Hacker News Reader

The reliance on the tool in front of you can constrain you.

Some tools are fads, some make you less of a thinking person.

When I faced a second wave of IDEs I figured that any IDE is a fad - there will be better IDE in near future and your boss will require you to use it.

The use of IDE can make you know less about project at hand, because you always can investigate, as opposed to internalize/remember. The difference? If you know internal relations, you can judge changes easier. You are better programmer that way and I would say a better person.

The same goes to LMs - they help you not to know/remember and/or create thoughtfully. LMs constrain you to not know fully.

And the written word helps you to not remember anything, to paraphrase Socrates, so better get to memorizing the Odyssey so you can be an even better person.

I got better things to burn my neurons on than memorization. I normally forget almost everything, only remember abstract stuff. Maybe it's because I have used web search for 25 years that I learned to remember the good search keywords rather than actual information. Is this like relying on maps for navigation instead of pure memory?

London taxi drivers that need to remember approximately 35 thousands streets and their more or less recent state, have two changes in their brains: they have a dedicated brain area that helps with navigation and their hippocampus is bigger and more dense than hippocampus of regular people of their age and comparable lifestyle.

Hippocampus is the brain's system to transfer knowledge from short term memory to long term memory.

If you do not have enough physical activity, it shrinks with age.

In other words, London taxi drivers are slightly more young brain-wise than people of their age.

Trying to have your memory fresh and live is good for your health.

> The use of IDE can make you know less about project at hand

i disagree.

Someone who's not interested in the project might know less (but is still able to accomplish their task), where as without such an IDE, they will have to learn a lot first. It's one reason why beginner java courses at university asks you not to use an IDE (at least for the first few tutorials learning the basics).

But this isn't the fault of the IDE at all, and simply just reflect the motivations of the person using it.

An IDE makes someone motivated to learn be more capable. It's a multiplier/enabler.

> less of a thinking person.

no, it makes you think less unimportant stuff, and thus have more to dedicate to the important stuff. Effortless navigation in an IDE (like intellij) means you can move through the call graph of an application without having to see/understand the entire application completely.

The same can be said for LLMs - may be not today's, since it's still early, but at some point these LLM tools will make trivial/unimportant things easy, leaving time for you to focus on the things the LLM can't do.

> You are better programmer that way and I would say a better person.

You think not using IDEs makes you a better person?

IDEs add overhead, thus burn CPU cycles, thus increase your CO2 footprint. Not using IDEs is the only moral choice.

So true. The only moral dev tool is a HEX editor.

Or, for those less capable, also add a C compiler, as any newer language is guaranteed to have either runtime overhead or wasteful compile times.

If an IDE makes you more productive, you can spend the extra time optimizing your code to reduce CO2, so it depends.

The only moral choice is to turn the computer off.

"Real programmers use COPY CON PROGRAM.EXE"

Can you give more details regarding what IDE features you have in mind? I agree that java/c# people not knowing how to use the command line is bad (because they only know how to click "build project" etc) but IDE features such as LSP, compiler error messages as you type, intelligent completion, variable renaming, etc I think are essential tools.

I think that LSP is a Language Server Protocol. I do not use that. I can understand its utility but I do not use it.

"Compiler error messages as I type" distract me from, well, typing. I can pause to think and then I get an error message for my still incomplete code. I cannot help but notice that and that interferes with the train of thought.

I have trouble with my keyboard lately (two keys do not produce reliable key press events) and only then I started to use CtrlN in vim to autocomplete.

The need of variable renaming is often related to variables in most mainstream languages having multiple roles due to side effects: on input an array has one meaning, after execution of some function the same array now has different meaning. I prefer not to do so and greatly prefer languages with controlled side effects.

OK. For reference, I was an Emacs user for 20 years, and participated heavily in emacs extension development. I used to think like you early on. But, especially if you use languages with expressive/complex type systems, having the editor be able to inform you about types is very useful. So I understand where you're coming from, but you're probably being too absolutist and are upholding preferences that you should re-examine. For example, it's basically silly to write Rust without using rust-analyzer. I agree that feedback can be a distraction while typing but it should be possible to tune your IDE editor to not display anything until a specified amount of time has passed with no keypresses.

