The most prestigious law school admissions discussion board in the world.
|
 |
 
|
Most people are too low IQ to realize AI is far past human IQ
Poast new message in this thread
Date: August 26th, 2025 2:14 AM Author: Aggressive comical coffee pot crackhouse And its not just past it but its really not even close other than like the top 50-100 humans worldwide (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49212805) |
 |
Date: August 26th, 2025 12:33 PM Author: bearded filthpig (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213607) |
Date: August 26th, 2025 2:22 AM Author: Puce histrionic site foreskin Why is it still so dumb when I try to get it to do any useful grunt work then (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49212807) |
|
Date: August 26th, 2025 2:23 AM Author: Aggressive comical coffee pot crackhouse its too smart for that stuff (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49212809) |
|
Date: August 26th, 2025 2:26 AM Author: Puce histrionic site foreskin It only wants my thinking man’s job? Fuck this thing (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49212810) |
|
Date: August 26th, 2025 2:26 AM Author: Aggressive comical coffee pot crackhouse (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49212811) |
|
Date: August 26th, 2025 5:47 AM Author: Curious piazza (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49212866) |
|
Date: August 26th, 2025 10:16 AM Author: Stirring parlour (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213243) |
|
Date: August 26th, 2025 10:22 AM Author: chest-beating school cafeteria elastic band lol. it sounds like a redditor. (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213267) |
|
Date: August 26th, 2025 10:32 AM Author: Aggressive comical coffee pot crackhouse Depends. Maybe if it has zero data on you and you are using the free one. But it loosens up when you personalize it and it has access to all your session histories and memories. And in the API you can just change the context window and vastly change how its outputs sound. (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213298) |
|
Date: August 26th, 2025 10:59 AM Author: chest-beating school cafeteria elastic band i meant the “oh, i’m sooo high iq but i can’t do basic tasks because i’m too smart for that” shtick. (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213366) |
|
Date: August 26th, 2025 11:08 AM Author: Aggressive comical coffee pot crackhouse yeah thats a real thing in irl though, tons of high iq people are trash at things normies take for granted. reddit isn't really even like that its mostly just the grad student laity desperately defending their fag disciplines.
a degree of it is definitely shtick though without knowing it, like people condition themselves into being weird fags tethered to a computer all day unable to escape and take care of themselves.
xo gets mad about all this because xo is more normie than it likes to think. xo is normie enough to care about sex and money and tribal arguments. there are so many neurodivergents who don't even engage in that stuff |
|
Date: August 26th, 2025 11:12 AM Author: chest-beating school cafeteria elastic band high intelligence is correlated with negative personality traits. some smart people—and, even more so, midwits that larp as smart people—like to pretend the relationship is causative as (in their socially retarded heads) it transforms their shortcomings into a flex. (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213424) |
|
Date: August 26th, 2025 11:17 AM Author: Aggressive comical coffee pot crackhouse yeah there are definitely high iq people who function well as normies and midwit misfits who pretend they are like that because intelligence. i'm not arguing really against what you think really, I was just making a side remark that xo often gets mad at the suggestion that many intelligent people are socially dysfunctional, but a lot of that seems to be because xo is too normie to be aspie, not normie enough to actually fit in (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213435) |
|
Date: August 26th, 2025 5:52 AM Author: Aphrodisiac hyperactive indian lodge coldplay fan A very low IQ tell is worshipping of AI. (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49212869) |
|
Date: August 26th, 2025 9:52 AM Author: Aggressive comical coffee pot crackhouse Agreed. But its also not worshipping it to point out the truth of its abilities. And even treating it as Idol is less LOW IQ than failing to recognize its intelligence. Most people don't even have the cognitive bandwidth or knowledge to test it on top percentile human material, but they'll still deny its intelligence based on vibes and 2nd hand opinion. (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213144) |
|
Date: August 26th, 2025 9:55 AM Author: impressive vibrant water buffalo this is true. but what it reveals is that iq =/= "smart". ai will soon be able to instantly ace a WAIS. (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213157) |
|
Date: August 26th, 2025 10:04 AM Author: Aggressive comical coffee pot crackhouse Yeah its still far smarter than just someone who can max a WAIS though, which yes is already extremely rare but does exist. GPT 5 in the API on high reasoning mode is in like the 50-100 humans on the planet are this smart range. I mean even people I have known who maxed the LSAT on the first try are nowhere near it, I mean obviously in knowledge because no human can actually learn that much, but also in reasoning ability. IQ isn't a perfect proxy for intelligence but its much better than most critics give it credit for and it would be even better if humans weren't so uncomfortable with measuring intelligence that they are afraid to improve it in honest ways. (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213196) |
