They don't pick what it's good at. If they did then of course they would choose money before art.The real problem with all these bullshit "A.I.s" is that they are automating the wrong things. These idiots think automating art is a worthy endeavor when automatically filing taxes is what they should be working on....
Try thinking about what having a rich life means in less mercenary terms.That's a rich man's perspective.
Why does his help?My brother is a high school history teacher. He's going 100% paper next year. If his district hadn't allowed it, he was going to quit. Says kids will spend more time trying to find an LLM that isn't blocked than it would take to just actually do the work.
One of the biggest uses is turning out to be replacing programming jobs, so those smug jerks are getting it too.The real problem with all these bullshit "A.I.s" is that they are automating the wrong things. These idiots think automating art is a worthy endeavor when automatically filing taxes is what they should be working on....
Why do I feel like you keep assuming I don't understand that perspective, or that all I care about is money? There's no reasonable way to expect to have a good life in this country if you don't have a good enough plan on where the money is going to come from, so why am I not allowed to point that out? I didn't say it was a perspective no-one could have, just that it was a perspective that only works if your immediate needs are already secure.Try thinking about what having a rich life means in less mercenary terms.
And yes: this is an attitude I can afford to have now, possibly because I chose to work four jobs while a full-time grad student and parent of a newborn when younger. It's kind of why I did those things, because I had that attitude then, too*. My priorities got out of whack from time to time, to be fair, and sticking with the topic of hard work and callouses building skills while shortcuts and cheats buy us nothing of lasting value, I tell myself the mistakes were worth making as long as I learned from them.
*Thanks, Dad!
LLMs use emdashes because human writers use emdashes. The fact that you used them yourself until very recently, but still believe they’re a giveaway for AI just makes my head hurt.Honestly I used to use them all the time back in the day because I thought they were a fancy dash, but I stopped using them after ChatGPT came around and they give away AI generated material.
Who it should really be replacing is the portion of management/admin who don't actually accomplish meaningful things. I've heard an anecdotal rumor that people at some company discovered that their management had been making strategic decisions for months based on information in reports or presentation materials generated by AI that turned out to be made up because the actual source data was never provided to the AI in the first place. It may or may not be a particularly true rumor, but if they were doing a better job at their role than AI, you'd think it would be less believable.One of the biggest uses is turning out to be replacing programming jobs, so those smug jerks are getting it too.
Also people are already using AI for filing taxes. A friend of mine was just telling me a horror story about how his tax preparer used AI and it botched his return.
I remember when assigning homework in the later years of grad school I would tell the students that I did not grade it because it was a "self-correcting problem." If you didn't do the homework, you'd do poorly on the tests. I gave regular ol' pen-and-paper tests.I dont really recall getting that many assignments in university: the majority was learning principles and then applying them in an examination. If I had a time travelling AI helper back then I’d only be able to influence about 30% of my total score. Not insignificant but it would be clearly suspect if I aced assignments but failed exams consistently. And I seem to recall that you had to get a passing grade on exams as well as overall to get that module.
one possible defense against excessive AI dependency would be to require each assignment to have a reference back to a previous assignment, I.e. “consider how this differs from the area reviewed week 3”. That is much harder to feed into a machine to write as it doesn’t know what week 3 was, unless you pay for a lot of tokens.
This reminds me of the survivors bias. I think we are looking at this from the wrong perspective indeed.It's a real problem, but I feel like it's touching on another one at the same time. There's a paragraph towards the end where the author notes that no student using a LLM actually thinks they're learning, and that they're using it as workload management.
This was true in my time as well, but it's only become worse since: students are doing triage. The old cliche goes, "Will this be on the test?" And it's an old cliche.
Yes, LLMs in education are a blight, but their use by such a large proportion of students is a symptom of the larger problem. Students, like everyone else, have to decide how to spend their time and effort to get the most out of them. Time management and prioritisation like that is a skill, and one they tend to be taught late or not at all. This while contending with authorities who usually don't have their best interests at heart, and will quite happily lie about what's important for the student to suit their own agenda, or more likely just not consider the student's needs at all.
This is not meant as a criticism of the teachers on the ground, like the author. They're the ones most likely to actually care, if that system hasn't worn them down too.
The one lesson every single student learns early is that the school as an institution is not there to help them. It's an obstacle course they have to navigate. They all want to have a good life for themselves, every single one of them, but they have to try and judge what will actually help them to achieve that, and what's only there to make them a more valuable commodity for someone else at their own expense. And since most of them are young, either very young adults or literal children, they don't have the experience to avoid making big mistakes when navigating that minefield.
They're using LLMs as a tool to try and manage the parts they've identified as unimportant or detrimental to them. Like the author, I think that, in itself, is one of those big mistakes. But I also think it's critically important to be aware of why they're making it.
I see what you did there. I, of course, am immune to infection with the Dunning Kruger virus. I reiterate the immortal words coined by my mother: "You people that think you know everything are very annoying to those of us that do". Also I am not an engineer or politician...The tendency of smart professionals whose field requires substantial investment of focus and learning to assume that expertise extends into areas they couldn't be bothered to care about in school is widespread, to the extent of infecting many of us in these forums, in fact. I don't consider myself immune to it, and do try to remind myself periodically the line between ignorance and knowing enough to be dangerous is super-duper thin and nigh invisible.
