Ask HN: The whole NextJS situation is overwhelming

[mbelen]

NextJS came to save us all, until it became a mess. The front-end framework drama has been going on for at least a decade. Theoretically, could we even have some sane resolution with clear choices within the next 5 years? What can it look like?

[jalapenos]

Yeah the solution is no framework and have an AI generate your front end

[solardev]

Sounds great until someone else needs to maintain it. I'd never want to touch a codebase like that...

[xenospn]

Does that actually work? I can’t imagine trying to find a bug in a sea of AI generated JS

[owebmaster]

I second that. React is not needed anymore, much less nextjs. People should break free from these shackles

[mbelen]

But using what, just js + html + css, for anything moderately complex this will become impossible. Even all AI agents will have easier time using some framework as it will be easier to comprehend most likely

[owebmaster]

> for anything moderately complex this will become impossible.

It won't. You actually remove complexity by removing unnecessary dependencies.

You can organize your code the same way you would with React: components, state management, reactivity. It is all there within JS+HTML+CSS without 100kbytes of minified JS.

[Hackbraten]

What situation and what mess? Would you mind being a bit more specific?

[mbelen]

I was mainly referring to the levels of complexity that keep piling up in the front-end world. The main reason, I wrote this is because I spent a day debugging and trying to understand one of the newer caching-by-default behaviours introduced by Next 15 in attempts to solve previous caching behaviours mistakes they made.

But in my post I was referring to the overall state of the front-end tools right now and the frustration that comes with it. From one side it's the best time to build things and see how far we've come. On the other hand it can feel like the unnecessary complexity of our tools only increases rather than making things simpler, I was interested to hear some ideas of where we might end up with all this and how soon

[mubou]

Blaming all of frontend for Next.js' issues is a bit much, but I agree; people have been complaining about Next.js for a long time. It's slow, complicated, and a PITA if you need to do anything outside the box.

Unfortunately, options for SSR in React are all pretty limiting compared to other languages' frameworks, unless you go the full DIY route which isn't usually feasible.

Of course, if you don't need SSR there's no reason to chain yourself to Next and Vercel's whims. Starting a React project from scratch is a lot less painful now with Vite (there was a time when just setting up Webpack & stuff took half a work day).

This is just my personal experience, however. I know a lot of people live and die by Next.js.

[mbelen]

nextjs is great in a sense that it allows to make full stack apps very conviniently. Making an api endpoint is just creating a file. If only there was less stuff in between though

[solardev]

You can use Next with the page router and it's lot simpler. Or Astro?

[mubou]

"Simpler" doesn't necessarily mean better. Next is convenient as long as you stay inside the box, but it's very limiting otherwise (that is, it doesn't adapt to a wide variety of usecases, but rather forces you to adapt to it -- that's my main gripe).