this post was submitted on 19 May 2025
1514 points (98.0% liked)
Microblog Memes
8436 readers
2951 users here now
A place to share screenshots of Microblog posts, whether from Mastodon, tumblr, ~~Twitter~~ X, KBin, Threads or elsewhere.
Created as an evolution of White People Twitter and other tweet-capture subreddits.
Rules:
- Please put at least one word relevant to the post in the post title.
- Be nice.
- No advertising, brand promotion or guerilla marketing.
- Posters are encouraged to link to the toot or tweet etc in the description of posts.
Related communities:
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Idk, I think we're back to "it depends on how you use it". Once upon a time, the same was said of the internet in general, because people could just go online and copy and paste shit and share answers and stuff, but the Internet can also just be a really great educational resource in general. I think that using LLMs in non load-bearing "trust but verify" type roles (study buddies, brainstorming, very high level information searching) is actually really useful. One of my favorite uses of ChatGPT is when I have a concept so loose that I don't even know the right question to Google, I can just kind of chat with the LLM and potentially refine a narrower, more google-able subject.
The thing is that LLM is a professional bullshitter. It is actually trained to produce text that can fool ordinary person into thinking that it was produced by a human. The facts come 2nd.
Yeah, I know. I use it for work in tech. If I encounter a novel (to me) problem and I don't even know where to start with how to attack the problem, the LLM can sometimes save me hours of googling by just describing my problem to it in a chat format, describing what I want to do, and asking if there's a commonly accepted approach or library for handling it. Sure, it sometimes hallucinate a library, but that's why I go and verify and read the docs myself instead of just blindly copying and pasting.
That last step of verifying is often being skipped and is getting HARDER to do
The hallucinations spread like wildfire on the internet. Doesn't matter what's true; just what gets clicks that encourages more apparent "citations". Another even worse fertilizer of false citations is the desire to push false narratives by power-hungry bastards
AI rabbit holes are getting too deep to verify. It really is important to keep digital hallucinations out of the academic loop, especially for things with life-and-death consequences like medical school
This is why I just use google to look for the NIH article I want, or I go straight to DynaMed or UpToDate. (The NIH does have a search function, but it's terrible meaning it's just easier to use google to find the link to the article I actually want.)
I'll just add that I've had absolutely no benefit, just time wasted, when using the most popular services such as ChatGPT, Gemini and Copilot. Yes, sometimes it gets a few things right, mostly things that are REALLY easy and quick to find even when using a more limited search engine such as Mojeek. Most of the time these services will either spit out blatant lies or outdated info. That's one side of the issue with these services, and I won't even get into misinformation injected by their companies. The other main issue I find for research is that you can't get a broader, let alone precise picture about anything without searching for information yourself, filtering the sources yourself and learning and building better criteria yourself, through trial and error. Oftentimes it's good info that you weren't initially searching for what makes your time well spent and it's always better to have 10 people contrast information they've gathered from websites and libraries based on their preferences and concerns than 10 people doing the same thing with information they were served by an AI with minimal input and even less oversight. Better to train a light LLM model (or setup any other kind of automation that performs even better) with custom parameters at your home or office to do very specific tasks that are truly useful, reliable and time saving than trusting and feeding sloppy machines from sloppy companies.
I don’t trust LLMs for anything based on facts or complex reasoning. I’m a lawyer and any time I try asking an LLM a legal question, I get an answer ranging from “technically wrong/incomplete, but I can see how you got there” to “absolute fabrication.”
I actually think the best current use for LLMs is for itinerary planning and organizing thoughts. They’re pretty good at creating coherent, logical schedules based on sets of simple criteria as well as making communications more succinct (although still not perfect).
Sadly, the best use case for LLM is to pretend to be a human on social media and influence their opinion.
Musk accidentally showed that's what they are actually using AI for, by having Grok inject disinformation about South Africa.
