alt Hacker News"Personal Software" i.e. programs that one writes for oneself, was the original vision of home computing back in the 1960s. The PC wasn't really anticipated, but the thought was that everyone would have a computer terminal at home, and write programs to do whatever was needed. It was imagined that programming would become easy enough that anyone could learn to do it. We're not there yet but with LLMs we're getting closer.
Replies
> We're not there yet but with LLMs we're getting closer.
I feel I'm there. Whenever I have a problem I definitely ask "should I vibe code an app for that?"
My current Swift app has 15K LoC (5K in tests, 10K implementation) and it is finished. Like, there are minor improvements now, but it does what it needs to do. It took me 20 hours. I think the actual thing would personally take me 500 hours as I haven't programmed in Swift.
I have told this to many friends who scoff at me, but using a computer is very clearly going to also mean "having the computer create programs for you". We won't even think or know about it.
To me, it isn't a matter of if, and the matter of when is also very clearly in "at most 10 years, probably much, much earlier", given that I have relatives already doing this without knowing how to code.
This is a future of computing I am absolutely in love with, and is so incredibly empowering!
My fears with the situation you are describing is that we end up without a common file format, if everyone has a propietary app and/or file system then that makes transition or collaboration a pain.
We probably won't end there due to how lazy most of us are, but it's certainly something to consider
That's going to be huge thing in the future I think
Everyone having their own hyperspecific apps or even different UIs/visualization in the same app
The whole idea of an application becomes a much more fluid thing
If your app is built with a dynamic language why not let users re-write the code themself and add whole new features
LLMs are great for problem exploration. Especially with the decline of Google I think we're at a point where it's less difficult to get an LLM to spit out something that'll accomplish a task sorta compared to actually finding that on the internet. But if the task is going to be repeated or modified then I think LLMs are at a permanent disadvantage to prebuilt software. Even if that prebuilt software is just someone else running an LLM and then passing the output through acceptance testing most people just don't want the headache of debugging weird edge cases and the novelty of "I'm a developer too!" wears off pretty quickly.
I'm excited that the weird grey-zone of excel sheets with business critical logic is likely going to disappear as LLMs slowly make the logic driving those too complex to be comprehended and managed and those get foisted off onto actual engineering resources. It'll be painful but probably for the best... but for actual tools people need to use day-to-day I think the assurance that the tool will work has a lot more value than the AI hype has comprehended.
It’s exactly what I use LLMs for as a non-computer professional.
> It was imagined that programming would become easy enough that anyone could learn to do it.
Arguably LLMs take us further away from that than we've ever been. All they do is automate copying and pasting in shit from StackOverflow.
We were closer to everyone being able to learn how to program computers in the mid-80s when everyone had one and they started up with a BASIC prompt.
The road not (yet?) taken is the full flowering of the HyperCards, the Visual Basics, the Macromind Directors and Flashes...
That is, the idea that a non-expert might create interesting software in an authoring environment with good, well-thought-out building blocks and easy-to-grasp metaphors, shorn of layers of accidental or over-engineered complexity.
In this vision software still requires careful logical thinking, but it makes it much less cumbersome to translate that thinking into running code, with no tooling and build system nightmares.
Instead, we've invented such powerful models that they can regurgitate and recombine complex incantations on our behalf. The complexity is still there, though, and it's still inscrutable to non-experts.
But maybe they can help us eliminate some of it?
I think that path is still possible, and it may even nicely complement the LLM world, where LLMs help generate software that individual humans can still easily comprehend and manually modify.