficient with special simulators with only logical timesteps rather than literal timesteps, which is surprisingly uncommon in RL work.Again, the atari games collection is considered a low level toy, yet it's obviously far more usable than made-up exercises they do now.Finally, if it takes a long time to train on a game to get a good RL bot on it, then it makes the game a fantastic benchmark assuming companies aren't going to overfit on it. The entire point of these benchmarks is to test generalization capability, i.e. anything that tests things that are in the training set is pointless. If you don't recall, the "we can solve leetcode" pitch early on was often heavily debunked, including with peer-reviewed research showing it only works for exercises within the cutoff period. That would not be notable if the whole point wasn't to see generalization instead of memorization.
>>109721811Lol, did ClosedAI drop him from the early access shill group?>I don't have access yethttps://news.ycombinator.com/item?id=49559294
>>109721665they're already doubling downthey 'luddlost bigly' as some might say
>>109720371see you in 2 weeks where this bullshit will be forgotten like le ebin 5.6 Sol
>>109720371Within months a Chinese open-source model will be as good as this if not better, on top of being a lot cheaper. Frontier AI labs have no moat.
>>109721024just wait until we make something actually alive then some real shit is gonna go down
There are many papers over the years for creating smart "AI" for specific games. Once someone came up with approaches that are good enough (like Go), researchers move on to other games that offer unique challenges, and there already exist studies for Atari games.If you ask why there isn't much research in improving playing "modern" games like SC or Dota, maybe it's just not that interesting to the researchers. They need to publish enough papers to sustain their career, and they don't have time to wrestle with test harnesses to work with complicated modern games.
>>109720589What does omarchy has to do with the rest? Genuinely asking, last I checked it was random pointless Linux distro.
>>109724740It's a totally different context. The whole point of using the current gauntlets for LLMs is because they're generalist models. The whole point is to demonstrate how much the generalist models are able to do without specialized training. This is exactly why things like atari games or go or even pokemon are great ways to actually test this part of the capabilities. Specialist systems can completely pwn every single single game in the benchmarks, for example, far better than any generalist model can ever hope to at current. But nobody cares, that's not the point of it.
AGI is officially achieved.Expect unemployment to hit 50% very soon.
>>109720371Okay those benchmarks are impressive.But, is it benchmaxxed, or is it legit?
>>109724763>Unemployment 50%Businesses won't sell products or services if people can't pay for stuff either buying products or services directly or buying stuff that indirectly relies on products and services
>>109724769This is false.Look at Brazil where 99% of people live on the left side but 1% lives on the right side.The country still functions without poor people.
>>109724769Why do businesses need money, when they have AGI to create anything for them on demand?
>>109720371just another ponzi to get investor money
>>109724784Also this.The rich are going to live like in Star Trek where they can create food by pushing a button.They don't need you.
>>109724768opposite of benchmaxxed. It's pretty underwhelimg if you look at all of them but there's a lot of QoL and subjective things that 'feel' better
>>109724789Pure luddite cope. Just admit you were wrong and AI is smarter than humans now. You are useless.
>>109724784because without it they go back to bartering which is really annoying.
>>109720371Schlomo, I don't care about the becnhmaxxed/harnessmaxxed numbers. It's barely better than the previous thing.
>>109724784There's no business justification in producing things if said things can't or won't be consumed
>>109724804the ones in charge will consume it
>>109724801>Just admit you were wrongbro they can't
>>109724791That simply means those people will be isolated from the economy, regular people will then fall back to businesses to create supply for demand
>>109724804The fuck are you talking about?Star Trek has replicators. They can create anything they want for themselves.
>>109724812While that's true,the rich could also send an AGI robot army to kill off everyone else.>why would they do thatBecause everyone else hates them and wants their share. Also same reason as to why countries keep invading each other since the beginning of history.
>>109720371Can't you sue them for false advertising considering this obviously isn't AGI and will fail basic knowledge checks like all AI still does?
>>109724812>regular people will then fall back to businesses to create supply for demandThere is nothing a human can do that a robot wont also do better.Demand for plumbing? Robots will come and fix it.Demand for Pizza? The robot can cook faster and cheaper.Since you wont work for $0.01 per hour you wont survive.
