Live data from Hacker News

I'd rather read the prompt

claytonwramsey.com

281–290 of 880 posts

Re: I'd rather read the prompt

#281
post #221

> I should hope that the purpose of a class writing exercise is not to create an artifact of text but force the student to think; a language model produces the former, not the latter. It's been incredibly blackpilling seeing how many intelligent professionals and academics don't understand this, especially in education and academia. They see work as the mere production of output, without ever thinking about how that…

Using llm’s for papers does not mean your brain is atrophying though. There are lots of ways to challenge the mind even if you use llm’s to write some papers.

If they use LLM for writing papers, they probably use it for other things as well. I have seen so many instances of adult actually skipping the step of "whys" and "whats" and go straight to "ask the LLM and we trim backwards".

Its basically adults producing texts of slop messages to each other. It is actually atrophying.

You might be in a circle of people that wants to know "why" things work. For example, when there's a bug, we go through several processes of:

There's a bug...why does it happen? What were they thinking when they wrote this? How to prevent this from happening?

This is true even for simple bugs, but nowadays you just vibe code your away into the solution, asking the AI to fix it over and over without ever understanding how it works.

Perhaps its just the way things are. I mean who uses their head to do calculations nowadays? Who knows how to create a blurring effect in physical drawing?

Re: I'd rather read the prompt

#282
post #4

I like the author's take: it isn't a value judgement on the individual using ChatGPT (or Gemini or whichever LLM you like this week), it's that the thought that went into making the prompt is, inevitably, more interesting/original/human than the output the LLM generates afterwards. In my experiments with LLMs for writing code, I find that the code is objectively garbage if my prompt is garbage. If I don't know what I…

100% agree.

LLMs may seem like magic buy they aren't. They operate within the confines of the context they're given. The more abstract the context, the more abstract the results.

I expect to need to give a model at least as much context as a decent intern would require.

Often asking the model "what information could I provide to help you produce better code" and then providing said information leads to vastly improved responses. Claude 3.7 sonnet in Cline is fairly decent at asking for this itself in plan mode.

More and more I find that context engineering is the most important aspect of prompt engineering.

Re: I'd rather read the prompt

#283
> I should hope that the purpose of a class writing exercise is not to create an artifact of text but force the student to think

Back in HS literature class, I had to produce countless essays on a number of authors and their works. It never once occurred to me that it was anything BUT an exercise in producing a reasonably well written piece of text, recounting rote-memorized talking points.

Through-and-through, it was an exercise in memorization. You had to recall the fanciful phrases, the countless asinine professional interpretations, brief bios of the people involved, a bit of the historical and cultural context, and even insert a few verses and quotes here and there. You had to make the word count, and structure your writing properly. There was never any platform for sharing our own thoughts per se, which was sometimes acknowledged explicitly, and this was most likely because the writing was on the wall: nobody cared about these authors or their works, much less enjoyed or took interest in anything about them.

I cannot recount a single thought I memorized for these assignments back then. Passed these with flying colors most usually, but even for me, this was just pure and utter misery. Even in hindsight, the sheer notion that this was supposed to make me think about the subject matter at hand borders on laughable. It took astronomical efforts to even retain all the information required - where would I have found the power in me to go above and beyond, and meaningfully evaluate what was being "taught" to me in addition to all this? How would it have mattered (in specifically the context of the class)? Me actually understanding these topics and pondering about them deeply is completely inobservable through essay writing, which was the sole method of grading. If anything, it made me biased against doing so, as it takes a potentially infinite extra time and effort. And since there was approximately no way for our teacher to make me interested in literature either, he had no chance at achieving such lofty goals with me, if he ever actually aimed for them.

On the other side of the desk, he also had literal checklists. Pretty sure that you do too. Is that any environment for an honest exchange of thoughts? Really?

If you want to read people's original thoughts, maybe you should begin with not trying to coerce them into producing some for you on demand. But that runs contrary to the overarching goal here, so really, maybe it's the type of assignment that needs changing. Or the framework around it. But then academia is set in its ways, so really, there's likely nothing you can specifically do. You don't deserve to have to sift through copious amounts of LLM generated submissions; but the task of essay writing does, and you're now the one forced to carry this novel burden.

LLMs caught incumbent pedagogical practices with their pants down, and it's horrifying to see people still being in denial of it, desperately trying to reason and bargain their ways out of it, spurred on by the institutionally ingrained mutual-hostage scenario that is academia. *

* Naturally, I have absolutely zero formal relation to the field of pedagogy (just like the everyday practice of it in academia to my knowledge). This of course doesn't stop me from having an unreasonably self-confident idea on how to achieve what you think essay writing is supposed to achieve though, so if you want a terrible idea or two, do let me know.

