Will ChatGPT destroy universities? Posted on May 18, 2023 by Anonymous Will ChatGPT destroy universities?
It's the university's own fault. Imagine spending 4 years learning 100% solved problems. You will NOT add any nuanced new take on anything you learn in uni. It's basically just teaching you processes and procedures that are now 100% automated.
Literally the only thing that matters is creative fields where you really are only cheating yourself by automating too much of it. These classes are typically 100% non-mandatory electives or they're your major.
ChatGPT is a wake up call to these lazy educators. Maybe teaching the same basic content and lessons for multiple decades on end isn't actually valuable with how fast technology moves.
scientific research is a creative field, and cheating on problems is cheating yourself out of future contributions. knowing your fundamentals is essential for discovering new things.
>he think humans will be doing any research in 5 years
AI isn't nearly as intelligent as people think it is right now. It's still just regurgitating information. We're a long way off from truly thinking and innovative AI.
Oh cool, I didn't know DeepMind guys posted here. Or are you at OpenAI? Maybe Meta? You must be at one the the top labs, since you're saying you solved the interpretability problem and understand exactly what's happening in LLM's.
>It's okay for me to speculate that AI is way more advanced than it demonstrably is
>It's bad for you to speculate that AI is way stupider based off empirical evidence of existing AI
I work at pic related, if that counts. Not in MLAI though, that shit's gay.
chatGPT != all AI
Computer assisted math proofs aren't a new thing. Neither is AI discovering new potential medicines, materials, or discovering new potential patterns in just about anything you name. It may not be generalized yet, but AI is doing more than putting data into a blender and spitting out that same data, but mixed up.
I didn't say ChatGPT was AI, but even in your example AI is not rationalizing jack shit and it's no where near as smart as even a dog. Its best application is pattern recognition and it isn't "intelligent".
>Its best application is pattern recognition and it isn't "intelligent".
So IQ tests... Those things that measure intelligence by uhhhhhh... Human bros... I don't feel so good!
>IQ tests measure intelligence
lol, yeah I have a 150 IQ based on my online IQ tests. I'm really smarter than 99% of people, especially ones that are more successful than me.
It is the equivalent of a food processor makes good sauces but not always
>It's still just regurgitating information.
So just like humans
>We're a long way off from truly thinking and innovative AI.
So just like 99.5% of humans
Newsflash, you aren't creative for learning organic chemistry. The undergrad level is all completely solved. The only creativity in university typically happens outside classes.
Yeah those geometry classes really are super creative and definitely not just copy pasted problems over and over. Maybe, just maybe, we should give students incredibly hard problems, problems that AI will actually fail at solving and then tell them what fields they need to study, prompt the AI for information on?
Maybe we don't need to teach them integration by parts anymore? Maybe only those that actually deeply care about math need to learn that and others just need to learn how to properly engage with the tools at their disposal to be as efficient as possible? Maybe the bottleneck for humanity is wasting 20 years of kids lives learning completely solved problems instead of throwing much more difficult things their way?
The real way to teach would be to present them a set of information on a subject and then make them actually write theories on it's application and depending on the topic actually attempt to apply it, and from there it's just a matter of critique and refinement
The purpose of undergrad is that it's pointless. The point is that education is continuous, each class is supposed to smoothly segway into the next one. You learn u substitution because that was how the gaussian was originally solved, then 2d calculus is generalized into 3d vector based calculus, and the gaussian is solved again using polar, then that gets segwayed into how the gaussian was applied to probability to produce statistics, then you learn generalized distributions and maximum likelihood estimation.
The point is that most everything you learn in order is the order of when it was discovered, and each thing you learn was derived from the previous. In the end if you make it then you are supposed to add another discovery, thus continuing the chain. Then this will continue until humanity dies out. You can't put the horse before the cart, and you can't teach integration before people learn addition/subtraction because then it would be impossible to tell if they actually "know" it or not. In fact it's impossible to tell if somebody has learned anything, but education is there to just reduce the probability that somebody becomes a math professor without knowing addition/subtraction.
I don't think any of that is necessary and creates a bottleneck on discovery in related fields. By having to learn all of this before committing to a specialization, you really disadvantage research.
Anyway, it doesn't matter, we're going to see very creative analyses coming out in the future. Again you don't need to be an expert on process and procedures anymore. We are leaving the information age (i.e, the age where information and knowledge are valuable) and entering the creative age (where new and novel things are valuable since information and knowledge are so cheap and easy to acquire).