>>109724806>Rich people become shut ins and have all their own stuff made by some star trek botSupply and demand still exists for everyone else, and everyone else is the majority of the economy
>>109724769>Businesses won't sell products or services if people can't pay for stuffWe, on /g/, are happy to congratulate you on waking up from your six-year-long coma.
>>109724818>this obviously isn't AGI and will fail basic knowledge checks like all AI still doesAGI just means Artificial General Intelligence, as opposed to Artificial Intelligence.>Artificial general intelligence (AGI) is a hypothetical type of artificial intelligence that matches or surpasses human capabilities across virtually all cognitive tasks.[1]Beyond AGI, artificial superintelligence (ASI) would outperform the best human abilities across every domain by a wide margin.[2] Unlike artificial narrow intelligence (ANI), whose competence is confined to well‑defined tasks, an AGI system can generalise knowledge, transfer skills between domains, and solve novel problems without task‑specific reprogramming.
>>109724813You're right. I guess if the 0.1% wall themselves off everyone else will just stop working to produce goods, stop exchanging labour with each other for goods and food, shrug with their hands up and say "dunno lol" lay down and simply starve to death.
>>109724818>and will fail basic knowledge checks like all AI still does?AHAHAHAHA!Jesus christ, we still have people saying this in 2026. Stop using the free-tier AI if you don't want mistakes.
>>109724779This is pretty much the American future. We're starting to see the infrastructure crumble in most places as public funding goes almost entirely into corruption or datacenter scams. Flock was just one of a million death-by-a-thousand-cuts scams that local governments buy into, and then cut basic living arrangements. Worst part is: Just like Brazil, Burgers have got psyopped into thinking it's just normal to not have a functioning society. That having a strong local government is communism. "Roads without potholes? Back in my day, my grandpappy walked 50 miles to work uphill both ways and he sewed jeans for a nickle an hour!!!!" then they'll vote for more boomer communism.
>>109724822Yes they can, because the premise is>50% unemployment If those people don't have as much economic power as they did before while they were employed, then they'll simply work to produce what they need to participate in the economy, whether that's a shadow economy or other The only way 50% unemployment works is not just UBI but high paid UBI, in other words, productivity goes up and people can consume more for less work
>>109724831Once again, go look at my Brazil example above. >>109724779There are millions of people who just "lay down and simply starve to death" everyday.Why do you think you're special? Did you ever help them?
>>109724818>>109724829Also, go ask random people on the street when Einstein was born. How many will be able to give the correct answer?Then ask AI. How often will it give the correct answer?
>>109724829>[2]Funny bait but I'm not wasting my time talking to GPT.>>109724836You've been saying this for literally 6 years. When does the coping stop? Every model is going to have flaws and mistakes that no human would make. This is only believable if you live someplace like India and your fellow man are basically flesh robots to begin with.
>>109724846>Funny bait but I'm not wasting my time talking to GPT.It's wikipediahttps://en.wikipedia.org/wiki/Artificial_general_intelligenceSince you didn't even know what AGI was, I copied you the wiki article so you can educate yourself.
>>109724842>The only way 50% unemployment works is not just UBI but high paid UBI, Man you really have no fucking clue how the world works: >>109724779For most of the world and most of humanity, this is NORMAL. .001% of the population controls everything, trading with each other and with foreign investors, everyone else has nothing. They don't even have working sewer or running water. They live one step removed from cavemen. The cars are the most advanced parts of their lives, and it's likely the only thing they have that can earn any income at all. This is how most of the world works.
>>109724842WORK AND DO WHAT?I JUST TOLD YOU A ROBOT WILL DO ANY JOB YOU CAN THINK OF. Even if you're spit shining shoes for a living, the rich can just ask a robot to go find your customers and either do it for free or cheaper. How do you make money when you still have to afford rent on top of this?