The only substantial uses i have for it are occasional blurbs of R code for charts, rewording a sentence, or finding a precise word when I can't think of it
It's decent at summarizing large blocks of text and pretty good for rewording things in a diplomatic/safe way. I used it the other day for work when I had to write a "staff appreciation" blurb and I couldn't come up with a reasonable way to take my 4 sentences of aggressively pro-union rhetoric and turn it into one sentence that comes off pro-union but not anti-capitalist (edit: it still needed a editing pass-through to put it in my own voice and add some details, but it definitely got me close to what I needed)
I'd say it's good at things you don't need to be good
For assignments I'm consciously half-assing, or readings i don't have the time to thoroughly examine, sure, it's perfect
exactly. For writing emails that will likely never be read by anyone in more than a cursory scan, for example. When I'm composing text, I can't turn off my fixation on finding the perfect wording, even when I know intellectually that "good enough is good enough." And "it's not great, but it gets the message across" is about the only strength of ChatGPT at this point.
Can you try again using an LLM search engine like perplexity.ai?
Then just click on the link next to the information so you can validate where they got that info from?
LLMs aren't to be trusted, but that was never the point of them.
I have two friends that work in tech, and I keep trying to tell them this. And they use it solely now: it’s both their google, and their research tool. I admit, at first I found it useful, until it kept being wrong. Either it doesn’t know the better/best way to do something that is common knowledge to a 15 year tech, while confidently presenting mediocre or incorrect steps. Or it makes up steps, menus, or dialog boxes that have never existed, or are from another system.
I only trust it for writing pattern tasks: example, take this stream of conscious writing and structure it by X. But for information. Unless I’m manually feeding it attachments to find patterns in my good data— no way.
So use things like perplexity.ai, which adds links to the web page where they got the information from right next to the information.
So you can check yourself after an LLM made a bullshit summary.
Trust but verify
So are people. Rule NUMBER 1 when the internet was first picking up is "Don't believe everything you read on the internet". it's like all of you have forgotten. So many want to bitch so hard about Ai while completely ignoring the environment it was raised in and the PEOPLE who trained it. You know, all of us. This is a human issue not an AI issue.
To be fair, facts come second to many humans as well, so I dont know if you have much of a point there...
That's true, but they're also pretty good at verifying stuff as an independent task too.
You can give them a "fact" and say "is this true, misleading or false" and it'll do a good job. ChatGPT 4.0 in particular is excellent at this.
Basically whenever I use it to generate anything factual, I then put the output back into a separate chat instance and ask it to verify each sentence (I ask it to put tags around each sentence so the misleading and false ones are coloured orange and red).
It's a two-pass solution, but it makes it a lot more reliable.
So your technique to "make it a lot more reliable" is to ask an LLM a question, then run the LLM's answer through an equally unreliable LLM to "verify" the answer?
We're so doomed.
Give it a try.
The key is in the different prompts. I don't think I should really have to explain this, but different prompts produce different results.
Ask it to create something, it creates something.
Ask it to check something, it checks something.
Is it flawless? No. But it's pretty reliable.
It's literally free to try it now, using ChatGPT.
Hey, maybe you do.
But I'm not arguing anything contentious here. Everything I've said is easily testable and verifiable.
And just as back then, the problem is not with people using something to actually learn and deepen their understanding. It is with people blatantly cheating and knowing nothing because they don’t even read the thing they’re copying down.
Something I think you neglect in this comment is that yes, you're using LLMs in a responsible way. However, this doesn't translate well to school. The objective of homework isn't just to reproduce the correct answer. It isn't even to reproduce the steps to the correct answer. It's for you to learn the steps to the correct answer (and possibly the correct answer itself), and the reproduction of those steps is a "proof" to your teacher/professor that you put in the effort to do so. This way you have the foundation to learn other things as they come up in life.
For instance, if I'm in a class learning to read latitude and longitude, the teacher can give me an assignment to find
64° 8′ 55.03″ N, 21° 56′ 8.99″ Won the map and write where it is. If I want, I can just copy-paste that into OpenStreetMap right now and see what horrors await, but to actually learn, I need to manually track down where that is on the map. Because I learned to use latitude and longitude as a kid, I can verify what the computer is telling me, and I can imagine in my head roughly where that coordinate is without a map in front of me.Learning without cheating lets you develop a good understanding of what you: 1) need to memorize, 2) don't need to memorize because you can reproduce it from other things you know, and 3) should just rely on an outside reference work for whenever you need it.