Re: I'd rather read the prompt

#284
post #265

Earlier quoted context omitted.

> I should hope that the purpose of a class writing exercise is not to create an artifact of text but force the student to think I'm there for the degree. If I wanted to learn and engage with material, I could save $60,000 and do that online for free, probably more efficiently. The purpose of a class writing exercise is to get the university to give me the degree, which I cannot do by actually learning the material (…

The work this degree will credential you for is so is so disconnected from the areas of study in your degree program - presumably in the same field as the job - that the majority of the things you might learn would not be valuable? I can’t imagine this in my own life. I use concrete things and ways of thinking and working I learned in my CS degree _all the time_.

Heck my degree was in biochemistry, and now I’m a programmer, but I still feel like I am constantly using skills I developed in school. The scientific method and good test design transcend all sciences.

Re: I'd rather read the prompt

#285

Earlier quoted context omitted.

Using llm’s for papers does not mean your brain is atrophying though. There are lots of ways to challenge the mind even if you use llm’s to write some papers.

> Using llm’s for papers does not mean your brain is atrophying though. There are lots of ways to challenge the mind even if you use llm’s to write some papers. Writing is hard. Sometimes it means sitting with yourself, for hours, without any progress. Leaning on an LLM to ease through those tough moments is 100% short circuiting the learning process. To your point, maybe you're learning something else instead, like…

Lately I've been saying often the phrase "the process is the product." When you outsource the process, then the product will be fundamentally different from what you would have delivered on your own. In my own case of knowledge work, the value of the reports I write is not in the report itself (nobody ever reads them...) but rather the thinking that went into them and the hard-won wisdom and knowledge we created in our heads.

Re: I'd rather read the prompt

#286
post #221

> I should hope that the purpose of a class writing exercise is not to create an artifact of text but force the student to think; a language model produces the former, not the latter. It's been incredibly blackpilling seeing how many intelligent professionals and academics don't understand this, especially in education and academia. They see work as the mere production of output, without ever thinking about how that…

An analogy I've heard is that it's like using a forklift at the gym. The point is not to get an object from point A to point B, it's to develop skills.

Re: I'd rather read the prompt

#287
As someone who doesn't have to grade student assignments, I'd rather read the bullet points. I've always liked using bullet points even in grade school, and also used words like dwelve, so I'd definitely fail the AI sniff test if I ever went back to school.

Re: I'd rather read the prompt

#288
post #4

I like the author's take: it isn't a value judgement on the individual using ChatGPT (or Gemini or whichever LLM you like this week), it's that the thought that went into making the prompt is, inevitably, more interesting/original/human than the output the LLM generates afterwards. In my experiments with LLMs for writing code, I find that the code is objectively garbage if my prompt is garbage. If I don't know what I…

> I'd love to hear any counterpoints from folks who have used LLMs lately to get academic or creative writing done

They’re great at proofreading. They’re also good at writing conclusions and abstracts for articles, which is basically synthesising the results of the article and making it sexy (a task most scientists are hopelessly terrible at). With caveats:

- all the information needs to be in the prompt, or they will hallucinate;

- the result is not good enough to submit without some re-writing, but more than enough to get started and iterate instead of staring at a blank screen.

I want to use them to write methods sections, because that is basically the exact same information repeated in every article, but the actual sentences need to be different each time. But so far I don’t trust them to be accurate with technical details. They’re language models, they have no knowledge or understanding.

Re: I'd rather read the prompt

#289
post #221

> I should hope that the purpose of a class writing exercise is not to create an artifact of text but force the student to think; a language model produces the former, not the latter. It's been incredibly blackpilling seeing how many intelligent professionals and academics don't understand this, especially in education and academia. They see work as the mere production of output, without ever thinking about how that…

It's been obvious since ChatGPT blew up in early 2023 that educators had to rethink how they educate.

I agree that this situation that the author outlines is unsatisfactory but it's mostly the fault of the education system (and by extension the post author). With a class writing exercise like the author describes, of course the students are going to use an LLM, they would be stupid not to if their classmates are using it.

The onus should be on the educators to reframe how they teach and how they test. It's strange how the author can't see this.

Universities and schools must change how they do things with respect to AI, otherwise they are failing the students. I am aware that AI has many potential and actual problems for society but AI, if embraced correctly, also has the potential to transform the educational experience in positive ways.

Post reply on HN