Live data from Hacker News

A coder considers the waning days of the craft

newyorker.com

161–170 of 1001 posts

Re: A coder considers the waning days of the craft

#161
post #92

Earlier quoted context omitted.

If software dev is simplified to the point people working jobs like you describe are able to do it, wages will also plummet, so it’s not like their situation is going to improve.

This is a zero-sum way of thinking about the world.

It’s true though. If anyone with a high school education can be a successful “programmer”, most programming jobs will be filled by the cheapest labor.

Why would any company pay more than they need to to keep their company functioning?

Re: A coder considers the waning days of the craft

#162

> At one point, we wanted a command that would print a hundred random lines from a dictionary file. I thought about the problem for a few minutes, and, when thinking failed, tried Googling. I made some false starts using what I could gather, and while I did my thing—programming—Ben told GPT-4 what he wanted and got code that ran perfectly. I mean ... IDK, if you can't write that on your own without research, I don't…

I guess the point is that there are different levels of coders. For coders who can't write code that print a hundred random lines from a dictionary file, GPT-4 is already a significant threat.

On the other hand, even if you can write it, does it mean you should write it? How much time does it cost you to write it? How much money your company is paying you in that amount of time? Is it more economical for GPT-4 to do it?

So I was curious and decided to ask the mistral-7b-instruct, and it gave me the following,

> Here is an example of a Python script that reads a dictionary file and prints 100 random lines:

    ```python
    import random
    
    with open('dictionary.txt', 'r') as f:
        lines = f.readlines()
    
    random_lines = random.sample(lines, 100)
    
    for line in random_lines:
        print(line.strip())
    ```
> This script uses the `random` module to select a random subset of the lines in the file. The `open` function is used to open the file and read its contents into a list of strings, which are then passed to the `random.sample` function to select 100 random lines. Finally, the selected lines are printed using a for loop and the `strip` method to remove any leading or trailing whitespace.

> Note that this script assumes that the dictionary file is located in the same directory as the script and is named "dictionary.txt". You may need to modify the script to specify the correct path and filename for your dictionary file.

Re: A coder considers the waning days of the craft

#163

Am I the only one becoming less impressed by LLMs as time passes? I will admit, when Copilot first became a thing in 2021, I had my own “I’m about to become obsolete” moment. However, it’s become clear to me, both through my own experience and through research that has been conducted, that modern LLMs are fundamentally flawed and are not on the path to general intelligence. We are stuck with ancient (in AI terms) tec…

It sounds like your initial impression was an overestimate and your current impression is a correction back down from that. You could say that it's "fundamentally flawed" coming from a very high initial expectation, but you could just as easily say "this is an amazing tool" coming from the point of view that it's "worthless" as many people seem to think

Re: A coder considers the waning days of the craft

#164

> At one point, we wanted a command that would print a hundred random lines from a dictionary file. I thought about the problem for a few minutes, and, when thinking failed, tried Googling. I made some false starts using what I could gather, and while I did my thing—programming—Ben told GPT-4 what he wanted and got code that ran perfectly. I mean ... IDK, if you can't write that on your own without research, I don't…

The problem is not that I don't know the algorithm... the problem is that to implement the algorithm you need to remember what method from what class does the thing you want to do. And knowing that is very difficult, even for seemingly simple problems. That's why Google is the way to program modern programming languages. (Of course I could do the same task without any libraries, but in that case it would be considere…

This isn't a complex problem - this is something you do with built-in standard libraries in pretty much any programming language. Like I would expect anyone who claims to know even basic file I/O in their given language to be able to produce a mostly working version of this in less than 30 minutes.