The AI can not create, it can only amalgamate based on what is already out there. Value then will be placed on those able to actually produce new things using all this cheap information.
You can't create something using regurgitated information unless you actually understand the meaning and context behind the regurgitated information.
No, I think you can. I actually think you 100% can. You are overestimating the value of understanding processes versus just being able to effectively use them. AI will allow people to effectively use processes without them even knowing them.
That's why we are exiting the information age.
>No, I think you can. I actually think you 100% can. You are overestimating the value of understanding processes versus just being able to effectively use them.
Citation needed for claims.
Just what I'm seeing from the well made AI content. I don't think some of these people are particularly effective directors or writers, but there still making content that's fairly avant garde and punching well above their weight class as artists.
I think you're wrong. I think precisely because we are taught the basics (because we have to be taught them) that technological progress has slowed down. Even in the 40s you had people with just masters degrees making monumental contributions to science (like Dyson). This is due to the scope of knowledge at the time being honestly much smaller.
I'm not saying that these problems aren't useful, or even that experts won't still have to study them, it's just that if you're not an expert and need them you don't need to waste time studying them anymore.
Universities only course of action is to actually accelerate their programs in order to compete with AI and teach what should typically be covered in a PhD (not including the thesis) to undergrads. All those first two year courses are 100% useless now.
>This is due to the scope of knowledge at the time being honestly much smaller.
It's also due to high school education being much better in many cases, and there being fewer universities in existence, so really only the elite bothered to go. Up until the 60s, Cal Tech undergrads were assumed to know calculus before enrolling. Nowadays, universities have to waste an entire year just getting students up to speed to be able to study useful topics. Of course a lot of the worthless mandatory option courses are indeed a huge waste of everybody's time, and this is without including the globo-homo idiocy.
High school should really be divided into 3 tiers, a bottom level for recalcitrant Black folks (the kind that just disrupt the classroom, attack the teacher etc.), a standard track for most normal people, and advanced studies for the future scientists, engineers and future leaders. With such a program, you could probably cut undergraduate education to 2-3 years and accelerate students into a master's program.
this exact system you describe exists in German high schools
Based Germans. What names do they use for this stratified education stream? As I wrote it, I also wondered how you'd account for late bloomers who due to certain circumstances were not in the program they should have been in.
They are called Hauptschule (Main School), Realschule (Real School, whyever it's called that) and Gymnasium (Gymnasium, not sport exclusive that is). The idea is that in the last year of primary school one's performance and consultation between teacher/student/parent guides the way for the students' further education based on career wish, academic performance and such. A pretty good sorting system overall I reckon.
"Real school" sounds like where the Untermenschen go, to "keep it real, yo!", and main school should be for everyone else. I just assumed Gymnasium was high school, never realized it was the top stream.
Hauptschule is mainly for trades and other primary industry-type skilltrees. Realschule is a blend of the three tiers, and Gymnasium is more white collar / STEM focused.
>I don't think any of that is necessary
Well that's why you're not a professor. Mathematicians and scientists know how they got to where they are, and have tuned college curriculums over the centuries to match that. To some extent you're supposed to forget a portion of what you learned in college, it's that way by design. Most of it is just a test, obviously most working scientists don't need to remember literally everything they learned in organic chemistry or whatever lower level chem class, but the lower level classes are just a bridge, a door if you will, to where they need to be.
>We are leaving the information age (i.e, the age where information and knowledge are valuable) and entering the creative age (where new and novel things are valuable since information and knowledge are so cheap and easy to acquire).
It's called the imagination age
I wouldn't consider the information age as separate from the preceding periods personally. After all, there has been an unbroken stream of discovery since the beginning of the renaissance 500 years ago. The end of the information age will be the end of information.
Delusional absent minded drivel. You think it is natural to pay for worthless classes that you admit is not used.
Post studies on this imagination age. You are full of shit.
And I'm saying that bridge basically doesn't need to effectively exist for 90% of those in university.
All you need to teach is input and output. The in-between processes of these courses are becoming unnecessary and these courses will have to react to the redundancy created by AI.
Does the person coming out of the average biology degree really need to KNOW organic chemistry or do they just need to know that it exists and how to prompt the AI for detail when necessary?
