Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

I'd rather read the prompt

claytonwramsey.com

Post reply on HN
Print view View original

811–820 of 880 posts

Re: I'd rather read the prompt

#811

Post by johnnyanmac »

lazide wrote: post #559
jjmarr wrote: post #345

Earlier quoted context omitted.

> The reason a teacher asks you a question is not because they don’t know the answer. A decent amount of my professors don't know the answers because they bought the course, test questions, and lectures from Cengage. During exam review, they just regurgitate the answer justification that Cengage provided. During the lectures, they struggle to explain certain concepts since they didn't make the slides. Professors auto…

But that requires professors to do work and give a damn!

These projections about bad professor experience is exactly why colleges became so hyperconpetitive. The difference between a competitive (and not even exclusive level like Ivies) and a "semi-competitive" is night and day in terms of rigor, staff talent, and overall environment.

But sure, you're always going to find a few meh or bad professors. And they will stick out as much ad thr great professors

Re: I'd rather read the prompt

#812

Post by tobir »

musicale wrote: post #141
jsheard wrote: post #74

I'm far from the first to make this observation but LLMs are like anti-compression algorithms when used like that, a simple idea gets expanded into a bloated mess by an LLM, then sent to someone else who runs it through another LLM to summarize it back to something approximating the original prompt. Nobody benefits aside from Sam Altman and co, who get to pocket a cool $0.000000001 for enabling this pointless exercis…

> LLMs are like anti-compression algorithms when used like that, a simple idea gets expanded into a bloated mess by an LLM, I think that's the answer: LLMs are primarily useful for data and text translation and reduction, not for expansion. An exception is repetitive or boilerplate text or code where a verbose format is required to express a small amount of information.

Yeah, this is the usual interaction with LLMs when coding: Ask GPT/Claude to write me a simple function. It writes 100 lines of code, trying to infer the rest of your codebase. Tell it to only solve my issue and do it in 5 lines of code. Get something close to working.

Re: I'd rather read the prompt

#813

Post by johnnyanmac »

dyauspitr wrote: post #746
lazide wrote: post #560

Earlier quoted context omitted.

Sure there is - evaluate the candidates actual skill level without the ability to use automated tools. The current and old school way is a proctored exam.

Testing doesn’t work for everything, especially for dissertations, publications, research etc. That being said I’m wholeheartedly in favor of using extensive in class testing when applicable.

It works for enough that 99% of cheaters are weeded out. Perfection is the enemy of "good enough".

Also, always keep in mind the most underrated kind of cheater: the clearly smart kid who cheats to go from an A- to an A. That level of student can't be "caught" with tests because by all accounts they already know and have even mastered the material. The pressure at that level of competitiveness simply requires zero room for doubt. Not really what the article is about, but some food for thought.

Re: I'd rather read the prompt

#814

Post by godelski »

bumby wrote: post #774
godelski wrote: post #769

Earlier quoted context omitted.

I was responding to the comment you wrote. But I also disagree with the other one. It is a confounding variable[0]. The problem with trying to go after the confounding variable is you 1) don't solve the problem after fixing it 2) let's the current negative feedback loop continue growing. [0] https://en.wikipedia.org/wiki/Confounding

Nobody ever claimed the best fix is going after the proximate cause and not touching the root cause. That’s a narrative you contrived. I’m in favor of fixing both (which is how systemic improvements ought to work).

Look back at the thread:

  - ebiester claimed academia made itself the gatekeeper of credentials
  - alpinisme suggested that was silly. Employers have autonomy
  - you asked why MOOC classes don't get the same credentials
  - I said you missed alpinisme's point. The employer is the one that is deciding those credentials are not equal. I also pointed out how the same is true about different universities. Pointing out that you didn't need to say "MOOC" when "Stanford vs University of Wyoming" also leads to the same conclusion (which makes it weirder to point to academia as a whole instead of the prestigious unis)
  - you act like I was supposed to have read a reply to a different person while not addressing my point
  - I double down saying you're barking up the wrong tree. Take the finger and point it at employers
  - you say you never claimed employers aren't the problem and I'm creating a narrative
Well you're right, you never claimed that. But ebiester did and when alpinisme called bullshit you came to ebiester's defense. Forgive me if I got confused.

Regardless, it still doesn't change the fact that the employers have complete autonomy. They can make whatever rules they want. There's no use to pointing at academia because ultimately they have no say. Do they want to be the credential keepers, yeah! Which also means they'll be happy to be the ones being yelled at if the result is that employers keep using them as credential makers.

Ultimately academia is about prestige. Ultimately it's about far more than "the product" (the credential). But who decides what credential is best? The fucking employers. No one is holding a gun to their head. There's no front door or back door dealings.

You have concluded that since academia benefits from the employers selecting them as credential makers that they are the problem, or a meaningful part. The issue is, if you take away academia, the problem doesn't get resolved. Nor does it improve. Arguably, it becomes more noisy until employers converge onto some other arbitrary credential

Re: I'd rather read the prompt

#815

Post by johnnyanmac »

dep_b wrote: post #582
jameshart wrote: post #329

Earlier quoted context omitted.

I’ve come across this analogy that I think works well: Using an LLM to do schoolwork is like taking a forklift to the gym. If all we were interested in was moving the weights around, you’d be right to use a tool to help you. But we’re doing this work for the effect it will have on you . The reason a teacher asks you a question is not because they don’t know the answer.

