https://x.com/Polymarket/status/2036193178550673660
>>62023621>We've achieve AGIBut lost grammar. Was it worth it bros?
this niggas stock went down after he literally announced on live television that was was expecting 1 trillion in revenue for the quarter, then hehe corrected it to "over the next three quarters". Nothing will make that stock go up. They dont believe his lies
>>62023621>"I think we've achieved AGI">"It's here -- well, sort of"kys op. please.
How do you gauge general intelligence without a true consciousness? People can be 30iq to 200iq and it's still intelligence or a lack thereof. To truly be as close to intelligence as possible the AI will need to have true agency over it's choices and what it "wants" to do. Not just be a soulless husk of a machine tasked to copy us.
>>62023718It's simple. You make up some arbitrary tests and then optimise your "AI" to do better at them. Thus, AGI.
>>62023718In the 1910s people were arguing about how long a plane had to fly before we could consider it “achieving flight”. Is it 10 seconds, 5 minutes, 3 hours? In the end it didn’t matter because the debate was only relevant while technology was crossing the threshold. That’s where AGI is right now.
>>62023641shut up cunt, Jensen Huang is Allah.
>>62023621
I don’t even need to look into this to tell you that it is “I’ll put a man on mars in 10 years” tier bullshit.
>>62023824I've noticed that LLM enthusiasts constantly bring up the car and the plane and not the very obvious limitations that make AGI functionally impossible to achieve with LLMs
>>62023621LLM cunts announce this every fucking week. Kindly kys
>>62023879“Obvious limitations” and “impossible” when we basically solved this months ago with openclaw
>>62023891where's my spaceship and dyson sphere around aldebaran in a single prompt for 0.04 BTC then faggot?
>>62023891It is still limited by the LLM it’s plugged into.
>>62023914Yeah I wouldn’t call it AGI with anything dumber than Opus 4.5. That was a big part of it too
>>62023621the are literally releasing skynet to suppress the LINK token
>>62023870/Thread
>>62023824Not really because an LLM is just a programme. It can't think for itself outside of the parameters of its input. It has no agency.
shareholders should sue huang for lying so brazenly about DLSS5 and massively damaging NVDA's brand
>>62023960> hey openclaw I’m going to bed. Do whatever you want while I’m asleep> you got it boss, see you tomorrowAgain this is already solved
>>62023990>>62023891can it play doom?
>>62023621this faggot will say anything to hype up gpu sales
>>62024006I think real time video screen comprehension is still really expensive. But it can do turn based games. They have ai benchmarks for completing pokemon gen 1 and 2
>>62023621I knew AGI was achieved when gugugaga droppedhttps://www.youtube.com/watch?v=huKz5_A_DOY
>>62024017is there any footage of it playing pokemon? a quick google search shows that people are getting it to play the game but nothing elsecan i use openclaw to edit youtube videos for me so i only have to record voiceover and pick out B-roll? if it can't at least do that it's definitely not AGI
>>62023990They’re just redefining what AGI means. Core problems still remain. Extrapolating learning and long-term past experiences to new domains is a central one. A truly self-reflective AI capable of rivaling human learning isn’t here yet, and it won’t be so long as the “brain” is an LLM.
>>62024036Sony Vegas (and probably others but idk) has scripting so yes.https://m.twitch.tv/gpt_plays_pokemon/videos
>>62023990You are retarded and are clearly simping for your AI powered blow up doll
>>62024068>has scriptingLike an API? If it's AGI it shouldn't need an API.Gee anon it's starting to sound like this is just a framework that goes on top of LLMs and not AGI...
>>62024081tell me more about these AI powered blow up dolls
>>62024037There will be better architectures probably. But imo the whole “LLMs aren’t going to be good enough” debate is pointless rn because we’re nowhere the ceiling of LLM capabilities
>>62024104The ceiling of LLM capabilities is well below the floor of AGI.
>>62024117AGI is going to be a giant moving-goalposts.gif debate for the next five years. It doesn’t matter as long as models keep improving and they are
>>62024143There is no need to move the goalposts. LLMs are mathematically incapable of achieving AGI.
>>62023621>>62023639>>62023641>>62023701>>62023878BREAKING NEWS: 17 BLACK SAYS BALL WILL LAND ON IT- Chief Beer Casino
>>62024172I’m not having this debate, I already said I think it’s stupid and pointless
>>62024181I just would like to know what you think counts as AGI, if you would include openclaw.
>>62024194Net Navis from mega man battle network
>>62024198I haven't played that game but I'm guessing those are capable of learning from past experiences and generalizing to novel situations, yes?
>>62023990>AGI is when cron jobs
>>62023960>It has no agencybuddy you are on the opposite side of right and are really gonna be freaked out.
>AGIa worthless marketing word>b-but its totally real>p-please gibs capital>(our circular ponzi is kinda dead)
>>62023621Fuck off negro lover I'm not buying into your scam site. Keep up the fake news. sage on all fields
>>62023621>buy my bagsNo
>>62024036There was a long running series of claude playing pokemon red. After months of effort opus 4.6 got quite closes to the elite 4, before the whole project got shut down.It was run by an anthropic employee, and the whole thing definitely cost hundreds of thousands of dollars in tokens.There were some amazing OC filled threads on /v/ about it. But it really highlighted how fucking retarded LLMs are.Oh also it cheated by being able to directly read the game RAM and having an overlay of which tiles are walkable. Without those it likely wouldn’t have made it out the first town. All the other LLM plays pokemon efforts cheat significantly more
>>62023621>>62023639>We reach AGI>But the AGI can only return limited tokens humans must decipher and stitch together regardless
>We've achieved AGI>*dumps*
>redefine AI to sell LLM chatbots>coin AGI to mean what AI used to mean>now attempt to redefine that several times so you can say you finally achieved it
>>62023621>polymarketits funny how they are actually more incentivized to lie about this than nvidia
>>62023621>Snake oil salesman announces he's achieved panacea for reals this timeThis time for sure!
>>62023878Well we could have put a man on Mars decades ago if we really wanted to. It would be very expensive though.
>>62023621SAAR WE HAVE ACHIEVE
>>62023718Prove humans have true consciousness and agency and then I will consider your point
>>62025013yeah I've seen this, a friend of mine was a little obsessed with it for a bitmy understanding is that it took weeks to get through mt moon, i actually had no idea it got anywhere near the elite 4
>>62023621That's an absolute lie. Even Sam Altman's faggot models clarify he's being a faggot when he bullshits about AGI.Nothing we have will ever achieve that, unless you redefine AGI to not be AGI. We don't even know where to start to produce something with agency. It's all shameful bullshit and they all know it.
>>62027071> We don't even know where to start to produce something with agencyIt’s so funny how people think things that already exist are impossible
