(Go stick your head in a pig!)
Come to think of it, “share and enjoy” is exactly the way I would expect an AI-generated YouTube video to end.
(Go stick your head in a pig!)
Come to think of it, “share and enjoy” is exactly the way I would expect an AI-generated YouTube video to end.
As a college student, yeah, ain’t nobody trying to avoid AI, lol. We ALL use that shit every single day
As a professor, we know.
Professors are most definitely also included in the ‘all’
Using it to your own detriment. Fucking idiots.
How do you figure? Please extrapolate for me.
You’re paying tens of thousands of dollars to learn how to think critically and do hard shit with your brain. Instead of actually putting in the work, you’re letting an AI do the hard parts. Which defeats the purpose of going/paying in the first place.
I’m sure I would have done the same back then too. But it is short sighted.
I asked my math teacher when I would ever use [whatever we were learning at the time] and his answer stuck with me.
“It’s not about learning how to [thing], it’s about learning to solve a problem in a new way”
I regret not taking it to heart in school but I try to remind myself of that when I “waste” a night working on something “useless” - you learn a lot more than the solution while solving a problem
It hadn’t struck me until now that some think that all you’re trying to learn in school is the solutions to arbitrary problems, but people thinking that makes a lot of sense and helps clarify why they have the posture they do towards education.
Some aspects of education overemphasize memorization so maybe a lot of people think that applies to all of education when it does not.
Yup. Unlike a calculator which just makes doing the things you already understand the steps of doing easier, AI just gives answers whether they are right or not.
I remember a teacher telling me the same thing about calculators way back when. “If you’re not able to do the calculation yourself you won’t know if the calculator’s answer is right or not”
The difference being that if you put in the right equation the calculator will give you the correct result.
AI might give you the right result.
Almost everyone who goes to college is paying tens of thousands of dollars to get a piece of paper so they can get a better paying job.
And considering how many people leave college completely inept I’d say they’re not doing the hard work of learning.
AI is ultimately just a tool, and whether it’s beneficial or detrimental depends on how you use it.
I’ve seen some of my classmates use it to just generate an answer which they copy and paste into their work, and yeah, it does suck.
I use it to summarize texts that I know I won’t have time to read until the next class, create revision questions based on my notes, to check my grammar or rephrase things I wrote, and sometimes I use Perplexity to quickly search for some information without having to rely on Google, or having to click through several pages.
Truly it isn’t much different from what we used to do around 2000-2015, which was to just Google things and mainly use Wikipedia as a source. You can just copy and paste the first results you find, or whatever information is on Wikipedia without absorbing it, or you can use them to truly research and understand something. Lazy students have always been around and will continue to be around.
It’s bad at that, because effective summarization requires an understanding of the whole, which AI doesn’t have.
The difference between what you’re doing and what people were doing 10 years ago is that what they were doing was referencing text written by people with an understanding of the subject, who made specific choices on what information was important to convey, while AI is just glorified text prediction.
LLMs can learn skills beyond what’s expected, but of course that depends on the exact model, training data and training time (See concepts like ‘emergence’ and ‘grokking’).
Currently the models tested in the study you mention (Llama-2 & Mistral) are already pretty outdated compared to other LLMs that lead the rankings. Indeed, research looking at the summarization capabilities of other models suggests that human evaluators rate them equal to or even better than human summarizers.
Well that’s a different argument from the first commenter, but to answer your point: The key here is trust.
When I use an AI to summarize text, reword something or write code, I trust that it’ll do a decent job at that - which is indeed not always the case. There were times when I didn’t like how it wrote something, so I just did it myself, and I don’t use AI when researching or writing something that is more meaningful or important to me. This is why I don’t use AI in the same way as some of my classmates, and the same is true for how I use Wikipedia.