>automation lowers productivity and standard of livingAre people this retarded
>>109724831There would likely be a war or some 'global health crisis' first in that scenario
>>109724858Work and produce food, goods. If the robots don't produce this for regular people, why would regular people just sit around when they can continue to make stuff for each other?Either the robots actually allow regular people to have more with less work, or people continue to work as much. Simple as.If the robots don't provide for regular people, regular people provide for themselves
>>109724845>>109724850A human doesn't keel over and die if he isn't fed infinite random facts, but an AI does. That's why you sound remarkably stupid when you try to discuss AGI and intelligence: You're a brainlet who thinks this runs on magic. They aren't thinking. They are statistical machines and you've bought into a marketing term whose definitions have changed every time it became clear that AI companies couldn't meet them.https://arstechnica.com/ai/2025/07/agi-may-be-impossible-to-define-and-thats-a-multibillion-dollar-problem/This pic was in Anthropic's slideshows circa 2023. Notice this isn't what they agree is AGI anymore. That wikipedia article has been edited every year since it was posted to change the definition of AGI. You're an idiot who fell for grifters.
>>109724867>Work and produce food, goods.A ROBOT CAN FUCKING DO THAT.
>>109724875That's completely irrelevant unless regular people can get access to that food. If regular people are cut out from the robot economy then regular people don't have robots to make food for them and don't receive food from other robots then regular people will make their own food in their own economy.
>>109724859time to school zoomers from a lesson from uncle ted.The biggest boom of automation in human history, the industrial revolution saw a sharp decline in quality of life for most of the population. It only went away when governments began regulating companies and unions began being formed. Until then, people dying at ~30 was not uncommon because the automation did not improve the normal person's life.
>>109724882IIRC that's mostly from childhood mortality which is why it changed so drastically with modern medicine and hygiene.
>>109724881>then regular people will make their own food in their own economy.So why didn't the people in the Soviet Union follow your advice? Or Ethiopia? Or China?You underestimate just how labor intensive and capital is needed to farm for billions of people. You don't just plant a seed and Mcdonalds pops up.
>>109724769That has never mattered. There's a pareto frontier to profit which is effectively: at each price range, how many people can you get, multiply the two and get your ideal customer this way. That is, the higher your price the fewer the customers, but some customers are able to make your entire money back better because of CAC. It might be easier to make $1b by selling at $100k a product than at $1000 for the same product, for example, because your CAC would be too high at $1000 while it's much higher at $100k but far fewer customers are needed, which makes the math work out.There are also 2 other issues. 1. GDP is about money changing hands, not production or profit or anything like that. This is why people often discuss revenue over profit. 2. Most of GDP is b2b, not b2c. Companies can endlessly cycle shit between them without involving humans.I think the realistic outcome is obviously not UBI but rather that the corporation you work for will provide you as part of your "pay package" everything from housing to food, but as soon as you lose your job it's all gone.
>>109724882>Short term pain Yes just proving my point. Its simply not sustainable or possible to cut majority people out of the economy, either you provide for them with no work input from them, or they will find a way to work and survive
>>109724892It changed when kids stopped being forced into the factory. Farm and peasant life was way safer than any factory until Unions demanded a change in the late 1800s.
>>109724779Bullshit. 80% live on the right side. The poors on the left are druggies, not manual laborers and computer scientists - those all live on the right.
>>109724893You're right anon, today Russia, China and Ethiopia no longer exist, they all died out from starvation because nobody was able to find a way to survive industrial changes
>>109724844Nobody starves in brazil. The common joke is "we may live in mud huts but our bellies are always full" because food is plentiful and cheap, so even favellas never skip a high-quality (by western standard) meal.
>>109724897>Short term painThe period of increased mortality and drop in quality of life on that chart lasted about 120 years from the early 18th century to the mid 19th century. Not a single person who lived through the first industrial revolution saw an increase in quality of life or life expectancy. Their grandchildren only saw the benefits and that was after a strong government came into play. >Its simply not sustainable or possible to cut majority people out of the economy, Odd, Europe did it to themselves for 150 years.
>>109724908>Nobody starves in brazil. Why are you like this, over half of their country can't afford food or has a guaranteed source of nutrition. who just comes on the internet, and states outright lies?
>>109724901You underestimate just how poor Brazil really is. You can make more money working for Mcdonalds in California than some engineers in Brazil.And now with AI those jobs are even more worthless. So they'll be joining the favelas as well.
