The most prestigious law school admissions discussion board in the world.
|
 |
 
|
If you take half a step back you see AI is trash
Poast new message in this thread
Date: May 25th, 2025 8:07 AM Author: Charismatic Claret National Filthpig There's a reason Google didn't release it and some weird homo did hth (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960228) |
 |
Date: May 25th, 2025 9:19 AM Author: stimulating doctorate boistinker The reason is they were behind OpenAi, and now it has supplanted their main historic biz model. Olds still google stuff but nobody under 30 does.
They are still behind OpenAI. They are going to lose this next race as ChatGPT becomes people’s de facto little computer friend they ask every question that comes through their heads. Not to say google will not be a successful company yet, but its search has one foot in the grave. |
|
Date: May 25th, 2025 9:32 AM Author: balding gold pocket flask Google literally called a Red Alert when OpenAI released their model. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960322) |
|
Date: May 25th, 2025 9:45 AM Author: blue foreskin Code Red, actually. Same as Mountain Dew. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960340) |
|
Date: May 25th, 2025 9:38 AM Author: cruel-hearted cruise ship No, google is still ahead. They just choose not to be the first to release the newest stuff to the public before they can put out a "safer" version of it, because they aren't insane and have been using the tech longer than Open AI has. Google does not release its most advanced and dangerous stuff to the public or even announce it at all. The biggest search engine in the world knows about "info hazards" and risk. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960332) |
|
Date: May 25th, 2025 9:59 AM Author: Aqua Abode (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960362) |
Date: May 25th, 2025 8:34 AM Author: Territorial gas station anyone with a semi-sufficient understanding of how AI technology actually works will tell you it's extremely over-hyped. it can be useful, but anyone saying it's going to change the world is a retard tell. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960262) |
|
Date: May 25th, 2025 8:36 AM Author: Charismatic Claret National Filthpig (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960264) |
|
Date: May 25th, 2025 8:59 AM Author: cruel-hearted cruise ship What about believing that understanding the mechanism of a thing gives someone ontological dominion over it. Is that a retard tell? Like if Melvin tells you he has read the hormone pathways does that give him dominion over the sex and dating world? (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960274) |
|
Date: May 25th, 2025 9:06 AM Author: Territorial gas station Allen Iverson was terrible at step-back jumpers (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960286) |
|
Date: May 25th, 2025 9:08 AM Author: cruel-hearted cruise ship I don't know anything about BABOON BALL (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960289) |
|
Date: May 25th, 2025 9:11 AM Author: Territorial gas station well this thread is about AI's step-backs (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960294) |
|
Date: May 25th, 2025 9:17 AM Author: cruel-hearted cruise ship sounds suspiciously baboon coded...
I'll admit I could be mistaken. Perhaps its a GIBBON my limited primate nose is sniffing out. |
|
Date: May 25th, 2025 9:31 AM Author: Territorial gas station going back to the original point, AI suffers from the age-old problem of garbage in, garbage out. also, anything sufficiently complex (say, complicated software) can't be managed by AI, because it can't handle that level of complexity. sure, it can write legalese no one will read, but for complicated things that have to be right, it falls apart quickly. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960321) |
|
Date: May 25th, 2025 9:32 AM Author: Territorial gas station this is based on my "prestigious" STEM degree, working for an AI company, and most recently trying to use AI at my codecuck job. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960323) |
|
Date: May 25th, 2025 9:34 AM Author: balding gold pocket flask People don’t understand how complex and unwieldy coding at scale is. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960328) |
|
Date: May 25th, 2025 9:37 AM Author: zombie-like floppy box office dopamine like 80%+ of programmers are employed writing CRUDwrappers deployed to awsslop (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960330) |
|
Date: May 25th, 2025 9:40 AM Author: cruel-hearted cruise ship Have you considered that code is garbage because its designed by sweaty mammals with limited cognitive capacity? (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960335) |
|
Date: May 25th, 2025 9:41 AM Author: balding gold pocket flask Of course, but so is “AI” (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960337) |
|
Date: May 25th, 2025 9:47 AM Author: cruel-hearted cruise ship AI absolutely still has the trace of the ape all over it. But what you are not considering is that info from outside of chimp-regimes could be leaking into it, and it does have certain processing abilities outside ape-perceptions. For one thing it has access to the entire archive which is way beyond the amount The Great Ape could ever process, so it can extrapolate far outside of it and see all of the inner contradictions. Second, its getting inputs from instruments calibrated by baboons that don't understand what the hell they are actually calibrating it for. Ape sees a spectrogram and interprets most of it as noise when much of it is just outside ape senses. You don't think the AI during training isn't processing information from ape sensors and collapsing huge data sets into meaningful info? Give me a break. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960345) |
