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

Learning to code is still worthwhile

stevekrouse.com

Post reply on HN
Print view View original

311–320 of 326 posts

Re: Learning to code is still worthwhile

#311

Post by bloppe »

maccard wrote: post #195
palmotea wrote: post #179

Earlier quoted context omitted.

> LLMs are an abstraction just like machine code -> assembly -> C/JVM -> some lang -> LLMs? People keep trying to make that analogy, but it doesn't really work because LLMs aren't deterministic like compilers and assemblers.

They don’t need to be dererministic, only reproducible. They’re not reproducible, nor are they even reliable right now.

I don't think there's a meaningful difference between "deterministic" and "reproducible" the way you just used them. How can something be reproducible if you can't accurately predict what it will do?

Regardless, you can make them deterministic by turning the temperature down to zero. Just nobody likes doing that for whatever reason. I guess it ruins some sort of illusion people seem to like.

Re: Learning to code is still worthwhile

#312

Post by fragmede »

jplusequalt wrote: post #308
fragmede wrote: post #293

Earlier quoted context omitted.

I'm not sure what archetype my comment makes me fall under, but my point is people spend time and effort learning about the things that are important to them. The things that are important to them aren't important to you and I'm saying that being interested in how a CPU actually works doesn't make you better or smarter than them being interested in how other parts of the business work.

>spend time and effort learning about the things that are important to them But Earlier ... >Or reading Hacker News. Reading Hacker News is important to you?

it's not for you?

Re: Learning to code is still worthwhile

#313

Post by tibbar »

bcrosby95 wrote: post #157
tibbar wrote: post #130

Earlier quoted context omitted.

The non-determinism is one of the relevant features of this layer of abstraction! And one can learn to validate that the translation is being done properly. Some of the tools you have include writing extremely detailed specs, generating visualizations of the internals of the tool, or (perhaps) reading the code, though that becomes less feasible with volume. Basically it turns out that code is full of incidental detai…

Writing specs validates the translation is being done properly? Was this written by an LLM?

Writing detailed specs is an important tool to help the LLM translate the idea in your head to code. The spec is an intermediate artifact.

Re: Learning to code is still worthwhile

#315

Post by piyush24rai »

gyulai wrote: post #163

I think that, even among those who have had the ability to code since before LLMs, the skill of coding has been in atrophy, and will continue to be so over the next 10 or 20 years. Learning to code, over the same period of time, will not be worthwhile. At the end of that process, cracks in the system will start to show: Namely, we will have nothing but LLM-generated code to train future LLMs on, and codebases for pro…

[flagged]

Re: Learning to code is still worthwhile

#316

Post by txzl »

taybin wrote: post #283
bazoom42 wrote: post #204

Earlier quoted context omitted.

The point of art (like music and litterature) is the art itself. Code is a craft, it is means to an end. It can still be beautiful and impressive and creative, but it is a different thing. It is not a value judgement. Art can be bad or bland and code can be a work of genius. But the moon lander or a handmade watch are beautiful because they actually work. It can’t really be compared to music.

Thank you for understanding my comment. Code is a craft and could maybe be more compared to pottery, which usually has a use, and is primarily judged by well it fulfills its purpose. I'm not saying that there isn't "beautiful" code, and certainly not denying clever code. But the goalpost that code is just as expressive as literature or music is just too much. Show me the code that is inspired by the reflection of a s…

https://www.dwitter.net/d/35814

Re: Learning to code is still worthwhile

#317

Post by cedws »

Zak wrote: post #284
cedws wrote: post #242

Earlier quoted context omitted.

The reason LLMs seem powerful is that they can churn out the Nth variation of a CRUD app in minutes. But it’s hard to imagine they’ll ever independently develop a compiler for a truly novel programming language, for instance. They don’t have creativity. They are pattern generators.

> it’s hard to imagine they’ll ever independently develop a compiler for a truly novel programming language I did exactly that using an LLM. It may not count as independent depending on how strict you are about that, but then LLMs don't do anything independently . I wrote a small sample program and expected output, then told the LLM to write a compiler for it in C using LLVM. I subsequently told it to extend the lang…

But what was novel about the language?

Re: Learning to code is still worthwhile

#318

Post by Zak »

cedws wrote: post #317
Zak wrote: post #284

Earlier quoted context omitted.

> it’s hard to imagine they’ll ever independently develop a compiler for a truly novel programming language I did exactly that using an LLM. It may not count as independent depending on how strict you are about that, but then LLMs don't do anything independently . I wrote a small sample program and expected output, then told the LLM to write a compiler for it in C using LLVM. I subsequently told it to extend the lang…

But what was novel about the language?

At the time the LLM generated the compiler, just syntax. The semantics it generated were the subset of C that the C version of the compiler already used.

You could set the goalposts such that it wasn't novel enough to count, but for a short time I had code running in a language that nobody had ever known. Getting it from that point to a language that's ergonomic to use, teaches something about computing, or both is a longer journey, and certainly not one an LLM could take on its own.

An LLM won't come up with an interesting CRUD app on its own either. Parts of that process are pretty mechanical, but we had skeletons and templates before we had LLMs.

Re: Learning to code is still worthwhile

#319

Post by kaffekaka »

nayhel89 wrote: post #256
OhSoHumble wrote: post #8

> Code is a beautiful form of creative expression, as rich as literature or music Something I'm trying to do right now is to build something and avoid using LLMs to write any code. I still use it to consult. I'm writing a Dota2 tournament match aggregator in Elixir that takes tournament streams and chronologically orders them in a format that makes it easier to watch them sequentially since I find YouTube hard to use…

I'm reading Seneca right now and in one of his letters he said: "That focused concentration on one's work is deeply enjoyable in itself; the pleasure one has in the finished product after the work is done does not equal it. Now, the artist is enjoying the result of his art; while he was painting, he was enjoying the art itself". And basically this is the problem I see in LLMs: they robbed me of these moments of enjoy…

I sympathize completely with you, but.. they robbed you of these moments of enjoying the art of programming itself at your day job.

You are still free to enjoy it in your free time, just like I enjoy drawing or guitar playing in my free time. The loss we have experienced is the luxury of doing at work something we enjoy for its own sake.

Re: Learning to code is still worthwhile

#320

Post by david-gpu »

drTobiasFunke wrote: post #117
david-gpu wrote: post #108

Earlier quoted context omitted.

> You are assuming that the amount of work is finite, so more productive people implies fewer people are required. Is that the case though? Has it ever been? How many horse farriers have you met? How many coopers, blacksmiths, or shoemakers?

Not a fair comparison. Unlike shoemaker or coopers who do ONE thing, a SWE doesnt do just one thing.. atleast in general. I churned through basic, logo, c, c++, java, c#, python, go and now agents. The executable building on single machine died when distributed systems came in. It again evolved when cloud took over. We have been reinventing our work every few years. Every job change requires us to learn new skills. H…

You asked for counterexamples, so I gave you a few. Did I misinterpret your question?
Post reply on HN