When using Wikipedia we trust that the contributors who wrote the information on the page didn’t just nitpick their sources and are accurately summarizing and representing said sources, which sometimes is just not the case. Even when not being infiltrated by bad actors, humans are just flawed and biased so the information on Wikipedia can be slanted or out of date - and this is not even getting into how the sources themselves are often also biased.
It’s completely fair to say that AI can’t always be trusted - again, I’m certainly not always satisfied with it myself - but the same has always been true of other types of technology and humans themselves. This is why I think that even in their current, arguably still developmental stage, LLMs aren’t more harmful than technological changes in information we’ve seen in the past.
The second study link you gave didn’t find that it was ‘better’ than human writers, it concluded that if you do a lot of fine tuning then it can summarize news stories in a way that six people (marginally better than n=1 anecdote in the first link, I guess?) rated on par with Amazon mturk freelance writers. And they also noted that this preference for how the LLM summarized was individual, as in blind tests some of them still just disliked it. There are leagues and leagues of room between that and “summarizes better than humans.”
You and I both know that 99.9% of people are not fine tuning LLMs that way when they ask for a summary, which means almost nobody is going to be getting that ‘kind of as good as a person’s summary maybe if you like that style of summarizing.’ They’re getting the predictive text slop. Like, good for you if you aren’t, but maybe you should be a bit more upfront about how little you trust it and how much work you have to do to get it to give you an accurate (maybe?) summary?
My problem with LLMs is that it is fundamentally magic-brained to trust something without the power to reason to evaluate whether or not it’s feeding you absolute horseshit. With a human being editing Wikipedia, you trust the community of other volunteers who are knowledgeable in their field to notice if someone puts something insanely wrong in a Wikipedia article. An LLM will tell you anything and phrase it with enough confidence that someone with no expertise on a subject won’t know the difference.
There are several problems with these arguments.
It concluded quite the opposite actually. From the “Instruction Tuned Models Have Strong Summarization Ability.” section:
You might be confusing instruction-tuning with fine-tuning for text summarization. Instruction tuning involves rewarding a model based on the helpfulness of its responses in a user-assistant setting, and it’s the industry standard ever since the first ChatGPT showed its effectiveness.
Also they actually recruited thirty evaluators from MTurk, and six writers from Upwork (See “Human Evaluation Protocol” and “Writer Recruitment”).
Their conclusions are also consistent with the study you linked to, since which they fine-tuned the Mistral and Llama models in an attempt to generate better summaries but the evaluators still rated them lower to the human summaries. Though I’m not sure that you will be convinced by this study either since, as they state in the “PHASE 3 – FINAL ASSESSMENT” section:
Even putting all of this aside, you can actually use custom ChatGPTs that have been fine-tuned specifically to write summaries and test them for yourself if you want:
The exact same thing can be said about the lower scores in the study you linked to, so what is the exact threshold? Would you only trust an AI to summarize things if 100% of humans liked it? Besides, even if you think the best model in the study was still not good enough, there are other, even better models that have been published since then, like the ones at the top of the aforementioned leaderboards, and others like GPT-4o, OpenAI o1 and OpenAI o3.
That’s why I linked to the first article where they specifically asked an actual lawyer to evaluate summaries of legal texts written by LLMs and interns - and as we can see he thought the AI was better.
Whenever I’ve gotten into debates about the philosophy of AI and its relation to things like art, reason and consciousness, the arguments I’ve seen always end up being rather inconsistent and condescending, so I’m not even going to get into that. However I will point out that if we take the general definition of reasoning to mean “drawing logical conclusions through inference and extrapolations based on evidence”, the Wikipedia pages on LLMs, OpenAI o3 and “commonsense reasoning” explicitly describe AIs as reasoning. You’re welcome to disagree with this assessment, but if you do I hope we can then agree that, as I stated previously, Wikipedia contributors and their sources aren’t always reliable.
But sure, let’s put that aside and assume that reasoning is a magical aspect of the human brain that inherently excludes AI, so LLMs simply can’t reason… So what?