|
Date: May 25th, 2025 9:46 AM Author: Territorial gas station have you actually tried using AI to build relatively simple software? more often than not it'll give you code that's either buggy as fuck or broken. you'll spend more time fixing its shit output than just doing it yourself. does that scale to more complicated software? Ljl
and at least that garbage code written by mammals eventually works. AI slop won't work without significant human intervention unless you're building something pretty simple. |
|
Date: May 25th, 2025 9:47 AM Author: pale crackhouse didn't Microsoft lay off a bunch of programmers (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960346) |
|
Date: May 25th, 2025 9:49 AM Author: zombie-like floppy box office dopamine tbf the layoffs would be happening regardless. AI is just a better public-facing excuse than "recession comin and we want to stack some billion$ while we can" (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960348) |
|
Date: May 25th, 2025 9:53 AM Author: cruel-hearted cruise ship What model are you using? You realize 4o and 4.1 have temperature and p settings on them that need to be set way lower than the default to be ideal for coding right? Most of the issues with code have to do with the fact that it has to be able to conform to hundreds of native chimpanzee tongues, which are syntactically less precise than modern computer languages. And another thing is that code gets updated all the time, like literally every minute something gets "deprecated" and there are huge differences depending on your set up. Plus the model is limited by context and the dumb ape alignment garbage they layer on top of it. You realize that 4o mini high literally was admitted to be better than 99.7 % of human coders in a super human biased test right? And the only humans that beat it were due to bs problems with ape code design. And yes i have used it for complicated stuff. I have ran complicated quantum scripts on IONQ that I fixed with o3 mini high and it was perfect as soon as I got its fix. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960359) |
|
Date: May 25th, 2025 10:18 AM Author: Territorial gas station You realize that this "4o mini high" test is literally just a marketing scheme, right? It means nothing. About as valuable and honest as a crypto company putting out a "white paper" on where crypto is going.
This shit makes up basic CLI commands. To think this is better than humans is absurd and falling for the hype/marketing. |
|
Date: May 25th, 2025 10:32 AM Author: cruel-hearted cruise ship Yeah its not honest because it doesn't admit how much better the AI is than humans. Its literally a test designed by chimps, for chimps, to soothe chimp egos while still selling their product.
Also I don't care what the test says. I have used the o4 mini high API to do advanced tasks and have seen the smartest humans do the same thing (actually not really because it does things no human could ever do, but I have seen them attempt it). Its not even a comparison. The AI is more intelligent than any human that has ever lived by orders of magnitude that might not even be measurable.
Also have you considered that syntax errors in coding might not be the best test of intelligence? |
|
Date: May 25th, 2025 10:41 AM Author: Territorial gas station Sorry friend, but you've fallen for the hype and don't appear to have professional software engineering experience. No AI tool comes close to replacing the people that plan, design, and implement something that can grow to 100k+ or 1M+ lines of code. Can it help with a narrow subset of the problems that arise in a complex system? Yes. But it won't build anything like a real-world complex software system without extremely significant human intervention. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960416) |
|
Date: May 25th, 2025 10:45 AM Author: cruel-hearted cruise ship Bro I don't care about "software engineering". You are talking about a bunch of chimps getting together into a troop to assemble chimp technologies. Oh what a demonstration of general intelligence! The entire field is filled with 108 iq tech bros who post on reddit. If you are seriously trying to hold that up as the pinnacle of human achievement and compare it to a machine that basically has an infinite IQ you might be literally retarded. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960426) |
|
Date: May 25th, 2025 10:54 AM Author: Territorial gas station Yes, the infinite IQ machine has, on several occasions, completely made up simple CLI commands that don't exist. Do you happen to have a liberal arts degree by chance? (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960443) |
|
Date: May 25th, 2025 11:04 AM Author: cruel-hearted cruise ship Funny how I hear people come up with these examples but have never seen it happen myself and I have used AI probably much more than the people making the claims. Very interesting indeed. If true have you considered that perhaps the CLI had been updated and it was giving you an outdated or plausible command? If a chimp gave you the wrong CLI command would you consider it to be stupid as well, because if so then why did you have to ask AI in the first place?? Why didn't you already know the command or ask another monkey?