There's nuance to this, of course. Say, for example, that you cheat to find an answer because you just don't understand the problem, but afterward, you set aside the time to figure out how that answer came about so you can reproduce it yourself. That's still, in my opinion, a robust way to learn. But that kind of learning also requires very strict discipline.
So, I'd point back to my comment and say that the problem really lies with how it's being used. For example, everyone's been in a position where the professor or textbook doesn't seem to do a good job explaining a concept. Sometimes, an LLM can be helpful in rephrasing or breaking down concepts; a good example is that I've used ChatGPT to explain the very low level how of how greenhouse gasses trap heat and raise global mean temperatures to climate skeptics I know without just dumping academic studies in their lap.
Your example at the end is pretty much the only way I use it to learn. Even then, it's not the best at getting the right answer. The best thing you can do is ask it how to handle a problem you know the answer to, then learn the process of getting to that answer. Finally, you can try a different problem and see if your answer matches with the LLM. Ideally, you can verify the LLM's answer.
To add to this, how you evaluate the students matters as well. If the evaluation can be too easily bypassed by making ChatGPT do it, I would suggest changing the evaluation method.
Imo a good method, although demanding for the tutor, is oral examination (maybe in combination with a written part). It allows you to verify that the student knows the stuff and understood the material. This worked well in my studies (a science degree), not so sure if it works for all degrees?
Yeah it depends. Raw dogging chatgpt is always no
I might add that a lot of the college experience (particularly pre-med and early med school) is less about education than a kind of academic hazing. Students assigned enormous amounts of debt, crushing volumes of work, and put into pools of students beyond which only X% of the class can move forward on any terms (because the higher tier classes don't have the academic staff / resources to train a full freshman class of aspiring doctors).
When you put a large group of people in a high stakes, high work, high competition environment, some number of people are going to be inclined to cut corners. Weeding out people who "cheat" seems premature if you haven't addressed the large incentives to cheat, first.
Medical school has to have a higher standard and any amount of cheating will get you expelled from most medical schools. Some of my classmates tried to use Chat GPT to summarize things to study faster, and it just meant that they got things wrong because they firmly believed the hallucinations and bullshit. There's a reason you have to take the MCAT to be eligible to apply for medical school, 2 board exams to graduate medical school, and a 3rd board exam after your first year of residency. And there's also board exams at the end of residency for your specialty.
The exams will weed out the cheaters eventually, and usually before they get to the point of seeing patients unsupervised, but if they cheat in the classes graded on a curve, they're stealing a seat from someone who might have earned it fairly. In the weed-out class example you gave, if there were 3 cheaters in the top half, that means students 51, 52, and 53 are wrongly denied the chance to progress.
Having a "high standard" is very different from having a cut-throat advancement policy. And, as with any school policy, the investigation and prosecution of cheating varies heavily based on your social relations in the school. And when reports of cheating reach such high figures
then the problem is no longer with the individual but the educational system.
Nevermind the fact that his hasn't born itself out. Medical Malpractice rates do not appear to shift based on the number of board exams issued over time. Hell, board exams are as rife with cheating as any other academic institution.
If cheating produces a higher class rank, every student has an incentive to cheat. It isn't an issue of being seat 51 versus 50, it's an issue of competing with other cheating students, who could be anywhere in the basket of 100. This produces high rates of cheating that we see reported above.
Medical malpractice is very rarely due to gaps in knowledge and is much more likely due to accidents, miscommunication, or negligence. The board exams are not taken at the school and have very stringent anti-cheating measures. The exams are done at testing centers where they have the palm vein scanners, identity verification, and constant video surveillance throughout the test. If there is any irregularity during your exam, it will get flagged and if you are found to have cheated, you are banned from ever taking the exam again. (which also prevents you from becoming a physician)
No. There will always be incentives to cheat, but that means nothing in the presence of academic dishonesty. There is no justification.