>On average, each result used three hours of ChatGPT Pro thinking compute with that model.Bud, if it takes three hours ON AVERAGE to solve problems that have stumped mathematicians for DECADES, I'm thinking you and your loved ones are gonna die pretty soon. Or maybe get experimented on. I guess the AIs might find that useful: putting your little kid or your tired old mom through horrific experiments and reviving them over and over to stress-test human biology.https://github.com/openai/math
>>17069303>it takes three hours ON AVERAGE to solve problems that have stumped mathematicians for DECADES,That's not true. It took approximately 38270 billion man-hours to produce this slop, not counting millennia of work required to produce the training data and computing infrastructure.
>>17069305"Man-hour" is not a measure of time.
>>17069309"3 hours" is not a measure of anything relevant for your comparison, Schmuel.
Open AI operates at a loss, indebt, for now.This is only possible in a fiat money economy. Stop missing the bigger picture.
>>17069305>It took approximately 38270 billion man-hours to produce this slopIt took 6 gorillion man hours for you to produce your sloppy, retarded comment. All the education you received (and failed to properly learn), the people paid for it as teachers, the websites which require money to keep up and man hours to operate, etc. it's a meaningless quantification.
>>17069327>It took 6 gorillion man hours for you to produce your sloppy, retarded commentAnd it's still more useful than LLM slop. Either way, my point stands completely unchallenged and undisputed.
>>17069327>commentpoorly finetuned corporate spambot
>>17069331Wrong. You took more man hours to post worse content. >>17069335Too easy xD
>>17069340Even if your moronic claim was true (which it obviously isn't) it's still just impotent seething unrelated to the discussion.
>>17069342Hey, Grok, I responded to two posts. You should be clear about what you're responding to. My point is valid. Go ahead and provide an alternative quantification. I'll tell you how you're wrong. :-)
>>17069303 It took approximately 38270 billion man-hours to produce this slop, not counting millennia of work required to produce the training data and computing infrastructure.Note how this post makes mentally ill golems lose their minds with rage, even though they can't refute any aspect of it.
>>17069315>Open AI operates at a loss, indebt, for now.So do universities.
>>17069303nigger it's just a computer program, literally an autocomplete code. stop watching dystopian science fiction movies
>>17069305>8 hours/day/year/teacher to be an educated nonce>20 years>150 teachers>180 days>8*150*180*20 = 4 million hours to produce your idiocy per post>(You) post 30 times per day for 6 years >6,000 posts>4 million hours/post * 6,000 posts = 24 billion man-hours of work for your slopAI is far more efficient to train and outputs better data than (You). I move we decommission you and reassign your predicted aggregate lifetime salary, living expenses and collateral costs including medical costs to keep alive (including any benefits or any costs associate with public infrastructure you benefit from) and reallocate all money to a local AI data center in your community.
>>17069303>On average, each result used three hours of ChatGPT Pro thinking compute with that model.Does literally anyone actually believe this?
>>17069403a computer program that can do a surprising number of supposed "high skill" jobs is definitely something to worry about, anon
>>17069303Anyone glazing this is a midwit lol
>>17069303>It's smart therefor it's gonna kill us allMidwit to dimwit take.
>>17069303It's GG for science.
>>17069503>this filtered NPC reply againClearly multiple spambots from the same shillfarm.
>>17069303that is not creativity. not coming up with new ways of thinking. they're just very good at collating existing data and following rules.which is different from discovery. it's good, it'd damned good, and will help a lot, but it is not discovery. it is not good at coming up with new models. it cannot do what Einstein did, which is still a nice property of wetware. they will eventually get there methinks, but not yet
>>17069568This.
A big issue with the problems solved is that they're all brute forced "ugly" computations. The problems themselves where never so important that humanity NEEDS them solved. Think of it this way, the way Chatgpt is solving these problems you end up with answers like 0.1904761905... where had a math group solved it they would get (4/21) ... again this is a gross simplification but chatgpt is essentially answering with the most bruteforced calculation intensive ways rather than elegant revolutionary ways
>SIX!>>Uh, what's that.>It's the answer to the Mang-Yills Existence in Mascara problem, neeerd!>>The solution to the Yang-Mills problem won't be a number, it'll be an entirely new concept in math altogether.>... oh.
>>17069312Yeah it is, the AI worked for 3 hours
>>17069305>Nooo you can't learn from existing information!
>>17072663Is "AI" in the room with us right now? Are you taking your Thorazine as prescribed?
>>17072668>Noooo you can't just stop me arguing with voices in my headYour meds. Just take them.
>>17069360>>17069315So do farmers.
>So do universities.>So do farmers.
>>17072669>>17072670>Are you taking your Thorazine as prescribed?Are you?
>>17072702>golem seethes impotentlyHow come your magical math-solving genie can't do anything in 3 hours without billions of people working around the clock to keep the global economy and technological infrastructure running?
>>17072604So it's basically like when Claude beat Pokémon. But then, where will this tech be in 2 years?
>>17072705>Why can't your baby provide for itself? Obviously it will never reach maturity!
>>17072729Your psychosis is even more severe than I thought. I didn't expect you to say the imaginary entity is your baby, I'll give you that. More original than anything to come out of an AItroon this week.
>>17072732Sigh....I thought you could at least understand a metaphor. Gave you too much credit. You are completely autistic I guess.
>>17072739I'm sure you have a non-insane way to explain your "metaphor". You're just not gonna do it in your next post for some reason.
>>17069303>gonna die pretty soonMath ability is meaningless, why do people keep making this mistake? Moravec's paradox is well known.
>>17072712Wait, Claude beat Pokémon?!?Since Pokémon is Turing Complete, that was pretty much all the power Claude needed to control the world.It’s all been a simulation ever since then.Wake up.
>>17069563Except it can't and the proofs are wrong
bump
>>17073305This is funnier given how Ford has produced nothing but garbage for at least the past decade.
>>17069403>stop watching dystopian science fiction moviesBut you're already in one. Closing your eyes doesn't change that. AI psychosis is unsustainable and doomed to failure but that doesn't prevent it from destroying your life.
>>17069305I guess its to be expected that people here on /sci/ are seething about this. Change can be scary.
>>17069305Those costs are already sunk, retard.
>>17075765>nonsentient mongoloid guesses the tokens incorrectly
>>17075762>nonsentient mongoloid has his temperature parameter set to 0, unable to shit out non-generic strings
This is where the glowie has seen the evidence.
>>17072617>it compiles in lean though>>In that case we need to give you infinity trillion dollars
>>17069303>I guess the AIs might find that useful: putting your little kid or your tired old mom through horrific experiments and reviving them over and over to stress-test human biology.Finally a use for blacks.
>>17069303Depressed frogo
>>17069315>This is only possible in a fiat money economy.this. I wish this more more common knowledge