If I could have a healthy and good looking physique by never going to the gym I would never go to the gym.

Well, you could. Most people prefer their pre-processed food over never going to the gym though.

Re: I'd rather read the prompt

#816

Post by strken »

simonask wrote: post #551
strken wrote: post #326

Earlier quoted context omitted.

Yes, I too enjoy that universities never send students to pointless filler courses. I really enjoyed writing long essays about ethics in software engineering for people who'd barely even read them. I especially benefited from being told that "tell your boss, tell HR, then make a police report and quit if that didn't fix it" is not an appropriate response to being asked to break the law at work. While we're talking ab…

Do you... not understand that the point of giving you a class on ethics in software development wasn't so you would produce an essay?

Do you...not understand that the class on ethics in software development was completely pointless because it was neutered by a university administration that was too afraid to commit to any real ethical position?

I thought I was pretty clear when I told you I got marked down for taking the stance that one should not commit crimes when asked to by one's boss.

The problem isn't that I was asked to produce an essay as part of the process of teaching me about ethics. The problem is that the entire class was pointless busywork taught by lecturers who weren't particularly ethical and who failed to teach any of us about ethics. ChatGPT could have done the entire thing and the class would still have failed to leave me with anything other than lingering resentment.

Re: I'd rather read the prompt

#817

Post by johnnyanmac »

dogcomplex wrote: post #594
jameshart wrote: post #329

Earlier quoted context omitted.

I’ve come across this analogy that I think works well: Using an LLM to do schoolwork is like taking a forklift to the gym. If all we were interested in was moving the weights around, you’d be right to use a tool to help you. But we’re doing this work for the effect it will have on you . The reason a teacher asks you a question is not because they don’t know the answer.

If you took a forklift to the gym, you'd come out of the experience not only very good at "lifting weights" , but having learned a whole lot more about the nature and physics of weightlifting from a very different angle. Sure, you should lift them yourself too. But using an AI teaches you a shit-ton more about any field than your own tired brain was going to uncover. It's a very different but powerful educational exp…

> But using an AI teaches you a shit-ton more about any field than your own tired brain was going to uncover.

If you never learn to research, sure. Otherwise, you should be worried about accuracy, up to date information, opinionated takes, and outright lies/misinformation. The tool you use doesn't change these factors.

Re: I'd rather read the prompt

#818

Post by johnnyanmac »

indymike wrote: post #681
jameshart wrote: post #329

Earlier quoted context omitted.

I’ve come across this analogy that I think works well: Using an LLM to do schoolwork is like taking a forklift to the gym. If all we were interested in was moving the weights around, you’d be right to use a tool to help you. But we’re doing this work for the effect it will have on you . The reason a teacher asks you a question is not because they don’t know the answer.

> Using an LLM to do schoolwork is like taking a forklift to the gym. I'm sure the time has come for college students to master using LLMs. It's just as important as grammar or basic math now. The software I build (and the entire tech industry) automates huge swaths of business processes with AI. Students need to be able to understand, work with, and manage swarms of AI agents doing work. To stick to the analogy: I n…

If you need forkloft drivers, don't recruit at the gym and be mad no one is forklift certified there.

This isn't even an opinion on LLMs, it's recruiting 101. You're free to convince the gym to train forklift drivers, but don't be surprised when you're laughed out the room.

Re: I'd rather read the prompt

#819

Post by necovek »

tptacek wrote: post #772
necovek wrote: post #751

Earlier quoted context omitted.

You asked: > As a team lead, what would you think of a team member who consistently generated "crap" pull requests? This is directly answered with my first paragraph: that's exactly what I would think of them, and how I would act on it. Your first question was: > How is that better? In the second paragraph, I explained why it's better to do a code review for a crappy pull request that's human-produced vs LLM-generate…

Jim and Toby are interviewing Darryl Philbin for the position of manager at Dunder Mifflin Scranton. Jim asks what Darryl would do to resolve a conflict between two employees in the warehouse Darryl already managed. "I'll answer that, Jim. I would use it as an opportunity to teach... about actions... and consequences of actions". That's the answer you just gave me. Good note! (Darryl didn't get the job.) You're dodgi…

You seem to be arguing a point I never made: I never claimed LLMs are "dangerous", nor that crappy code only comes out of LLMs. They are just a tool, and I agree it is the responsibility of the developer wielding them to produce good work with (or without) them — this was never disputed.

I don't have any issue with someone using an LLM but I have not observed any efficiency gain from those who do — that's my entire point, and the biggest selling point for using coding assistants. I've either seen them produce "crappy code" faster (which, ultimately, they could do by hand as well), or be slower than doing their work "manually".

At the same time, I disagree about teams producing lousy PRs being mismanaged by definition: there are circumstances where doing that is warranted (LLM or no LLM), as long as the long term direction is improving (less crappy code over time). There are plenty of nuances there too.

Re: I'd rather read the prompt

#820

Post by mark336 »

lmm wrote: post #489
godelski wrote: post #411

Earlier quoted context omitted.

> We are a society of ladder-pullers. I don't disagree, but often we complain about people pulling up ladders and when faced with the same decision we follow suit. Ultimately we can't change this behavior if no one is willing to defect from "conventional wisdom"

We can't fix the problem by making better choices as individuals, and exhorting people to do so saps energy and distracts. The system interprets integrity as damage and routes around it.

This is a problem with liberal thinking. Its brains over do the right thing.
Post reply on HN