> Those working in specialized fields will hold on for much longer.
The thing about AI up to this point is that it has replaced, well, very few jobs, but what jobs it seems to be most capable of replacing are extremely counter-intuitive. Eight years ago, everyone thought it would be: self-driving, machines, hard labor, etc. Turns out that stuff is really hard, and the first industries to fall were actually the more creative ones like writing and art.
If you have an intuition for how the next ten years will look, I'd implore you to be open to the reality that there is no way you could have predicted the world-state of 2024 from the perspective of 2019; and that's only 5 years.
Here's the counter-intuitive take that I believe to be true: Specialized coding might fall faster than generalist CRUD apps, if either falls at all. The value in specialized engineering is biased a lot heavier toward knowledge of the specialized thing you know about. Versus, the problems generalist CRUD coders deal with every day aren't actually technical or coding problems; they're business problems, coordination, resource allocation, and politics. AI has demonstrated itself as being pretty good at knowing things, even highly specialized things; it has not demonstrated itself as being very good at taking responsibility for its actions.
We just don’t have a massive breakthrough in robots at a level comparable to LLMs. Which is kinda weird because there would be of even higher demand than language models
Both are technical problems with technical solutions that have to fit 100%. You can't say this for art or writing. It's more mushi and anything can fit there.
Oh nah, I just use that term incorrectly to mean, its likely someone somewhere has lost their job directly because of AI. I've heard a couple horror stories I'm biased toward believing about copywriting and e.g. video game concept art.
There's anecdotes [0] [1], surveys [2], analyses [3], direct statements from companies after layoffs (Duolingo and Klarna for example), a decrease in job postings and a decrease in freelance work. But HN just shrugs and says none of this counts because it's not hard empirical data (as if data about jobs was ever reliable). So let's just keep pretending everything's fine and tell the people who lost their jobs to stop gaslighting us.
I work for a company that builds apps for companies, most of our apps start as "a simple crud app" and turn into massively complex projects. I feel as if the business community overestimates the ability of AI code generation. It'll get better and better, but I don't see it getting good enough not to have a developer cleaning up its mess behind the scenes.
ive worked as a software person for 15+ years, largely in an agency setting, and literally the first "complex" project that came to mind for me was a company that makes custom doors. just doors. they used an excel spreadhsheet that was 200+ megabytes and 40+ sheets of extensive calculations and was used for both estimating prices and also the extensive component sourcing and a billion other parameters and in their 100+ employee company there was maybe one, two people who fully understood this shithole of a spreadsheet. as an agency, half a million dollars was understanding this stupid spreadsheet and documenting what this looked like as real sofwtare, and $150k was actually the development. AI would have zero chance of doing this work any time soon
CRUD isn't the complexity with CRUD apps. Working with the rest of the business to make sure the processes get built out in ways that make sense and are actually worthwhile is.
So maybe one dev can handle writing five or ten times as many CRUD services as before. The idea that the biz person is gonna get rid of those devs entirely and babysit the coding agents/deploys/etc themselves strikes me as wildly unrealistic, though. Nor is it realistic that SAAS as a market will just dry up because individuals will just have their own coder agents that do everything for them.
Question is if there's going to be an "enough is enough" point feature/change-wise. But I can't think of the last time I worked at a company where the product team and senior management threw up their hands and said "we did it, we finally cleared the entire product development backlog" as opposed to generating ideas at a rate many multiples of the rate of implementation.
The baseline "here's what you can get with a shoestring team" level of output will get higher. But it will also get higher for your competitors. So expectations will just be much higher. So you just gotta move that much faster now.
> But CRUD apps, apps requiring gluing bunch of APIs with backend and database wil go away.
They were already going away with the improvements in no-code tools, Zapier, scaffolding frameworks etc. LLMs are just another automated tool in this context.
Those working in specialized fields will hold on for much longer.
But CRUD apps, apps requiring gluing bunch of APIs with backend and database wil go away.
Most apps do not need much scaling so highly specialized scaling masters aren't really needed
Currently using zed editor, I am blown away by its Ai integration.
Though completion via FIM throut custom LLMs is lacking
And there are other problems like lacking git integration, I prefer vscode for that!