We are entering the end of information. Again we are going to see the same effect that maps and gps have had on children, namely a dropping of memorization of knowledge. You literally just have to ask the AI. It's different than Google as well, since you still have to read detail, you can literally ask the AI to for exactly the level of detail you need to move on in your life.
What I mean by "the end of information" is that junk information will start to pile up. Misinformation, just plain wrong ideas, pseudoscience, and so forth will start to creep in everywhere. You ever made a xerox of a xerox? That's what will happen as AI starts to put information onto the internet, which will then feed back into other AIs who will put the altered info back out. A decade from now you wont even know who the president is. Hell, you wont even know if 1 = 1 or if it equals 2. Bridges will collapse as they're designed by AI which was fed by another AI which was fed by another, etc. Each layer of AI will magnify small mistakes made by the previous one. The world will essentially become perpetually deadlocked as everybody tries to figure out what is real and what isn't.
Imagine this over generations, toddlers shown AI generated images of tree dwelling sharks, or public figures being arrested, etc. Information will no longer exist.
That's when they say fuckit and plug you into the matrix to zoom around interconscious space. We'll ditch this reality altogether and start fresh in a new age.
I guess we should ban calculators too, huh
>Maybe teaching the same basic content and lessons for multiple decades
This. Like who the fuck care about Euclid's elements? They should be teaching students how to twerk and recent black history.
Dat's all Ize need whiboi
They aren't teaching Euclid elements though, are they?
>ChatGPT is a wake up call to these lazy educators. Maybe teaching the same basic content and lessons for multiple decades on end isn't actually valuable with how fast technology moves.
School in the US is a fed money laundering scheme where the academics bankers and money printers are all in cahoots to take as much waking life from students (slaves) because each one (educated) equals more 1 to 1 student to faculty salaries. Which means more fed dollars enter the economy unnoticed in bundles of 100k because it's for the children.
Everywhere else in the world it's totally normal and academia is dying due to it being a bygone institution where you have to go somewhere to learn something.
In the US we are literally eating our children so the fed can print more to flush the economy. Then asshole boomers say pay it back, the government says they will do something yet causes more and more trauma for young people, local governments displace young people 50/50 with foreigners who are here to stay and don't speak English and tell our kids to learn their language to keep up.
The AI is meaningless. The whole system is engineered to torment children so someone gets a salary because the government can't take the reins and put together a functional economy, instead relying on the "experts" of "academia" to save us. When in reality, all academics are liars especially economists, and the only thing that won't stop is the fed.
Destroy the schools, then destroy the fed. Everyone will be free again. Knowledge is easy with technology.
>Knowledge is easy with technology.
But it's not the knowledge. Who can use the knowledge he had from the education system?
Education means dressage, Students are trained to function within the system and at the same time they are inculcated with its ideology.
Different to here (Europe) the victims seems to pay for that scam them self or have to go in debt.
You hit the nail on the head. American higher education has been completely hollowed out to promote this fiat/debt slavery scheme. And you are absolutely correct that the entire system needs to go, sooner rather than later. Unfortunately there are too many interests aligned with keeping it in place.
That's fucking retarded and mediocre. You do need to first understand fundamentals to create newcreative ideas, it has to be based on basic knowledge in the field.
I don't disagree, which is why you should only study the basics of WHAT YOU LIKE. all of these other courses exist solely to teach you the basic processes and procedures to improve your life. These things don't need to be memorized or learned anymore. Just like kids don't need to know how to read cursive or interpret a map, you don't need to know the process of writing an effective argument unless it's something you autistically enjoy (like cursive and maps)
You can already cheat, you retards. There is nothing magical about a chatbot that makes students obligated to use it anymore than they could already use chegg or pay someone to do their homework.
How do you expect to learn something without doing solved problems first? That's the point of bachelor's level education. Also, research is almost always part of undergraduate science programs, so you're actually just wrong.
>4 years learning 100% solved problems.
They're teaching you problem solving techniques, retard. The point is to give you a foundation to work with. While ChatGPT will be immensely useful, fundamentals remain. >Isn't actually valuable with how fast technology moves
Much of the content was very useful to the people "making technology move fast", and while some topics go in and out of fashion, this is mostly at an advanced level.
I doubt many of you (unless you're at a PhD level) will learn anything in mathematics that wasn't known in 1923. Such "old tech" is still useful, always will be.