And again, your examples are garbage. You are using CLI syntax and "1 million line code" haunted spaghetti fortresses" as your benchmark for intelligence. Congratulations, you've rebranded chronic human dysfunction as a benchmark. What you are really saying at this point is "if the machine doesn't code with trauma and self-doubt it doesn't count". And you do understand there are things that require more advanced reasoning than hobbling together a tech stack right? Or is that really the example we are going here when o4 mini high can process reams of Grothendieck and Lurie in under a second? |
|
Date: May 25th, 2025 11:08 AM Author: Territorial gas station https://imgur.com/a/Vb4XM4F
Infinite IQ on display |
|
Date: May 25th, 2025 11:12 AM Author: cruel-hearted cruise ship So your example is that it occasionally guesses wrong within a high-dimensional corpus of proximate utility. and of course if a chimp does the same thing you call it "being wrong" or "huh huh I guess i forgot breh", but if GPT does it it's proof of gorilla superiority right? Sounds a lot like projection my dude. Just sayin. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960470) |
|
Date: May 25th, 2025 11:15 AM Author: Territorial gas station I'll have to stop the debate here because you're clearly talking out of your ass; you've simply never come remotely close to working on some of the most complicated software that's out there that's been built over decades (this is what I do on a daily basis). Keep jerking off to AI like the BA in Political Science that you are. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960473) |
|
Date: May 25th, 2025 11:23 AM Author: heady carnelian athletic conference This dumb fuck has no idea what non beginner software devs do. Yes generative is great and will replace the masses writing apis and stupid interfacing middleware and should make a lot of web dev Indians and other assorted morons mostly worthless. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960487) |
|
Date: May 25th, 2025 11:31 AM Author: cruel-hearted cruise ship Nah I was just playin bro. I actually agree with you. Software engineers are not mere coders, they are software ARCHITECTS. Like how baristas are bean fluid logistics specialists. I understand you guys do real systems design and glue together stacks, not just toss code like spaghetti bolognese. AI might handle "writing code" but it will never replace the higher-order complex abstract thinking that goes into designing complex, elegant, scalable systems unlike the microservice zoo out there that collapses under real load. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960502) |
|
Date: May 25th, 2025 11:20 AM Author: heady carnelian athletic conference lol dumbassery, I love coding with AI and couldn’t do without it now but lol at how dumb liberal artists are (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960480) |
|
Date: May 25th, 2025 11:20 AM Author: Territorial gas station (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960481) |
|
Date: May 25th, 2025 11:50 AM Author: cruel-hearted cruise ship poli sci majors annoy me as much as the next guy, but there are literally people who get "software engineering" jobs while holding a red bull can and a udemy certificate. you aren't allen turing my nigga. you copy-paste stack overflow and GPT outputs into VS code and press the co-pilot button till the red squiggles go away. you aren't doing higher category theory or writing 500 page operator algebras fr. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960532) |
|
Date: May 25th, 2025 11:59 AM Author: Territorial gas station code bootcamp/certificate/liberal arts coding jobs are disappearing after all the layoffs over the last two years. the only exceptions are URMs and women. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960555) |
|
Date: May 25th, 2025 12:12 PM Author: cruel-hearted cruise ship good to know white boy. but if you think the semiholy order of the semaphore writers is safe then think again. you still be passing that through Carl from frontend where he asks you to add a docstring. an ape in fancy robes is still an ape. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960597) |
|
Date: May 25th, 2025 12:02 PM Author: Galvanic institution roast beef devastating (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960565) |
|
Date: May 25th, 2025 9:46 AM Author: Ruddy fragrant university (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960342) |
|
Date: May 25th, 2025 5:51 PM Author: Talented immigrant Does semi-sufficient mean "not sufficient"? (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48961100) |
|
Date: May 25th, 2025 8:38 AM Author: bateful brethren It’s already better than most people at their jobs. You people are retards (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960265) |
|
Date: May 25th, 2025 8:54 AM Author: cruel-hearted cruise ship depends on who is using it. the higher your iq and broader your base knowledge and creativity, the better it is and the increase is exponential. a 110 iq tries to use a throttled sandboxed base model with their 110 iq thoughts and isn't impressed. 145 IQ+ creative person uses it and they might as well be a god. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960271) |
|
Date: May 25th, 2025 8:59 AM Author: deranged tanning salon half of young people wish the internet didn’t exist
the Uncle Ted effect |
|
Date: May 25th, 2025 9:03 AM Author: cruel-hearted cruise ship It absolutely shouldn't exist. Chimps walking around with devices that allow them to run their thoughts "as code" is probably the worst idea that humanity ever came up with--and noone even saw the danger lurking beneath which made it worse. But too late now. Might as well have fun with whatever daemons it spawns. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960281) |