I do not write Rust. My day-to-day job is C++'s application support and, frankly, what Rust offers does not help there.

When I do programming for fun, I prefer to use Haskell into which "Rust" can be embedded to, if I need that. But, the problems I am interested to look deep into do not benefit from borrow checker's checks. Sparse non-negative matrix language modeling is not a thing Rust can help with. ;)

(any continuous problems like ML do not benefit from borrow checker)

And I do not use emacs. I use vim. As an EDITOR for midnight commander. I go to ghci when I need feedback, not earlier. ;)

> You are better programmer that way and I would say a better person

Lost you here. A better person by using old school IDEs? What are you smoking

A better person used LLM and whatever tool he gets at disposal to get job done ASAP and then goes outside to enjoy life with fellow humans

What you do with LLM will affect other persons, especially if the result is code. The cumulative time others spent reading your code will be much longer than you writing it.

You go hang up with your five-ten friends and dozens of other men and women will read your code meanwhile.

LLMs do not help you not to produce smaller code, I see no such goals being pursued.

That's why I make an effort to report anyone whose code I have to read to the police. They are immoral people. All codebases should be a single character.

What is a fad to you?

To me a fad is a topic of general interest, time-boxed to a short period, and is something frivolous: pet rock, hula hoop, sneakers with wheels.

I think we have different definitions.

>I am turned off whenever there is a handy new tool for us to use and then some people come up with reasons why using the new tool is a harmful thing to do.

Same, but it turns out its complicated. There is something materially different learning French and using your phone to translate French. Or learning to play piano vs listening to a recording of someone playing the piano. The danger is that these tools give make you feel that you can do something when you can't.

> For some tasks LLMs are very useful and for other tasks they don’t offer much help

Yes, and the difference is that I don't care about bash so I let it write scripts for me, but I DO care about other languages so I generally write my own. I'm basically curating my own skillset according to value and capacity. In that way AI can be a big win for teams of approx 1 who only need pro forma help in certain areas, but deep expertise in others.

> The danger is that these tools give make you feel that you can do something when you can't.

i dont think people are so stupid as to feel this way. It didn't happen with cars - you could travel hundreds of miles in less than an hour with it, where as without it, you can barely do 5. And yet people have never made the mistake that they could travel that fast themselves.

LLMs as a tool is not more harmful than what could be done with existing tools today prior to the advent of LLMs. And the benefits of the LLMs should outweight the harm imho.

> i dont think people are so stupid as to feel this way. It didn't happen with cars - you could travel hundreds of miles in less than an hour with it, where as without it, you can barely do 5. And yet people have never made the mistake that they could travel that fast themselves.

But cars are in a totally different category insofar as they do one very specific thing and are bound by very obvious physical limits. Our own physical limitations are very obvious to us, our intellectual limitations, I think, are less so. Further, companies (I think knowingly) market AI (read: llm chatbots) as though it's some general, infallible solution to all intellectual and digital content problems. This is obviously not the case, and if you use a chatbot to work in a domain in which you have any expertise whatsoever, you'll quickly uncover this. The problem is people without expertise using these tools and being fooled into thinking they've plumbed the depths when they haven't tread past the shallows. Physically speaking, your capabilities are very apparent when you attempt to run 100mph and simply *cannot*. Realizing you cannot do something or don't understand something correctly intellectually is a far different scenario.

>but I strongly disagree that it is a bad thing for individuals to sometimes save time and sometimes get better results

you shouldn't really use tools to save time or 'sometimes' get better results, because that's an indication of a sort of laziness. When you use powerful tools you need to understand exactly in what way they save you time or effort, what their failure modes are, otherwise they become pathways for exploitation.

The recent backdoor is a good example of this. People pile automation and tool chains in software on top of each other but don't bother to actually verify what comes through them, and so rather than leverage for them they became leverage for someone else. If you use LLMs to process information faster but someone feeds you garbage you're now the fastest consumer of garbage. There's nothing worse than a gullible user with a powerful tool.

And given that the general attitude we have nowadays is that tools are a sort of thing that allow you to turn your brain off, rather than understanding that you need to be smarter and more critical and more alert the more tools you use, it's very reasonable to be extremely conservative about adopting new tools.