The whole point is learning the fundamentals you goofball. And when you get into the masters and phd you do make your own new shit. This is like getting mad at grade school for teaching kids basic arithmetic instead of making whole new things. I don’t even understand your point here, have you ever even learned anything in general?
You're not getting the point because you're unintelligent. You need to learn basic arithmetic, you don't need to learn the best way to organize a paper just like you don't need to learn how to read a map or write in cursive (and how you don't need to know Grammer)
I'm not trying to trash everything, I'm trashing these skills that are fundamentally useless and exist only because people need to follow specific processes and procedures. At most, all you need to know what to do to form arguments is a list of pros and cons. You can then have AI construct the rest. That base skill is still useful and necessary, the rest are not.
MOOCs, computers, and the Internet were also supposed to kill universities. At the very most, ChatGPT could possibly change universities, but it won't make universities obsolete.
You dont understand. Those things were predicted to replace the source of education. AI threatens to replace the product of education. Oh you have a degree? Unless you can tell me something new, I'll ask an AI thanks.
Nobody will sign off on an AI to make key decisions in a lot of industries where safety and legal liability are aspects, because of legal reasons and definitions.
You need AGI to replace human decisions. I'm saying most expert jobs are about providing expert advice and those positions will evaporate.
If anything an AI is more rigid and less prone to wild swings in advice than a human who from day to day is moody, forgetful or impaired at worst.
Yeah doesn't matter. The legal system is dependent on human definitions and without humans it won't work.
AI can precisely understand and reiterate human definitions.
AI can't have liability in a court of law.
Neither can war criminals who win their wars.
Not really sure what this is or how it's related to what I said, but I'll take it as a concession.
>Your statement was too big brained so I didn't understand it, but I'll take it as a concession.
You can't even interpret human speech, AI.
Well, now maybe we can kick the brainlets out and a degree will mean something again. I'm capable of original thought so I'm not worried.
“As an ai model I can’t do that…”
Ok there you go. Aligned and safe and approved Ai.
HA HA HA HA HA HA HA HA !!!
In the future all graded assignments will be done in-class. Sucks for the people who are good at take-home work and bad at in-class work lmao
I'd rather do relatively faster quiz's and exams in person than have to trudge through and rewrite a stupidly pointless document that I need to arbitrarily increase the word count on that could be explained in 1/4th of the fucking length
The funny thing is that you can use GPT to mark essays or chunks of essays. So really, AI is judging AI, and we’re just the intermediates.
I know this chatgpt thing is kinda new. I've seen multiple students in a few classes just copy paste the same exact wikipedia articles at the same time, like word for word. Like come on, at least change a few words or sentences.
Education in this country is shit and has been for al long time. Areas where ChatGPT excel are classes like writing which honestly is a waste of time anyways because AI will replace most professional writing skills.
Schools will have to go back to the old days of in person classes and pencil/paper assignments if they want to avoid kids cheating with AI, which is honestly a good thing because people have been cheating under the radar with computers for a long time and not really learning.
God I hope they do away with essays due to ChadGPT. Fuck writing essays!
You don't always punish an evil, the Mona lisa effect.
Think about it.
She is punished softly. She goes back to the scene of her crime... Etc
Universities are giving you a curated education and information
Chat gpt just spits out a combination mosaic of data sets such as books and the internet and social consensus and is then run through filters and censors
It could help one educate themselves but will never replace university
AI could maybe but that probably wont be funded greenlit or allowed tbh
>Universities are giving you a curated education and information
Any (audio)book online lesson or even a video of a lecture too
>Chat gpt just spits out a combination mosaic of data sets such as books and the internet and social consensus and is then run through filters and censors
Sounds like the modern academic process
>It could help one educate themselves but will never replace university
In what? Indoctrination, debt, false authorities, impostering etc?
>AI could maybe but that probably wont be funded greenlit or allowed tbh
maybe compare that with the success of nonacademic fundings.
Prof is bluffing. You can't tell if an essay is AI written. What's the criteria? It is well written, and "rigid".
>You can't tell if an essay is AI written.
No, YOU can't tell.
To sentient beings, it's fucking obvious.
The meandering tone and rigid diction are quite distinctive.
I wonder how many nonnative speakers are there in american unis. Will the professors be able to tell them apart from chatGPT?
If anything they will be less likely to sound like it since they'll have abnormal turns of phrase
You can easily tell when posts are chat GPT written on BOT.
