I love code, you don't, fine. So be it.
But as is: it just output stuff that will not scale properly, and wrangling it takes just as much time as me coding it up myself. So it's in my hands until then. I've found LLM's somewhat useful as a first-order discovery in documentation, and not much farther. Try to trust it as an oracle replacement of all search/documentation and you'll be burned once every few days.
I think you’re letting your pride get the best of you. Writing software should not be about ego. It’s about making something helpful, or something beautiful, or at the very least financially useful. Interacting with software through a natural language interface does not mandate you are creating junk. I won’t tell you how to use the tool. But I am sure if you were sufficiently motivated you’d learn how to make it work for you.
> I think you’re letting your pride get the best of you. Writing software should not be about ego. It’s about making something helpful, or something beautiful
The things I'm the proudest of are, by far, the most useless and unhelpful ones. The ones I did for myself. The ones that made me think "wow, now I rock!" and the machine spit its guts crying "mummy help me". That's what the demoscene (to name one subculture) was all about. That's what hacking is all about in the first place.
> or at the very least financially useful
These have always been the ones I'm the most ashamed about. These were the things I would not have done for myself, but for other people. The ones that kept me living, not the ones that made me feel alive. What did you think I was talking about?
> But I am sure if you were sufficiently motivated you’d learn how to make it work for you.
I'm motivated to code. I'm not an architect, a QA tester, or whatever, I'm a coder. Now, wanna know what kind of person I am and how I made LLMs work for me long before you suggest it here? Read that:
https://news.ycombinator.com/item?id=48072319#48073350
Then, come back and we can talk. I WILL NOT ASK A MACHINE SOMETHING I LOVE TO DO MYSELF, WHICH I TAKE PLEASURE FROM. If that looks like an idiotic and stubborn stance to you, FINE.
Also, humans didn't evolve to write or read code, even the best programmers are terrible at it from an objective point of view, if we compare it to what is theoretically possible.
And, no, it is not a question about style or preferences, the code is messy and contains to me as a maintainer obvious bugs.
It's interesting you feel comfortable to comment with such authority on the capability of a tool you don't even use... Lisp is a bad programming language because my 10 year old brother wrote some really bad stuff, btw!
Basically, they still need a human guiding them pretty closely unless you’re happy to just ship a mountain of vibe coded mush.
That said, I haven’t tried Astra in earnest yet. We’ll see in a few weeks whether or not it is a step change.
These companies have many billion dollar investments and yet their main product is a bloated meh.
Just as they produce writing that isn't structured well for humans, they do the same with code. Sure, if you're willing for that code to be write-only and for agents to be the sole interface through which you interact with it, that works fine, but it is not better than what a professional, skilled human could produce. It's often full of pointless crap and design decisions that over index heavily on local structure that make very little sense when global system structure is considered.
I think the question is less about the capabilities of LLMs and their outputs (because obviously that can change), but because they represent a workflow which codifies exactly that role I mention. You fulfill the role of a manager more than a programmer. You have a vague idea of the codebase, but lack any kind of intuitive mechanical understanding. There are many cases in which I can see an issue, identify the problem, and have it fixed faster than you could prompt an LLM to investigate. After a certain point of familiarity, all reasoning is local reasoning. That's the power of mechanical sympathy.
Many, many large shops don't give a shit. They've long since galloped into the dark woods of bureaucratic know-nothing hell. Let the LLM figure that shit out. Those shops are already optimizing for maximal output and churn, so the value proposition seems obvious. I've been smugly saying it for years, but LLMs have me saying it quite a lot more these days: "Sure looks like it sucks down there!"
That's a weird way of looking at it. Programming lives at the end of a well-worn path uncountably many millenia long, defined entirely by our biology. We didn't "evolve to do it" in the same way a duck didn't evolve to swim in a specific river.
Current models are in fact much better than the top models from a year ago. I have no idea how anyone could claim otherwise without a huge amount of motivated reasoning.