Re: A coder considers the waning days of the craft

#166

> At one point, we wanted a command that would print a hundred random lines from a dictionary file. I thought about the problem for a few minutes, and, when thinking failed, tried Googling. I made some false starts using what I could gather, and while I did my thing—programming—Ben told GPT-4 what he wanted and got code that ran perfectly. I mean ... IDK, if you can't write that on your own without research, I don't…

The problem is not that I don't know the algorithm... the problem is that to implement the algorithm you need to remember what method from what class does the thing you want to do. And knowing that is very difficult, even for seemingly simple problems. That's why Google is the way to program modern programming languages. (Of course I could do the same task without any libraries, but in that case it would be considere…

if you don't know the algorithm you could easily get it subtly wrong

but the only method involved here is string.strip(), so i don't think this is an example of the problem you describe

    import random
    n = 100
    ws = [line.strip() for line in open('/usr/share/dict/words')]
    for i in range(n):
        j = random.randrange(i, len(ws))
        ws[i], ws[j] = ws[j], ws[i]
        print(ws[i])
(did i fuck that up)

and yeah probably calling random.sample would be better. and usually import random is better than writing your own rng. but not always

Re: A coder considers the waning days of the craft

#167

> At one point, we wanted a command that would print a hundred random lines from a dictionary file. I thought about the problem for a few minutes, and, when thinking failed, tried Googling. I made some false starts using what I could gather, and while I did my thing—programming—Ben told GPT-4 what he wanted and got code that ran perfectly. I mean ... IDK, if you can't write that on your own without research, I don't…

I'm the opposite. I feel like it is true. I very much will NO LONGER THINK about "simple" problems. My most recent thing is for a game I'm working on. I got a bunch of free sound FX from freesound.org, but they were AIFF format. I googled and found an online converter, then asked ChatGPT to use ffmpeg to write a script that goes through the directory converting the AIFFs to MP3s. In the past if I would have written s…

a whole script? surely it's just something like `find . -name '*.aiff' | parallel ffmpeg -i {} {}.mp3`

Re: A coder considers the waning days of the craft

#168
post #124

Earlier quoted context omitted.

> If the profits from AI (or robots) replacing Job X were distributed among the people who used to do Job X, I don't think anyone would mind. Why on Earth would you expect something so unjust and unfair? Do you expect to pay a tax to former travel agents when you buy a plane ticket online? Do you pay to descendants of calculators (as in profession — the humans who did manual calculations) every time you use a modern…

AI is trained off the intellectual output of the people who did Job X, so it seems 100% fair to me.

In 90% of cases, these people have consented to sell their intellectual output to their employers, and in remaining 9,9%, they have consented to release it under an open source license. In both cases, it's completely unfair for them to expect any additional monetary reward for any use of their code above what they have already consented to — salary in the first case and nothing in the second.

Re: A coder considers the waning days of the craft

#169
post #6

Maybe I’m in the minority. I’m definitely extremely impressed with GPT4, but coding to me was never really the point of software development. While GPT4 is incredible, it fails OFTEN. And it fails in ways that aren’t very clear. And it fails harder when there’s clearly not enough training resources on the subject matter. But even hypothetically if it was 20x better, wouldn’t that be a good thing? There’s so much of t…

I’ve never found GPT-4 capable of producing a useful solution in my niche of engineering.

When I’m stumped, it’s usually on a complex and very multi-faceted problem where the full scope doesn’t fit into the human brain very well. And for these problems, GPT will produce some borderline unworkable solutions. It’s like a jack of all trades and master of none in code. It’s knowledge seems a mile wide and an inch deep.

Granted, it could be different for junior to mid programmers.

Re: A coder considers the waning days of the craft

#170

Earlier quoted context omitted.

Also, I think we are quite a ways out from a tool being able to devise a solution to a complex high-level problem without online precedent, which is where I find the most satisfaction anyway. LLMs in particular can be a very fast, surprisingly decent (but, as you mention, very fallible) replacement for Stack Overflow, and, as such, a very good complement to a programmer's skills – seems to me like a net positive at l…

It's also where I find most of the work. There are plenty of off the shelf tools to solve all the needs of the company I work at. However, we still end up making a lot of our own stuff, because we want something that the off the shelf option doesn't do, or it can't scale to the level we need. Other times we buy two tools that can't talk to each other and need to write something to make them talk. I often hear people…

Then the CIO itself gets fired … after all, average per job life of a CIO is roughly 18 months
Post reply on HN