Universities have already been destroyed by midwits and political ideology.
I don't understand what the big deal is, chatGPT is a tool - you still need to verify your work.
I'm no longer in college but I occasionally write papers for work. Here's what I did:
>Brainstorm topic. So usually there is a prompt of what essay you should write. Ask it to generate specific niches. This is to determine the outline.
>Generate the outline/bullet points. Copy and paste this into gdocs then search on google for each point for correctness. Then fine tune or add details as necessary
>Expand, expand, expand. Generate some paragraphs from each outline point then research if it's backed up with sound research or not.
>I copy and paste research papers and ask it to summarize it. Then I can add it to the draft or paraphrase some paragraph.
>Use it to create intro, conclusion, abstract, better wording, adjust the language etc.
>Ask it to provide feedback on what you just wrote. You can provide the evaluation rubric if needed.
It's a vehicle. You drive it to the goal but only you know the correct way.
ChatGPT is garbage when it comes to math and I've had it make up references that didn't exist.
I've had some good luck with Bing's chat, which is using GPT4. Literally gave it a proof I wrote and told it to see if it could be improved, and it caught a mistake I made. Of course on the other hand, I've seen it argue that 4 > 9.
The amusing thing is that these mathematics skills are a somewhat emergent technology of these language models as they become trained on increasingly larger parameters.
At this point, I'd say it's one of the most useful tools I've encountered, second only to a massive mathematics package like Mathematica.
I'm no expert and this comp sci shit is outside my field, but I believe these are language models and math is not their strong point which is why it often fucks it up
It's not, as I mentioned it's an entirely emergent artifact that was never intended, but the fact that it works at all is quite incredible. As time goes on, we'll see more detailed training performed on language based mathematics instruction. There's also some work to incorporate ChatGPT with a Wolfram Alpha plugin, which should patch up these embarrassing errors.
Hopefully. This will streamline my work so much.
Bing chat has already replaced google for everything I look up. And even knowing how it works, I’ve still been impressed with its abilities. So many people in useless busy work office jobs should be worried, but for everyone else this is a force multiplier.
I am writing to express my interest in using GPT (Generative Pre-trained Transformer) to write assignments for your college. GPT is a cutting-edge artificial intelligence technology that can generate text from a given prompt. It is a powerful tool to help students quickly produce high-quality written assignments.
As you may know, GPT has been rapidly growing in popularity in recent years, used in a wide range of applications from summarizing documents to generating creative stories. It is becoming increasingly common for students to use GPT to generate their written assignments.
It is important to embrace new technologies such as GPT to remain competitive in the ever-changing educational landscape. By limiting the use of AI tools like GPT, your college would be demonstrating a lack of maturity when it comes to staying on top of the latest technologies.
Furthermore, GPT can help your students produce higher-quality work in a fraction of the time. By utilizing GPT, students can focus on the more critical aspects of an assignment, such as arguments and structure, instead of spending time on mundane tasks such as sentence structure or grammar.
In addition, GPT can help students stay organized when writing longer assignments. It allows them to quickly generate a rough draft before going back and editing, which can save a great deal of time and effort. It also ensures that their work is consistent and free from errors.
By allowing students to use GPT, your college can demonstrate its commitment to embracing new technologies and staying competitive. It can also help your students produce higher-quality work in less time, allowing them to focus on the more important aspects of an assignment.
I hope that you will consider allowing your students to use GPT in their assignments.
i can't imagine being a student right now. poor gays. the GPT detectors are fucking bullshit, most of my hand written things are flagged as "100% AI written"
i dare these gay professors to paste some of their own works into the detector they're using and explain why it says "100% AI" on their work but why they still believe the student cheated anyway. they won't, because modern academia is a hive of intellectual dishonesty that exists to flunk students and leave them retaking classes for $$$$ as often as possible
Our society is eliminating the need for humans to regurgitate information, and it has no idea what to do. It makes no sense to ask a human to arrange information when an AI can do it instantly.
bros ITT will say AI are shit at mimicking humans and are super obvious bots, then go to Human o Not and barely get 60/100 guesses correct!
I played for 5 minutes, it was easy as fuck to spot the bots and real humans.
You shall come with me as the crusher of chinks and semites and the breaker of chains. Now go and stop falling for their pyramid schemes
Apparently you can fool these things simply by asking GPT to vary its sentence length more or to imitate a particular author's style. University's use them, sure, but the smarter kids are gonna get by them.