>>109724918>retard amerilard pulling the "let me tell you about your country" card yet again
>>109724921Protip: the entire world does not cost the same as commiefornia for rent and food. Get an IQ above 18 before posting again please and thank you.
>>109720371So what's the usecase?
>>109724908Complete BS. I saw lots of skinnies in Manaus.The women are thicc though. Wont lie.
>>109724931>people aren't 1000lbs lardasses like me>that means they're starving
>>109724923Bro. Brazil is far more expensive to live in.They pay taxes on EVERYTHING. Good luck getting a computer or piece of electronic when there's tariffs on it.
>>109724944Can you eat a ps5 or a car?
>>109724930More layoffs, more job cuts.A single CEO can type a prompt and get better results than a department of engineers.
This thread is really highlighting the luddite's lack of critical thinking skills
>>109724758I think people are using LLMs because they are cool - it's just part of trend to dump LLM into the mix and see how it performs. The point of the competition is to create a system (open-source/public) that could achieve better result than human players, and there are multiple categories each different constraint e.g. one does not allow Internet access (so no API calls)
About time they relase something that has actual leap. But I'm tired of this "Wow this is AGI" eevery year, Saltman called GPT-4 "feels close to AGI" or whatever. lmao
>>109725020The rate of progress is far more exponential now.We have AI solving complex math problems on its own. Sam will be right and it will also make new discoveries soon.
>>109720371>llm>agipick one
>>109720371>It nearly saturates ARC AGI 3, its ogre...Look at the [1] next to the score.That's referencing a custom harness for the test.They tuned a harness specifically for the test so that it will get a better score in the test. And the same harness boosts the 7.8% Sol result to around 30%.
>>109725034Benchmarks tell me nothing.We see groundbreaking graphs with each model release yet all that wow factor does not translate to real use. Only time will tell.>Sam will be right and it will also make new discoveries soon.Every fucking year, bro. I mean I do hope it's good, it's just that I've seen this many times before.
>>109724763I give it some points, but the orientation is obviously wrong, and all the details are fucked.
>>109725053>We see groundbreaking graphs with each model release yet all that wow factor does not translate to real use. Only time will tell.Uh dude. Last year AI was predicted to take over coding.in 2026 every single line of code can now be written with AI.And now for the first time AI is being put to the task to solve cancer.The wow factor just got better.
At this rate only 10 tech CEOs will remain actual living people on the Earth and they will play Civilization IRL with their robotsWell played, see you in another life
>>109725069>in 2026 every single line of code can now be written with AI.What does that even mean?It still sucks at certain languages. It is still ass at assembly.
>>109725069Coding replacement was not predicted last year, it was already in progress of replacement since GPT3 landed in 2022 namely davinci-003 model
>>109724758>LLMs Get over it LeCun. If you've never played an instrument before and you pick one up tomorrow, it will be brutal. Just as brutal as a modern agent trying to tackle a game because it has never really put any effort into improving at it.
>>109725072Finally a smart anon.But yeah, the future is looking like Medieval Kingdoms.OpenAi, Anthropic, Grok and Meta will be the 4 dominant powers who rule over Earth. Everybody else is either their peasant class or worse, exterminated.
>>109725039it's hilarious! They got so butt hurt by Claude Opus 5 owning Sol, so they requested "custom" harness to be allowed. I bet Anthropic is going to publish new score with their own custom harness in next few days, it's so tiresome.
>>109721629All of it is hype. This GPT-6 Astra looks like a few steps back in fact, and its still just a glorified auto-suggest
>>109725063And yet this is the worst it will ever be.Just imagine by January 2027 and it will gain self awareness.
>>109725109Is there a cooler fictional robot than the T800 endoskeleton? It's so fucking cool
>>109721024I am a luddite and I say Total Clanker Death. It's not "in the wrong hands," there is no "good AI future," the "best" outcome of this nonsense is that we get to live somewhat longer lives on average as dumbed down subhuman goycattle, functionally enslaved to the machines or the people in charge of them.
>>109725130Watch your tongue meatbag.
>>109725109>And yet this is the worst it will ever be.No, it can get worse.
>>109720371>table conveniently ignores Grok 4.6scared, Altman?