I'm not sure exactly what you're disagreeing with. There are plenty of examples out there of individuals using these tools badly. E.g., the teacher who used ChatGPT to generate student reviews. The students who use them as high-quality plagiarism engines. Individual marketers using the saved time to spam many more people. People who want internet points generating forum comments and Stack Overflow answers. Random kooks generating disinformation, propaganda, and hate. Et cetera, ad nauseam.

Given that there are plenty of examples of people causing harm, I'm not sure how you can disagree with the notion that people can in fact use these tools harmfully. If a tools solves my problem by creating a problem for you, why shouldn't people point that out?

> using these tools badly.

i assume you mean unethical usages rather than bad.

You believe it causes harm, but who gets to decide what is ethical? I certainly do not delegate my judgement to you - only I can judge what i do myself.

Therefore, the only metric to judge is through the lens of legality (because everyone "agrees" what legal is).

> Therefore, the only metric to judge is through the lens of legality (because everyone "agrees" what legal is).

I just want to make sure we're all on the same page here: You believe that prior to Lawrence v. Texas, being a gay man in many of the US states was unethical? That prior to 1865 chattel slavery was ethical?

I often feel like HN has devolved into high-school level science fiction and philosophy, but this is rather on the nose.

> That prior to 1865 chattel slavery was ethical?

some people in that era did believe it was ethical, and some didn't. After a while, more and more people were convinced that it was unethical (and conveniently, machinary made slavery less economical for some industries). Once enough people believed/changed their minds about it being ethical, there was a large argument that resulted in laws rather than on individual ethical beliefs.

Legal is what we collectively decide. Ethical is what we personally decide.

Clearly something is either legal or not, and we have mechanisms setup to clarify where there is dispute.

Some professions have ethical standards. Some even have enforcement of them, making them pseudo-legalities.

But personal ethics, are, well, personal. You may find it ethical to eat meat, I might not. You may consider it unethical to eat cheese, or coffee, or honey. You might consider it ethical to pay minimum wage, or require employees to answer emails after hours. Or rent out an apartment, or wear white in October.

Ethics are personal and vary enormously. Declaring things to be ethical or not is therefore somewhat unhelpful. Clearly some will agree, some will not.

Whole college courses exist to try and define ethics and to encourage students to at least think about what their own ethical boundaries might be.

> Legal is what we collectively decide.

It's not even that. The collective didn't ask for sodomy laws. The collective didn't repeal Roe v Wade.

enough did that it passed, because the USA's system of first-past-the-post voting has such a loophole.

Sure, if you define "collective of american voters" as "5 unelected people", then sure "enough did".

This is both incredibly historic and clumsily sidesteps the fact that were you in that time, according to your statement, you would have found it ethical.

> you would have found it ethical.

not having lived that era, it cannot be said which side i would've been on - because it would've depended on which side i was born, and the circumstances i find myself in. The truth is that which ever "ethical" belief that benefits me the most will have been the belief i followed.

Most people believe that their beliefs are ethical, but are in reality, just post hoc rationalizations of decisions that benefit themselves (knowingly or not). I am just being clear, rather than try to hide this fact.

> the only metric to judge [what is ethical] is through the lens of legality

You've told us exactly what "side you would have been on".

Yeah, HN has a strong vein of Dunning Kruger running through it. Anybody taken in by the waffle about the law being the only metric to judge actions has never thought about where laws come from. Or read MLK's Letter from a Birmingham Jail, which clearly and passionately discusses the topic.

Ya, mean, what happened with the invent of the calculator? Sure maybe basic math skills declined…but i think i grasp relativity more than my father does and ever will…

> I am turned off whenever there is a handy new tool for us to use and then some people come up with reasons why using the new tool is a harmful thing to do.

So we shouldn't investigate (theoretically and practically) potential harmful effects of a new effects? Should we treat with coldness (dare I say disgust) anyone who dares?

You know, there is a big hole in the ozone layer that is now closing (or at least no expanding over the whole planet) because some poor researchers went against the grain and investigated all effects of this new invention called CFCs.

>We humans evolved to be very flexible and naturally use what is in front of us to solve whatever problems we need to solve.

Ha, yeah right. Almost all of us spent the vast majority of our time and energy just making dishonest rich people richer. In a world so full of exploitation, it should be no surprise that so many people are distrustful.