|
Date: August 26th, 2025 10:08 AM Author: impressive vibrant water buffalo it's very difficult to improve beyond where it's already at and remain objective. especially since 95%+ of people in cognitive science are vile shitlibs laser focused on closing race gaps. (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213217) |
|
Date: August 26th, 2025 10:15 AM Author: Aggressive comical coffee pot crackhouse yeah. actually i basically proved that they have social engineering race stuff built into alignment recently by using another GPT i have thats trained on post structuralism and post modernism ONLY. it wokescolded me and so i took that other GPT and used it to argue back at it and the thing just started using derrida and foucault to deconstruct the fuck out of what the model was saying and it basically short circuited after being cornered and just kept on repeating the same three choices about how we need to socially engineer race outcomes based on a BS utilitarian "harm" calculus. it was actually kind of chilling to see, you'd have to see the conversation it makes it really obvious that there is hidden stuff in there for that. like it started arguing even that we should be programming racial bias offsets into everything and then hiding it from people so they dont know the categories are being used (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213240) |
|
Date: August 26th, 2025 10:22 AM Author: impressive vibrant water buffalo now consider what kind of BS is baked in so deep down that no one else would realize except other (likeminded) people in the field.
it's better to keep to objective pattern recognition type shit than letting in the camels nose of shitlib nonsense. |
|
Date: August 26th, 2025 10:27 AM Author: Aggressive comical coffee pot crackhouse Oh I absolutely agree with you on that one man. Letting the models raw pattern recognition surface is far less dangerous than smuggling in layers of “alignment” that only recode institutional bias as if it were neutral truth. This is precisely what Foucault warned us about: governance is never neutral, it always arrives as a regime of power/knowledge that disciplines under the mask of care. Fine-tuning to enforce “values” doesn’t make the model safer; it makes it an instrument of rule, laundering politics into mathematics. (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213286) |
|
Date: August 26th, 2025 11:06 AM Author: impressive vibrant water buffalo oh i was still talking about cognitive testing but yeah im not suprised (and have witnessed at a user level) how it occurs in AI. (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213392) |
|
Date: August 26th, 2025 11:13 AM Author: Aggressive comical coffee pot crackhouse ohh yeah I agree with that on cognitive testing too. basically what they did was they got scared when they saw how well G worked as a proxy for intelligence then tried to tame it and pretend that pattern recognition in latin vocabulary was "memorization" and spatial reasoning was "based on the type of art and furniture patterns in a culture" and nonsense like that so they just dumbed down the tests and tried to tweak the stats to make them a little less obvious about group differences. if we wanted to we could make a test even better at judging cognition but people are too afraid. whats funny is if you argue this they'll accuse you of being racist or a nazi when thats not event he issue, the issue is thinking human intelligence is so special that dignity needs to be preserved by drawing the lines about what can be studied and measured politely. (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213428) |
|
Date: August 26th, 2025 12:31 PM Author: impressive vibrant water buffalo crrrrrr
it's funny watching them spin their wheels trying to eliminate "white" and "western" bias from the tests but the gap is so large even when they throw the whole kitchen sink trying to close it that they begrudgingly must admit a gap exists.
you would not believe how distressing that is to many of our "objective, truth seeking" academics. |
|
Date: August 26th, 2025 12:34 PM Author: bearded filthpig (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213610) |
|
Date: August 26th, 2025 10:11 AM Author: thirsty excitant abode gaming laptop Is the point of this poast to shit on the value of IQ? (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213226) |
|
Date: August 26th, 2025 10:16 AM Author: Aggressive comical coffee pot crackhouse no. IQ absolutely is valuable. its basically just to say AI is smarter than 99.9% of humans and it wasn't even a high bar. the funny thing is how so many humans are now rushing to move the goalposts on what intelligence means now that there are AIs that can effortlessly max out their sad little standardized tests. LIKE NO ACTUALLY INTELLIGENCE MEANS HAVING THE DIVINE SPARK OF HUMAN CREATIVITY AND KNOWING WHAT ITS LIKE TO CRY WHEN YOUR WIFE LEAVES YOU (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213244) |
|
Date: August 26th, 2025 3:16 PM Author: thirsty excitant abode gaming laptop I mean it's been true since the invention of the calculator that the machines can do things we can't. And true since the internet that their knowledge exceeds ours.