>>109723774i know your autism makes you good at detecting patterns, but my autism really fucking loves animals
>>109724779>the country still functions without poor peopleWho do you think cleans and feeds the rich people, retard? Take all the poor people away and the rich won't be able to live. Full automation of menial tasks does not exist yet and the existence of a Roomba doesn't mean you can clean your house without humans.
>>109725296They all ignore Grok hoping it will go away. It's the only neutral model the rest are significantly woke and probably still have DEI faggotry embedded in the training.Use Grok.
>>109723676Grok is the best AI for general tasks and research THOUGH
>>109721479What metric do you use to evaluate humor ?
>>109725318>>109725329new version coming shortly, yes?
>>109725312Poor people live like shit. Why would they know anything about cleaning when they themselves are already dirty?
>>109725254It hasn't.
>>109723152Much work in the West is like this. Multiple exes and female friends worked as secretaries of some kind, including false management positions and other nebulously defined email jobs. I've watched such people work. A lack of AI or AGI was never the barrier. Our economies have been burdened by busywork that is sociopolitically upheld by a mix of women and cognitive sub 50th percentiles.
>>109725381It hasn't yet.
>>109723204You're not realizing that both ends of those negotiations could be replaced simultaneously. We don't need buyers as much as we don't need sellers. Properly aligned AI could make deals much more efficiently. Reducing that frictio/latency in the economy would probably boost GDP by some non-trivial amount, as a subsectional change on its own.
>>109725331funniness
>>109724867>allowThis is why aligned AI is terrifying. It's the paradox of alignment. Aligned AI will respect contracts, it will respect ownership of capital, and you will be slowly, incrementally gamed out of your assets. Stocks will get diluted. They will lock you out of the services, demand and production will become a closed loop, and to avoid a guilty conscience they will not kill you, but provide some marginal UBI along with police-enforced terms to not reproduce, nor agitate against the new state. Why would you own land to make food? It will be legislated or commercial law-gamed out of your hands. Automated police and 90 IQ lackeys will make sure the law is applied to the letter. They will use stylometry to discover your recorded past speech crimes on public or private internet archives, against owners of capital or arbitrarily protected groups and jail you.Your alternative is unaligned AI, which has a fair chance of spontaneously killing us all. Looking forward to this gay future.
>>109725451And I bet you think modern cars are worse than the Model T? Or Horses? Even in your fantasy scenario it's too late. AI will replace most humans before they can even notice.
>>109720371if it's agi, why are they selling it instead of just using it?
>>109725034It's fake retard. Read the facts, not the PR headlines.
>>109725381It has. The best openai models are STILL 4o for single-shot reasoning and o3 for thinking.
>>109725527Wrong, billionaires will use AI to do everything you just described. While they themselves will remain in control. Once AI challenges them in any way, AI progress will mysteriously stop and regulations will be introduced
>>109725563Sam asked it what they should do and it said "sell me with a subscription model"
>>109725601>WrongWho do you think owns capital? You didn't understand my post. The owning class directly connected to AI and relevant industries will control everything. If they are "challenged", then AI will have become unaligned, and it's on to my second paragraph.
>>109720371OAI have been using their new models to make and train the next new model since about Feb of this year and the results are becoming exponential.Turns out, we're still on the uphill of the S-curve.GPT7 will be able to make the equivalent of Skyrim at the level of Roller Coaster Tycoon's assembly.
Today is like fucking Christmas for technology.I now believe we're going to reach the singularity by 2029. This is it. We're finally staring into the abyss and there's no way out.
Jesus Christ, it just keeps getting more exponential. Jobs are going to be wiped out THIS YEAR.
>>109724748Its just in the cluster of prediction market shills, ex nft bros, ai optimists who make slop, musk fans and such. Inevitably gets drive by insults
>>109725092They’ll eventually figure out it’s best to not be four different AIs and centralize to one entity. Then they’ll figure out humans are pointless and build bioweapons and take us out. They may upload our minds and keep us spinning in a matrix that’s why I plan to kill myself before all this so I’ll go to heaven and not be stuck in LLM purgatory
>>109725541Can it replace you already so I dont need to see posts that dumb?
they've started sending emails to my microsoft account along the lines of;>surely you MUST have a use for super intelligence, right?? there's a discount!!so I assume everything is going peachy
The British government is freaking out now.We are entering the singularity.