For the record, if anyone suspects I might be talking about you and your fellow [insert profession], you are probably correct—as always, of course![]()
I was seeing a lot of em dashes in this article and was laughing internally because I was thinking that you probably see paperwork with em dashes a lot of the time and shake your head, but then I was wondering if you put that in your article as a hidden joke also or do you really love to use them? Honestly I used to use them all the time back in the day because I thought they were a fancy dash, but I stopped using them after ChatGPT came around and they give away AI generated material.
IIRC it takes approximately 24 km/s of delta-v to hit the Sun from low Earth orbit. It takes about 16.5 km/s of delta-v to leave the Solar System.I get and sympathize with the sentiment, but you probably meant "make me want to launch myself into the Sun".
I've long been an advocate for launching things into Jupiter instead, as a cost saving move.IIRC it takes approximately 24 km/s of delta-v to hit the Sun from low Earth orbit. It takes about 16.5 km/s of delta-v to leave the Solar System.
Anger is no excuse for inefficient use of propellant.
There's evidence that this is not correct. The act of writing seems to enhance learning in and of itself.
https://pmc.ncbi.nlm.nih.gov/articles/PMC8641140/
https://www.scientificamerican.com/article/why-writing-by-hand-is-better-for-memory-and-learning/
For example, labor-intensive oral exams didn’t become an endangered species just because of the swelling student-to-instructor ratio. Pen and paper (or keyboard and mouse) exams make it easier for each student’s experience to be the same and remove some of the potential for bias in scoring.
Well writers use them, but when you see them excessively used in comment sections and it's an elaborate expose on the human condition with prose likening that to a skilled post doctorate student, i'm generally pretty confident that its AI. Especially since em dashes written outside of a word processor require a special ascii code as it's not on the keyboard and its extra work on a mobile device. It's a sense you get when you get used to seeing AI a lot. I never presume that something is AI just because it has em dashes in it. Precisely why i didnt call out the article or anything. It's a combination of indicators and often situationally based.LLMs use emdashes because human writers use emdashes. The fact that you used them yourself until very recently, but still believe they’re a giveaway for AI just makes my head hurt.
Yep, keep crying about how your job needs to change. I'll wait. No, time won't turn back so you can just do what you always did.
Guess it's time to create a new, good path for your field since the old assumptions don't hold. Or switch fields.
It doesn't seem impossible to monitor and watch while people at home work on "homework". Will it be perfect and free? No. But if you're trying to see when someone creates something, the obvious answer is to watch them create it. Right?
Are we finally to the point where colleges test us to figure out our existing knowledge, then they let us skip what we already know? I think that'd lead to no more "classes" where you have a cohort of students and a particular teacher managing them. Instead you'd have students working on their own to learn, and what they produce would need to be reviewed by topic knowledgeable people. With some tutors to help explain topics beyond what text or videos manage. It never made sense to me why we wanted every teacher to recreate the wheel (poorly often) under this misguided idea that kids only learn from that.
Depends on the device and the keyboard layout. Maybe Windows makes it hard, but on macOS with the US International layout it's just holding down option while you press the hyphen key. In iOS you just long-press on the hyphen. It's also not that unusual for two hyphens to be auto-corrected to an em dash.Especially since em dashes written outside of a word processor require a special ascii code as it's not on the keyboard and its extra work on a mobile device.
Written exams standardise the delivery of questions only, they do not standardise student experience because the actual process of interpreting questions from the written page is very dependent on the student. For example, if there is terminology that has a subtly different meaning to the writer and the reader, or a language barrier. This is the same reason that writing survey questions is so damn hard, but we rarely think about exams as survey instruments. Oral exams are considered better at getting access to what students understand rather than what they can regurgitate because the option to ask follow-up questions, and for students to seek clarification of questions themselves, gives more information to the examiner.Pen and paper (or keyboard and mouse) exams make it easier for each student’s experience to be the same and remove some of the potential for bias in scoring.
How about "Are chickens dinosaurs?"The problem is that for most students, anything they know sufficiently well to actually ask in-depth questions on is probably going to be something so generic and widespread that LLMs have good training data on it and thus provide good enough output to still fool these students. You can't ask them to ask deep questions on Dynamics engineering or viral propagation or something because students won't yet have the knowledge to judge the output by.
But that's not an in depth question, nor one they're going to have the in-depth knowledge about to actually know which answer is correct. Because it's a very complex thing and... there's arguments for both yes and no.How about "Are chickens dinosaurs?"
That's something that most kids today know (ask any ten-year old). However, the DuckDuckGo AI says "Chickens are not dinosaurs, but they are descendants of dinosaurs and share a close evolutionary relationship with them, particularly with the T. rex." where Google says "Yes, chickens are technically considered living dinosaurs. According to evolutionary biology and modern cladistics, all birds are avian dinosaurs, specifically descending from small, feathered theropods—the same group as Tyrannosaurus rex. Chickens are considered closely related to T. rex due to shared skeletal anatomy and genetic data."
So maybe rather than just ask one LLM, have them ask several?