AlphaFold can’t reason, but it still can predict the structure of proteins better than humans, so it would be naive to not use it simply because it doesn’t reason. In the same vein, even if you want to conclude that LLMs can’t reason this doesn’t change the fact that they are useful tools, and perform either equal to or even better than humans in many tasks, including summarizing text.
LLMs are, for all intents and purposes, just really complicated functions that model some data distribution we give to it. Language obviously has a predictable distribution since we don’t speak/write randomly, so given proper data and training there’s no reason to believe that an AI can’t model that even better than humans. Hell, we don’t even need to get so conceptual and broad with these arguments, we can just look at the quantitative results of these models, and assess their usefulness ourselves by simply using them.
Again, I don’t trust everything these AIs generate, there are things for which I don’t use them, and even when I do sometimes I just don’t like their answers. But I see no reason to believe that they are inherently more harmful than humans when it comes to the information they generate, or that even in their current state that they’re dangerously inaccurate. If nothing else I can just ask it to summarize a Wikipedia page for me and be confident that it’ll be accurate in doing so - though as the links I mentioned demonstrate, and as you may have come to believe after considering Wikipedia’s assessment of AI reasoning, the Wikipedia contributors and their sources aren’t 100% reliable.
We both know that humans fall for and say absolute horseshit. Heck, your comment is a good example of this, where you moved the goalpost again, failed to address or outright ignored many of my arguments, and didn’t properly engage with any of the sources cited, even your own.
If you just dislike AI on principle because this technology inherently bothers you that’s fine, you’re entitled to your opinion. But let’s not pretend that this is because of some logical or quantifiable metric that proves that AI is so dangerous or bad that it can’t be used even to help university students with some basic tasks.
You’re relying on something else for comprehension and composition. Those neural pathways that are made by reading (or listening) to information and digesting it are essentially becoming vestigial. Despite my personal feelings on AI (it has no place in the arts or to replacing voice actors), you cannot always rely on it. It’s already proven fallible for simple things and summaries of any kind are no substitute for reading, listening, or watching it yourself. Doesn’t matter if it’s Cliff’s notes, spark notes, Dead Meat Kill Count, or a garbage AI summary and essay.
Man, back in my day when we wanted to get something wrong on an assignment we had to do it ourselves.
Maybe you should considering your career could be at stake.
AI is quickly becoming an integral part of basically every career imaginable. Those that actually take the time to learn how to use it properly are going to inevitably be in a far better position than those too scared to figure it out. The real challenge is finding the balance between using AI as the tool that it is and just getting an easy answer (which, considering all the downvotes I’m getting, is probably the part yall are justifiably concerned with). We need to teach the world (ourselves) how to use AI, not avoid it, and run away like we keep doing. This cat is out of the bag and ain’t never going back.
We had a fresh CS graduate who could only function with ChatGPT. He gave up thinking completely, ChatGPT whatever-latest-model was his thing. He was always arguing with us that the AI told him to do it this way or another. He could not take input from folks with two decades plus of experience during review. He bragged that AI would replace us all in a year. He did not last two months with us - my boss cut him loose after lots of bugs and hideous refactorings. He was more of a drag on the team than any help. Don’t become that guy.
I work with Linux and was recently obligated to work with “Linux admins” from another company. One of them had apparently never used Linux before. I don’t begrudge anyone their lack of experience, but they shouldn’t be in positions that require fairly extensive experience.
Anyway, at one point they were doing a screenshare of some (very simple) code that I wrote but that I’m pretty sure they didn’t know I wrote. They were all collectively trying to figure out how the (again, very simple) script worked (it literally just changed permissions and renamed some things, IIRC). For every single line, they would copy and paste it into ChatGPT and ask what the line did. It was kind of amazing to watch.