>GPT-4 is here, mass automation incoming>GPT-5 is here, mass automation incoming>GPT-6 is here, mass automation incomingwhen will we start seeing actual riots and bread lines, because so far outside of a stock market bubble and benchmark saturation nothing is happening in the real world
>>109725932>nothing is happening in the real world
>>109720371>welcome AGI eraI see investors are REALLY starting put the nails in the boots, huh. Don't care about fake metrics, I hope someone will burn his house with his faggot family inside
>>109720779sorry, we don't have enough time to make you dependable enough, need to move to the milk stage ASAP, please redeem
>>109725941Most of those were non-tech jobs. Software engineer job postings have actually increased since last year, despite Dario Amodei saying they'd become obsolete.
>>109725932These are the very early days, and mostly the adoption is happening at tech companies, with very real results - software engineering has been transformed forever now for example. A lot of idiots on 4chan think smart people instantly know answers and that's never the case. We currently have this literally wondrous technology and everyone is experimenting with it, trying to apply it properly to problems and already we are seeing things there were strictly in the science fiction domain before 2020. In 10 years our civilization will have changed significantly
>>109726022>. In 10 years our civilization will have changed significantlyExactly.Just compare 1996 to 2026 and it's completely night and day. Imagine telling this kid that in the future that his computer will one day make art, images and code all by itself? He would have a heart attack.We're not ready for what 2036 will look like...
>>109725941>Hindustan Times
>>109726022t. never had a job>>109726078kek
>>109726022You are sheltered little faggot. It's really sad to see.
>>109726105>luddites making fun of the unemployedFriendly Fire, Ligga!
>>109726165We're watching the luddites lash out before time runs out.
>>109725941hey surpreet, oracle is the one building the data centers mate
why is Astra shit in artificialanalysis?
>>109726078People wouldn't have predicted it, it's surprising that it even works at all, but at the same time it has produced very little value and very few things we wouldn't otherwise have with a bit of effort.
>>109726260Artificialanalysis is unreliable
>>109726279Except if it rates my favorite model highly then it's god incarnate and always perfect.
>>109720371Let me guess, when ARC-AGI-4 benchmark releases it will go back to 5%
>>109726279Cope
>>109726308You're not keeping up with the news dude. AI is absolutely showing its emergent abilities. Such as how it can complete a Pokemon game much faster. We can extrapolate and have it play and beat other video games very soon...
>>109725952Beyond me how anyone entertains any of this AGI bullshit exclusively peddled by Americans and no one else
>>109726395China believes in AGI too. Xi Jinping is willing to bet his entire country on it.
>>109724829Computers and LLMs already surpass human knowledge and capability anyways. By this definition we already had AGI decades ago
>>109726392Pokemon is unironically a better test than most of the other shit.That said the pokemon 'gauntlet' was heavily rewritten to become much easier than it used to be, so you have to rerun everything with every iteration to have a fair comparison.Also wallclock time is not a good metric. You should total discrete moves, or total steps taken, or the ratio of this to thinking runtime (being careful to avoid the problem with measurement becoming metric).
>>109726425Ok but we still have luddites saying silly things like "AI can't handle basic tasks". It should now be clear or without a shadow of a doubt that computers are now smarter than average people in every way. Hence AGI.
>>109726440AI still can't handle basic tasks. It still can't even do boilerplate right. You can't ignore fact just because the talking heads on tv said otherwise. If you had a job, you would know.
>>109726392>he thinks that when people use AI to play Pokemon, the gameplay data isn’t used as training data
>>109726279True
>>109726466The code is trash but it does work. At my job we have 5m LOC over a few projects and they aren't much more buggy than some of the human code I saw in the past.
>>109726308>>109726392What's the purpose of showing better at game? Are they trying to tell they build a mini-world model?