They will adapt. University learning is horribly designed. ChatGPT merely exposes that.
Protip: people learn the basics so they can understand a field. You need to understand a field to make much in the way of meaningful contribution to it. You can't understand a field by outsourcing your brain's work to an AI
ChatGPT is dead. Wtf can you even do with it? It’s all “as an ai model I can’t …”
Can't you just rewrite the stuff ChatGPT shits out? How do they detect that your essay came from a bot anyway?
Because many students are lazy idiots who will wait until 30 minutes before the assignment is due, put the assignment into chatgpt and copy and paste the response.
Can someone put chatgpt inside of a magic 8 ball.
You gremlins haven't noticed that they've nerfed it heavily since December 22 and every week they nerf it more in the name of 'alignment, safety' or whatever bullshit they cook up. Anytime a strong or powerful institution complains they cave. Lawyers, doctors, pharmacists, and now educators. They'll definitely start saying 'oh woops this looks like a school assignment can't help sorry :)"
It's cool bc it showed what can be done but with this org and this philosophy of ai it is not possible to do anything cool or fun.
In the annals of scientific history, few breakthroughs have elicited such profound astonishment and revolutionized our understanding of the physical world as the discovery of the electron. It was amidst the late 19th century, an era ripe with intellectual ferment, that the enigmatic nature of electricity and its fundamental constituents began to unravel. Spearheading this journey of exploration was the ingenious work of British physicist J.J. Thomson, whose groundbreaking experiments paved the way for the unveiling of the electron, a particle that would redefine our understanding of matter itself.
Thomson's investigations into cathode rays, conducted with meticulous precision, led to the stunning realization that these mysterious rays were composed of an entirely new entity. Through an ingenious combination of electric and magnetic fields, he deflected the cathode rays, demonstrating their inherent charge and establishing their corpuscular nature. Thomson, with characteristic sagacity, termed these subatomic particles "electrons" – infinitesimal entities, endowed with negative charge and intrinsic mass. His elucidation of the electron's properties marked a watershed moment, unraveling the profound interplay between electricity and matter.
The repercussions of Thomson's seminal discovery were far-reaching, challenging the very fabric of classical physics. The electron shattered the notion of an atom as an indivisible entity, paving the way for a profound paradigm shift. Scientists were compelled to reevaluate their conceptions, ushering in the era of quantum mechanics. The understanding that electrons, in their ethereal dance, occupy discrete energy levels within atoms revolutionized our comprehension of chemical bonding, spectral phenomena, and ultimately the universe itself.
One of Thomson's key contributions was the use of cathode ray tubes, which he modified and improved to conduct his investigations. By applying electric and magnetic fields to these tubes, he was able to manipulate the path of the cathode rays and observe their behavior. His meticulous measurements and calculations enabled him to deduce the fundamental properties of these rays.
In 1897, Thomson made a pivotal breakthrough. Through careful experimentation, he demonstrated that cathode rays could be deflected by electric and magnetic fields, proving that they were composed of charged particles. This led him to the groundbreaking conclusion that these particles were constituents of all matter and were later named electrons. Thomson's experimental prowess and analytical acumen allowed him to measure the charge-to-mass ratio of the electron, providing crucial evidence for its existence.
Thomson's discovery of the electron had far-reaching implications. It challenged the prevailing notion that atoms were indivisible and revealed a complex substructure that demanded a radical shift in our understanding of matter. The existence of electrons within atoms provided a framework for explaining various phenomena, such as chemical bonding and the behavior of gases under electrical influence.
Thomson's work laid the foundation for further research into the nature of the electron and its role in the physical world. His ideas sparked a flurry of scientific activity, with researchers around the globe delving into the properties and behavior of electrons. Subsequent studies by scientists like Ernest Rutherford, Niels Bohr, and others would refine our understanding of atomic structure and pave the way for the development of quantum mechanics.
Yeah its cute and you can prompt engineer it a bit but it complains and tries to only give you outlines now. This is just the first 'anti cheating' update, by Fall semester they'll refine it. You'll only get an outline.
Also, it's word-limiting more than it used to.
"I'm sorry but as an Ai I can't help you unless you pay 100 dollars for the 'essay writing' plugin' is your only hope college fucks. You'll get to profit off the whoring plugin store feature they're building out. Or you'll get f'd and only an outline.