My job for the last decade has been working with sysadmins on Linux systems. Notice I didn’t say “Linux sysadmins” because most of them aren’t. They know a few commands by rote, but anything beyond that is impossible magic. The concept of the working directory, navigating the file path, permissions, and networking are all beyond their understanding.
I call them “turtles on posts” because they couldn’t have gotten themselves in that position and are now stuck. And since this has been happening for years it’s got nothing to do with AI.
Fortunately for me, I’m probably in the lower half of my company in terms of qualifications; it’s one of the best workforces in which I’ve ever participated. It actually bothered me a lot when I started, but as the saying goes, if you’re the smartest person in the room you’re probably in the wrong room.
The underqualified staff were with another company with whom we were required to work.
That sounds super painful to work with. But also a hilarious anecdote so you got that.
Honestly, it was painful, but mainly because of the ridiculous number of meetings they forced on us. Watching them bumble through messing up their tasks was pretty entertaining.
When I code using AI I get the best results by being very specific and write a class with pseudo code for it to fill out with the missing code.
If I just ask for it to write me a class that can X I often get some simple example code directly from stackoverflow.
It’s decent at writing simple string tools etc., because that’s what is out there, the day it starts writing code from API documentation will be a big milestone.
Currently it’s just a parrot that knows Python.
This right here is the big, glaringly obvious problem with AI, especially in academics. But, it’s also exactly why this whole issue isn’t really a big deal as long as long enough people learn to use AI correctly. Those that don’t learn and fall into the trap of easy solutions and laziness will always, inevitably fail as soon as they get to the real world and must then either learn or fade into obscurity. Those that do learn how to utilize AI will find far more success and will hopefully be able to pass on their skills and knowledge. Thus, the system, given enough time, kinda corrects itself eventually. It’s just a bit dangerous until then, hence why we need to teach and learn rather than fear what’s coming.
I feel like you are so close to realising why your argument is rubbish.
AI is absolutely a good tool. But only for people who understand what they are doing already.
Its good to help you with arduous tasks but you need to be able to review what it does with knowledge and experience or you wont understand what it gets wrong.
I use it in my job to help me to write large access lists. If i give it the parameters, the addresses i want to give access to and in what ports and protocols etc it can dump a hughe ACL and i can review it and correct any errors i find.
If i didnt know how and ACL was written, didnt know the correct syntax and didnt understand where it should be placed i could very easily apply a dodgey ACL to a live network and fuck things up for everyone.
You keep saying you need to learn how to use it and then its fine.
But its not. You need to learn that its mostly dumb and you need to scrutinise everything it does.
That scrutiny is exactly what I’m getting at when I say we need to learn how to use it. AI is really powerful, but is so incredibly far from being the magic bullet that people think it is. It is just a tool that needs to be applied carefully and responsibly, of course only the people that understand what they are doing are going to succeed. My argument is that we need to be building that understanding and sharing it as widely as possible so that even more people can use the tools properly. And, yes, that means check the fucking output, use your brain instead of replacing it.
I am reminded of an argument i had years ago about people relying on google to do their jobs.
I argued that using google to give you the answer to a problem doesn’t help you in the long run. Instead of understanding the solution and being able to use that understanding to solve problems in the future, you just become dependent on google to get you through the day.
It is much more important to learn why a solution fixes a problem and the steps you take to understand the elements of the solution. It opens more doors, and you learn how to use your brain.
Both thinking and googling will get people far, but if google ever went away, only the thinkers would survive.
This is happening again but this time its AI.
The funny thing is, the people who made google search, and the people who created AI, are likely the thinkers.
AI is quickly replacing a lot of careers.
And it will continue to do so.
I’m amazed that you think otherwise when it’s happening right now.
Also, “taking the time to learn to use it” takes all of what, a couple of days of reading at most if you want it to do something really unusual? We’re not talking about advanced coding here.