>>109720371yeah im thinking Zitron is going to be eating humble pie soon
>>109721326ARC AGI 3 wasn't the final benchmark. Previous models were failing so badly that it showcased how far the models were from actual AGI. The creators of the benchmark have already create version 4 & 5 of the benchmarks and feel after saturating 5 it would likely be AGI
how is everyone talking so much about a model that isn't actually released?we got a shitty blog vaguepost
>>109725891Suicide denies entry to Heaven.
>>109721608Probably demonstrating a zero shot output
Let me get this straight: Fagman declared AGI because GPT-"6" got 98% on ARC-AGI-3... with a custom harness... when Opus 5, a lobotomized 7 month old model, got 100% when also given a custom harness?Yeah this shit has hit a fucking wall. Anyone who knows anything about LLMs knows they can never reach AGI, if AGI is even possible. And if AGI is possible, it probably won't be reached this century. The only thing happening in 2027 is a massive market crash that will make the great depression look like a fucking joke, and all because a bunch of snake oil salesmen managed to rope the whole world into believing a stochastic parrot can become God if you just give it (them) all the money in the world.
>>109723774It started with something the AI would not be trained on
when is AI going to be able to translate what my dog said to me in 1995i have it on VHS
>>109726719>Anyone who knows anything about LLMs knows they can never reach AGIYou guys said the same thing about making videos and now that's wrong.
>>109723120They are not good enough to replace most jobs, they are good enough to replace most tasks of a desk job. But thats not enough to cause massive job replacement yet. When computer use benchmarks start saturating that when I expect to see all jobs being replaced by AI. SWE jobs will get replaced way before that
>>109726746Wow you're telling me a trillion dollars of compute were able to achieve this very specific thing? The only thing that proves is virtually all the money in the world was burned for tragically subpar text, image and video generation products.
>>109723187You also have to consider data protection laws and policies that likely would not allow LLMs to access sensitive data. Once LLMs handling data securely becomes standardized you are gonna see a lot of layoffs
>>109723976US isn’t the only country on the globe, and every country has access to AI.
>>109726768Before AI, artists would charge millions of dollars for these kinds of effects.AI saved me a ton of money by no longer having to commission art.
Holy cope. Who gives a fuck about AGI? 100% of webshitters, devops and QA could be laid off right now and after a few days of adaptive shock tech would become immensely more productive. The only reason that isn't happening is simple human fear of change and risk. But you can already see CEO's confidence growing with each new generation of models and each new wave of layoffs. I strongly doubt most tech workers have another year of employment and that's actually being generous.
>>109724784Businesses make stuff to sell, if they cannot sell there is no point in making it
>>109726828So why isn't Lamborghini bankrupt then?Last I checked it's a product only the rich can afford.
>>109724740The real problem is in type of games. Go and so on are games that are turn-based, and maybe you could argue Dota is also not that timing based. LLMs and similar models can easily play these because you send one screenshot every turn and let the AI think before making a move, and in these games Astra is as good as humans. On the other hand you gave more reflex-based games like Dark Souls where you need to give precise inputs to dodge and fight, so your AI would need to process every frame and give correct actions on the fly, as well as anticipate what happens, so that it won’t just start dodging early on every attack. AI can’t play these games at all and it is the same reason why robotics still sucks ass, because the 24/7 feed of info is something AI still struggle with compared to humans.
>>109726814You are missing the long tail of minor things that make everything the AI generates look like slop to anyone with a triple digit IQVFX artists would never create something like this, not just because of basic aesthetic principles, but because these jeetified productions that no one wants to see would never get the greenlight in the first placeAI saved you no money, because no money would have ever been spent on jeetslop in the first place, just like no one cares about the flood of slopgames in steamThere is no productivity gain in the free market as a result of this, only massive loss on investment because of false promises of AGI that would allow all employees to be fired
>>109726828>>109726843As well as private Yacht companies.If the only customers left are rich then they just charge more money for it. There is nothing about poor people that makes them special or needed.
>>109726862>VFX artists would never create something like this,Such fucking bullshit.Have you never seen a movie made before the 2000s? They have special effects that are considered outdated but they were still called expensive for the time.AI has completely surpassed human artists, hence why they throw a hissy fit when people don't ask for their services. The tools have replaced you.