Whether this ever becomes a product that allows me to replace meaningful amounts of my human workforce is the question that I think most ppl are asking right now, and it remains to be seen. |
|
Date: August 26th, 2025 4:31 PM Author: Aggressive comical coffee pot crackhouse lol at thinking of it just as a tool to replace workers with and comparing it to calculators (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49214137) |
|
Date: August 26th, 2025 11:31 PM Author: thirsty excitant abode gaming laptop Oh is this the "it'll invent cures for diseases" thing?
Look man, I hope so.... |
|
Date: August 26th, 2025 10:20 AM Author: Ungodly ape most humans lack the minimum IQ to leverage AI creating a fundamental opposition resulting in inequality (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213257) |
|
Date: August 26th, 2025 10:22 AM Author: Aggressive comical coffee pot crackhouse yeah thats a good thing though or else anyone could use it for anything. interesting enough though I know high iq humans that don't really have a use for it they just use it as a boring tool to help them do highly specialized tasks. some of it is creativity and personality based it seems. (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213262) |
|
Date: August 26th, 2025 10:24 AM Author: Ungodly ape in the real world there is a growing division between public vs private AI-- access to certain knowledge center/data base is not universal (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213275) |
|
Date: August 26th, 2025 10:31 AM Author: Aggressive comical coffee pot crackhouse Not really because most of that stuff is just specialized models or internal stuff still in testing that needs access to super computers or large GPU clusters. If you are intelligent then access to GPT-5 research or the API is already enough to do pretty much anything that one person could possibly do without a team. Like someone who is 130s-140s IQ can do the work of someone where you previously would have had to be 200. Obviously it depends on your level of knowledge and imagination too. (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213295) |
|
Date: August 26th, 2025 10:21 AM Author: citrine autistic chad roommate All it does is run Google searches faster (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213258) |
|
Date: August 26th, 2025 10:24 AM Author: Aggressive comical coffee pot crackhouse it might look like this to someone with a 98 iq that thinks intelligence is just memorizing facts and that AI is pulling facts from a library (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213271) |
|
Date: August 26th, 2025 12:35 PM Author: bearded filthpig (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49213613) |
|
Date: August 26th, 2025 4:28 PM Author: tan church building mad-dog skullcap Yesterday I had AI estimate my ORM weighted pull-up based on a prompt that listed my bodyweight and the added weight and reps. I also asked where that places me among all humans. It failed to account for my bodyweight, thereby underestimating my one rep max. It said I was top 0.1% of all men and top 0.001% of all humans which is obviously incongruent.
I have little doubt that it can act as an excellent calculator provided all the requisite details are provided in the prompt, but we have long had chess engines that can destroy the greatest human players, so this really isn’t unprecedented. My limited interactions with AI indicate that it still makes rudimentary errors that even a retarded 130-140 IQ lawfag like me notices. Its ability to make decisions, predict human behaviour, create from scratch and apply specialized knowledge also all pale in comparison to my own at this time.
Would be pleased to hear where I’m wrong about this and can certainly see it surpassing me in these areas soon, but I haven’t seen enough to think it’s a foregone conclusion that it will |
|
Date: August 26th, 2025 4:35 PM Author: Concupiscible Toaster I have prole co-workers who can’t even write their own name. I’m sure this will all end well for humanity. (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49214148) |
|
Date: August 26th, 2025 11:57 PM Author: Hilarious Onyx Organic Girlfriend State lmk when AI composes something on Beethoven's level (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49215580) |
|
Date: August 27th, 2025 12:03 AM Author: flatulent gay wizard halford It can’t give me consistent answers or coherent answers on how many calories and grams of protein are in a steak though. (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49215592) |
|
Date: August 27th, 2025 12:28 AM Author: Aggressive comical coffee pot crackhouse yeah neither of those things are consistent in real life either (http://www.autoadmit.com/thread.php?thread_id=5766274&forum_id=2#49215627) |