AI is not replacing much of anything, not yet anyway. It is evolving and forcing the world to evolve with it. While AI is used to write notes, summarize content, generate content, integrate data, organize life, etc., all of that still requires input of some kind from someone. Careers are going to be all about performing that input and interpreting the result. People will not be replaced (except the ones that refuse to keep up), they will just fill a different role.
You clearly understand nothing about AI if this all you think it is. Sure, anyone can type a prompt and get a garbage result in about 30 seconds, but there is a hell of a lot more to it if you want to actually solve a real problem using AI. Learning advanced coding isn’t actually a bad idea for the future.
Maybe you can understand a different perspective if you stop thinking of AI as gimmicky solution and start thinking of it as what really is, a powerful set of tools meant to make finding the solution easier, nothing more.
https://www.cnn.com/2023/03/01/media/axel-springer-ai-job-cuts/index.html
https://www.wypr.org/2024-05-30/wall-street-journal-layoffs-continue-despite-lucrative-ai-deal-and-record-profits
https://www.ccn.com/news/technology/biggest-tech-layoffs-in-2024-2025-focus-on-ai/
Edit: Here’s one more for kicks- https://www.hollywoodreporter.com/business/business-news/buzzfeed-ai-creators-news-shut-down-1235483607/
First, we are discussing careers, not individuals. No shit people are losing jobs, but guess what, that is exactly what happens when careers evolve or new ones are created. Every. Single. Time.
Think about when precision machining was invented, when printing presses where invented, when cars where invented, when computers where invented, when the fucking internet was invented, etc. Yes, a fuck-ton of people were suddenly out of a job. But then suddenly there are also a whole bunch of brand new jobs and careers to fill. People either learn to adapt and fill those roles, or they don’t, and they get left behind. AI isn’t really any different except that it is happening right now and it’s therefore hard to see what’s to come.
That’s just how the world works. It’s sad and frustrating, I know, but being scared and hiding your head in the sand doesn’t change that fact. Learning how to live and thrive with the new stuff does, though, so maybe let’s try that instead.
Second, and this isn’t to discount everything you linked, but you understand that there is a huge bias going on here, right? People are understandably scared about the future, and the media latches onto that fear and creates articles that feed the narrative beast. But often times, the articles completely neglect to talk about the other side of the coin, which is what we are discussing here.
Okay, well maybe you see things like the death of journalism and the death of criticism and the death of voice-over acting and the death of music composition to be good things, but I don’t know that you’re in the majority there.
And you may also see the massive ecological disaster that AI is becoming is a good thing too. I certainly do not.
https://en.wikipedia.org/wiki/Environmental_impacts_of_artificial_intelligence
What I’m trying to get you to understand is that this isn’t the death of anything, not really. Though it can certainly feel like it, especially right now. This is just the growing pains that goes along with literally every single major advancement in human history, and we always have this same exact unproductive argument. Yes, people get hurt along the way, but that is exactly why it is our collective responsibility to learn properly and mitigate as much of the damage as possible, being scared is never the way to do that. This isn’t the doom of society. It’s simply the dawn of a new version of society.
Continue being scared and wither into obscurity, or learn to adapt and thrive with what is inevitably going to be an integral part of our lives, career or not. The choice is really yours alone, but I want to see everyone succeed, including you and anyone else that reads this garbage.
Also, the ecological impact of AI is an entirely separate discussion, and I would appreciate it if you didn’t pretend to know my stance on the matter to support your arguments. If you want to have that discussion, we can, but not like this.
As a professional developer, same. It saves me so much time. My colleagues also use it. Lemmy is a bubble just as much as (or maybe even more so than) Reddit. Mention a use for AI and you’ll end up downvoted to hell. You just said “use AI” and people jump to “this guy switched off his brain and does nothing but blindly copy-paste ChatGPT output into his assignments.”
Yeah, I’m discovering that AI is one of those no-no topics in this particular echo chamber. Disappointing really, this whole thing is a lot more fun when people actually want to talk instead of just following the crowd. It is in the name I guess, lol.