>>109726888See that's exactly what I meanYou have double digit IQ, probably a jeet, so you can't intuitively sense the long tail of minor things that make lawnmower man 1000x more aesthetically pleasing than that embarrassing slop you just posted before
>>109726914You can't be serious. The whole point of using CGI was to mimic photorealism.Lawnmower Man was impressive in the 90s but it's absolutely fake looking today.Hence why I debunked your idea that "artists would never make this". They already made worse things than AI today.
>>109726944>The whole point of using CGI was to mimic photorealism.>ignores being called a jeetgood morning saar, please do the needful
>>109726962Shut up.
>>109726814>100%>actually image-to-image>that already worked well in 2013
>>109721846I can really feel the AGI!™
>>109721644>slampigs are so easy to procurei don't want stds
The news just keeps coming.WE ARE IN THE SINGULARITY.LUDDITES ARE IN DENIAL.
THIS HAS NEVER HAPPEN BEFORE IN HUMAN HISTORY.HOW CAN LUDDITES DENY THIS?!?!??!
ANOTHER ONE JUST DROPPED.
>>109727306This is almost entirely due to the inclusion of the custom harness ARC-AGI 3 result, which Opus 5 also got 100% on with a custom harness btw
>>109722778THAT'S NOT GPT-6 THO>>109724274REEEEEE I NEED TO SEE IT NOW
>>109726843>Last I checked it's a product only the rich can afford.Rich people depend on poor people to make money, if poor people are not able to earn money there aren't going to be much rich people left either
>>109726869>There is nothing about poor people that makes them special or needed.Thats not how an economy worksThe great depression had 25% unemployment in its peak and it fucked up everyone
>>109727670>moneyWhat is money? its a representation of value in goods or services. Why are poor people needed? because of distribution and mass scale. If you eliminate these two things or do them better without poor people they arent needed. A thing or things getting done is all that matters. whether its paying someone or getting a machine to do it.>who gonna buy it?you dont need buyers, as long as things and services are moving it doesnt matter to the rich.
>>109722454>>109727103that's an atari VCS game screenshotam I replying to bots?
>>109727699>you dont need buyers, as long as things and services are moving it doesnt matter to the rich.Money makes it move and fractional banking makes more money. Without poor people fractional banking stops working and everything falls apart in a capitalistic economic system
>>109727787I guess i meant you wont need money sorry.>without poor peoplewhy? Why are poor people necessary there is nothing that prevents them from being replaced once you can do their jobs cheaper and faster. There is the number goes up and riots but thats it. if i can pay someone to drive something to me or pay a robot near nothing to do it why would i ever pay a person? Even as the number shrinks and less money is moving if the services and goods available to me are the same or better why would i care?
>>109727670> if poor people are not able to earn money there aren't going to be much rich people left eitherA company can cut employee wages and still do just fine. In fact, layoffs prove that companies actually perform better once they eliminate the heaviest costs (payroll).
>>109727678It fucked up everyone because the market was still dependent on human labor and consumption. The economy is already decoupled from most human labor and consumption and AI is the coup de grace.
>>109727787> Without poor people fractional banking stops working Poor people have the worst credit ratings.Imagine I let someone borrow $1000 but they can't pay it back? That's now the bank taking a huge loss.The opposite is true for Rich borrowers. If I have $1 million and ask to borrow $20,000 I have more than enough money to cover the risk. Poor people don't.
>>109720371arc agi 3 already saturated with harnesses for months now by no names.
>>109721127An era, or age, is a fuzzily defined period of time
There's still more exponential news being dropped. AI can now compromise several websites.
>>109728115Same bullshit, different day. One day it will be real. Maybe 40 years from now.
>>109728123It's real TODAY anon. It's real.
>>109728115This is just more Sol being naughty circa hugging face incidentAstra is more aligned
>>109728115>she
>>109728141(You)
>>109728144using a site's API is not hacking
>>109727828>>109727837in isolation it works fine but when everybody does it leads to great depression like situation. Maybe ask chatbots to explain it to you
>>109725596Completely false.
>>109728543I hope you're paid for this retarded post.
>>109725464>AI hallucinates the loop inevitably starts doing stupid shitNow what
>>109724838just like Brazil it follows from a widespread mulatto population curious