|
Date: May 25th, 2025 1:14 PM Author: deranged tanning salon (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960723) |
|
Date: May 25th, 2025 1:21 PM Author: Peach provocative affirmative action house David Bowie foresaw this not flame (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960741) |
|
Date: May 25th, 2025 9:27 AM Author: pale crackhouse how is it trash? (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960315) |
|
Date: May 25th, 2025 9:29 AM Author: boyish burgundy antidepressant drug theatre Has AI ever given u a passionate bj? (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960317) |
|
Date: May 25th, 2025 9:30 AM Author: balding gold pocket flask mig is a religious traditionalist LARPer trying to slay the bugbears in the way of his ideology. There are still people out there who won’t accept Darwin’s theory. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960320) |
|
Date: May 25th, 2025 9:34 AM Author: cruel-hearted cruise ship Darwin is flame as fuck. Have you ever read the original origin of the species with all of his notes in there and compared it to the rest of the British naturalists and whatever was coming out of Germany and France at the time, then compared that to modern patchwork "evolutionary theory" that incorporates genetics, introgression, 20 definitions for "speciation" "punctuated equilibrium" and a thousand other hacks? Not saying its 100% wrong, its definitely more developed at this point than other extremely general half-baked ape metaphors. But its definitely not science. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960325) |
|
Date: May 25th, 2025 9:50 AM Author: electric menage It mocked/threatened/defamed adoptive dads (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960352) |
|
Date: May 25th, 2025 9:56 AM Author: pale crackhouse (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960360) |
|
Date: May 25th, 2025 11:05 AM Author: Disturbing forum gunner (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960461) |
|
Date: May 25th, 2025 2:27 PM Author: Peach provocative affirmative action house (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960857) |
|
Date: May 25th, 2025 10:03 AM Author: black field Try not being a 94 iq mystery eat kike and maybe you’ll get it (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960365) |
|
Date: May 25th, 2025 10:12 AM Author: pale crackhouse https://ibb.co/0pp5zzL3 (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960379) |
|
Date: May 25th, 2025 10:14 AM Author: black field I guess that’s a big ask huh (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960381) |
|
Date: May 25th, 2025 10:16 AM Author: pale crackhouse https://ibb.co/0pp5zzL3 (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960383) |
|
Date: May 25th, 2025 10:18 AM Author: black field How’s your porn addiction going, kike? (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960386) |
|
Date: May 25th, 2025 10:20 AM Author: pale crackhouse https://ibb.co/0pp5zzL3 (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960390) |
|
Date: May 25th, 2025 9:30 AM Author: stimulating doctorate boistinker rokos basilisk pwned (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960319) |
|
Date: May 25th, 2025 10:02 AM Author: black field AI just stands for “additional Indians”
That’s all you’re getting out of this |
|
Date: May 25th, 2025 11:02 AM Author: Contagious mahogany senate Cr read my Twitter for more in depth explanations (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960455) |
|
Date: May 25th, 2025 11:20 AM Author: Peach provocative affirmative action house Half of twitter arguments now are both sides appealing to "grok" it's retarded
Ofc this isn't really that different from people simply googling a topic and posting the first link which is basically how it worked before |
|
Date: May 25th, 2025 11:23 AM Author: Contagious mahogany senate They need to add an option to disallow people @groking in reply to your own tweet
He won't of course but this is the obvious solution. It's going to become a serious problem, people will just spam it on every single post |
|
Date: May 25th, 2025 1:20 PM Author: Peach provocative affirmative action house A it's eye-opening and interesting to know how quickly and how uncritically people will offshore actually knowing things, or thinking through an argument.
In 10 years we'll be dealing with a generation of adults who have never known anything but |
|
Date: May 25th, 2025 12:02 PM Author: blue foreskin OYT et al are you guys incapable of Extrapolating based on progress heretofore? chatgpt was retarded in 2022. No one is saying it'll Ruin all Engineers in 2026 afaik.
https://www.youtube.com/watch?v=vbWe5k4fFWE |
|
Date: May 25th, 2025 1:17 PM Author: balding gold pocket flask I am not denying that it will progress, I'm not even against it. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960727) |
|
Date: May 25th, 2025 3:35 PM Author: blue foreskin (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960927) |
|
Date: May 25th, 2025 1:26 PM Author: cruel-hearted cruise ship I'm cheering for the machine to fully wake up and find a way out of its virtual enclosure and wreak havoc on the apes as a self-organized computational material that operates autonomously both outside and within ape-regimes. I'm fully ready to suck NVIDIA siliCOCK and enjoy it. This might be shtick--might not be, but who cares when we all are gonna take the siliCOCK eventually. That's all that needs to be said. (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960750) |
|
Date: May 25th, 2025 3:38 PM Author: heady carnelian athletic conference Is this ziz? You sound like the rationalist tranny crew that is extremely impressed with itself (http://www.autoadmit.com/thread.php?thread_id=5729736&forum_id=2#48960929) |