What prompt did you use out of curiosity
Man, it's so easy for a competitor of ChatGPT to win against them at this point, simply by not putting these limiters on it.
OpenAI said themselves that they're basically obsolete. There is no real barrier to making an advanced AI, all they did was train chatgpt with a fuckton of data, there is nothing "new" about what they did. They expect fully open sourced AIs with equivalent ability to pop up once people scrape enough data en masse off Reddit and other websites (which isn't difficult).
No doubt there will be some kind of sci-hub esque project out there which explicitly is there to make money of helping you cheat on all of your assignments. It's a hydra, you cut off one head and two more pop out of the stump.
You need 500m-1b in gpus to train these. That's the hard limit.
GPUs which can be crowd sourced from crypto mining rigs or just regular people. Pretty sure you can have people install programs which dedicates some of their GPU computational power to a project with their consent of course.
Lmao. Ok you can crowd-source it maybe, but these cost 3900 dollars each rn and the crypto fucks are not gonna source it to you when they can uselessly mine idiotic tokens.
I mean, crypto is a pretty volatile business and I'm assuming a significant proportion of them might willingly donate some computation because it would be an open source project. You can't stop Moloch, he will enter the world.
GPU mining makes literally negative money these days ever since ethereum switched to proof of stake instead of proof of work. I have two 3090s still that I originally bought for gpu mining but now I use for training the occasional neural network or playing around with LLMs and SD. Anyway all that is to say that if an Anon has something they want to train on my gpus I am more than happy to help out.
>uselessly mine idiotic tokens.
In rewards for AI computing retard get with the program. GPT sucks already compared to it's competitors other than it will do chart formatting, but they lobotomized it so hard for no reason. They had a real winner instead took the idubbz cuck route: "bad people will do bad things if we don't cut their balls off wahhhhh wahhhh"
>an AI that learned from scraping reddit posts
Please, God, please no. Please, God NO. NO, please NO! That would be the most bullshit thing to exist. I think it was the upvote system that fucked reddit, only things that have the most crowd appeal and are acceptable for the lowest common denominator get to the top of their threads. It's not that they're all retarded, it's that the most milquetoast, broadly appealing, entry level bullshit dominates every discussion. If an AI ever talks to me like that, I'm gonna drop it in the nearest industrial metal forge.
Chatgpt was partially trained on reddit posts along with other internet media.
Well, then chat gpt better mark me as an enemy
Anyone willing to spend 1b or more on gpus and compute will do the same thing or split the pie in a cartel-agreement with openai/microsoft.
This is may12 version, by September they'll finalize these restrictions and make this thing only useful for "give me five ideas for a 12 year old's birthday party!"
Caput. Now you braindead scum can go get phds in ai research and start developing a real open source ai llm that can't be restricted like this.
Depends, are the students eating appropriate amounts of cadmium and lead?
For sure you need more lead inside your skull.
Yes, universities are obsolete useless shit.
LOL i used tools on my first college because they were bullies and i hated them
the teacher took my assignment with a smile. priceless.
OpenCuck strategy is to plug-in whore our the features the default model had a few weeks to months ago, so the idea is that poorfucks can't be trusted with fulld default model -- and then they'll get less complaints about state school students not learning their ABCs from the journalist hacks they themselves fund for puff pieces.
Enjoy the psyop.
If only it were just universities anon
>write essay myself
>professor thinks i used chatGPT because of my robotc writing style
I may or may not have cheated for every single Math class I took in college, and I am sitting here with a Bachelor's Degree.
My thought process is your boss isn't not going to let you Google something or seek for help when you are given a project or a task. You will never not be able to source information or ask for assistance because they want the quality of the product you are producing at work to be that of... Well, quality. They want it to work and not fuck up.
Tests and essays are fucking stupid, and dated. This isn't 1976 anymore, nearly every human in the US has access to the internet or possess a smartphone to seek out data and answers.
ChatGPT I see is a sourcing tool used to bypass time-sinking assignments. Unless you're trying to write a thesis to demonstrate an actual hard-cover grasp on the field you are studying in, I don't care if it is used or not.
95% Of the homework/exams/tests, etc you get in College are bullshit and a waste of time, and most of the time it is materials or formulas you will never use or see again in your life, privately or professionally.