Developer and refugee from Reddit

  • 1 Post
  • 113 Comments
Joined 1 year ago
cake
Cake day: July 2nd, 2023

help-circle



  • They all got greedy. All of them. We wanted a streaming service to watch our shows and movies on, and they all decided to pretend that what we really wanted was a return to paying $100+ a month for a collection of channels with content that we mostly don’t watch on them, only this time with a bunch of additional apps you have to install for each one, most of them remarkably shitty. Like cable, but stupider.

    Remember when the streaming setup was simple? There was basically just Netflix, it paid for licenses to content from Disney, Paramount, etc., and provided guaranteed income for those companies. Small income, sure, but steady.

    Then each of them said, “Hey, why don’t we replace Netflix, only all we’ll stream is our own stuff! And sure, most of it’s trash, but people will stick around for the good shows!”

    No. No, they won’t. They’ll go back to pirating it. No one is paying for Netflix, Hulu, Disney+, Paramount+, Max, Peacock, AppleTV, ESPN+, Prime, and whatever other shitty “exclusive” streaming service pops up.









  • I use Copilot in my work, and watching the ongoing freakout about LLMs has been simultaneously amusing and exhausting.

    They’re not even really AI. They’re a particularly beefed-up autocomplete. Very useful, sure. I use it to generate blocks of code in my applications more quickly than I could by hand. I estimate that when you add up the pros and cons (there are several), Copilot improves my speed by about 25%, which is great. But it has no capacity to replace me. No MBA is going to be able to do what I do using Copilot.

    As for prose, I’ve yet to read anything written by something like ChatGPT that isn’t dull and flavorless. It’s not creative. It’s not going to replace story writers any time soon. No one’s buying ebooks with ChatGPT listed as the author.


  • I didn’t find it surprising, because I do stuff like it every day (except I’m using Copilot integrated into my IDE, not ChatGPT).

    What he did was very cool. But he’s a game developer who already knows all the parts he needs and what to ask for, and he still has to do a lot of work by hand. He glossed over it quickly, but there are parts where he had to add code to specific, already-existing blocks of code in his program, and in order to do that, he had to know and understand what the current code was doing.

    And throughout the video, he had to know not only what to ask for, but how to ask for it. That takes experience and understanding.

    It’s possible that eventually, programming for a lot of people will mean expertise in interacting with large language models + lesser expertise in the actual programming language, but I don’t see that as likely to end programming as we know it. In fact, it might cause a surge in developer demand as the bar to entry lowers again, much like it did in the 2000’s. And there will always be demand for people with a deep understanding of the actual code, because they’ll be necessary for things like performance improvement, bug fixing… and writing the next generation of large language models.


  • In the early 2000’s, there was all this panic about how these newfangled languages and tools were going to obliterate the developer job market. They were too easy to use! Too simple for non-developers to pick up! Why, you could almost code in plain English now! Developers are doooooooomed!

    Instead, demand for developers shot through the roof, because the barrier to entry for developing applications had been lowered enough that adding staff developers to your employee roster became a no-brainer.

    Part of the problem is one of precision instructions. We call instructions that are comprehensive, specific, detailed, and precise enough to be turned into programs code, and ChatGPT doesn’t change that. It can only do what you tell it to do.

    Maybe someday, a large language model will be so sophisticated, you can say something like, “Write a program to do X, Y, and Z. It uses (address) for authentication, and (other address) for storing data. Here are the credentials to use for each. (Credentials). Your repo is at (address). You deploy the front-end at (address) and the back-end at (address). Your pipelines should be written using (file) as a template.” And maybe what it does with that will truly be able to replace me.

    But I genuinely doubt it. I glossed over an enormous amount of detail in that example. If I add it in, what it’ll start looking like is, well, more code.



  • In software development? Not many - and certainly not at smart companies.

    ChatGPT is a tool. It goes in the developer toolbox because it’s useful. But it doesn’t replace the developer, any more than a really good screwdriver replaces the construction worker.

    More and more, understanding how to use LLMs for software development will be a job requirement, and developers who can’t adapt to that may find themselves unemployed. But most of us will adapt to it fine.

    I have. I’m using Copilot these days. It’s great. And the chances of it replacing me are roughly 0%, because it doesn’t actually know anything about our applications, and if told to make code by someone else who doesn’t know anything about them either, it’ll make useless garbage.


  • There’s more to it than that, even. It takes a developer’s level of knowledge to even begin to tell ChatGPT to make something sensible.

    Sit an MBA down in front of a ChatGPT window and tell them to make an application. The application has to save state, it has to use the company’s OAuth login system, it has to store data in a PostgreSQL database, and it has to have granular, roles-based access control.

    Then watch the MBA struggle because they don’t understand that…

    • Saving state is going to vary depending on the front-end. Are we writing a browser application, a desktop application, or a mobile application? The MBA doesn’t know and doesn’t understand what to ask ChatGPT to do.
    • OAuth is a service running separately to the application, and requires integration steps that the MBA doesn’t know how to do, or ask ChatGPT to do. Even if they figure out what OAuth is, ChatGPT isn’t trained on their particular corporate flavor for integration.
    • They’re actually writing two different applications, a front-end and a back-end. The back-end is going to handle communication with PostgreSQL services. The MBA has no idea what any of that means, let alone know how to ask ChatGPT to produce the right code for separate front-end and back-end features.
    • RBAC is also probably a separate service, requiring separate integration steps. Neither the MBA nor ChatGPT will have any idea what those integration steps are.

    The level of knowledge and detail required to make ChatGPT produce something useful on a large scale is beyond an MBA’s skillset. They literally don’t know what they don’t know.

    I use an LLM in my job now, and it’s helpful. I can tell it to produce snippets of code for a specific purpose that I know how to describe accurately, and it’ll do it. Saves me time having to do it manually.

    But if my company ever decided it didn’t need developers anymore because ChatGPT can do it all, it would collapse inside six months, and everything would be broken due to bad pull requests from non-developers who don’t know how badly they’re fucking up. They’d have to rehire me… And I’d be asking for a lot more money to clean up after the poor MBA who’d been stuck trying to do my job.




  • For my job, I use Copilot, which is built on GPT-4, and I have zero concern that it’s going to replace me.

    It’s very useful, don’t get me wrong. It makes generating new code in applications that already exist a breeze a lot of the time (minus hallucinations and other forms of mistakes, of course). But it simply can’t create whole new applications of any complexity from scratch, and requires actual developers to check the code it does create. It doesn’t actually know what you want, it’s just auto-completing based on what its model decides you want.

    Again, it’s very good at that. But it’s not so good that you can replace a team of developers with just one… Or worse yet, with an MBA who thinks he can figure it out without paying anyone.