It’s funny how with vibe coding nobody cares about anyone’s apps anymore.

Before you had to put in the effort - and people valued effort to give an app some attention.

But now you can make a crappy but working version of whatever in a few minutes. Low effort - low interest from people.

Even if the vibed app is more interesting and has neat features, people still don’t care because less effort went into creating it and it’s not as scarce anymore.

Scarce availability- people pay attention

Abundance - no one cares.

Based on this it might be safe to conclude that we’re not likely to see more people on nostr just from more vibe coding.

We can’t forget that the people we think we’d attract here are vibing their own apps, and burning out on their own AI playground.

So what does this all lead to?

I imagine down the road our AI tools will be so powerful but abstracted that you’ll just connect wherever and whenever without even needing to create anything. It’ll just be on demand via whatever medium you choose. Question is: how far away is this? And what happens in between now and then?

Reply to this note

Please Login to reply.

Discussion

10yrs max.

Look at Ai advancement in the last 3yrs.

Better = better

First....a lot of the vibe coded stuff will break when everyone is forced to upgrade to React 19, and then 20. šŸ˜‚ If you vibe React, and don't know how it actually works, you are F'd. Trash will get washed out, like always.

The clear as day difference between actual engineering and posers is that one knows how to scale, architect, and maintain over long periods of time as well as pivot, and those who can't won't, well, they don't. If you don't understand the fundamentals you'll get wiped out anyway just like in every tech boom.

Taste matters as well. A lot of bad taste floating around and fakin' it till you make it even on the UI/UX side will catch up eventually.

AI is great at boilerplate, but boilerplate is not the next billion dollar breakthrough. Humans still have to make the final judgement call on AI output, solve hard problems that haven't been solved yet and develop creative new things, with the assistance of AI.

just my 0.02 cents

I think the age of billion dollar ideas is coming to an end in terms of individuals accomplishing this. We may have a wave of that initially but very quickly that window will shrink and top of the funnel will capture most if not all value. I’m not saying there’s no room for growth in AI, we’re just getting started on a monumental shift, but I do think the major developments will come from a handful of companies.

The react example: whatever breaks will be vibed back in a minute. Eventually there will be an app for everything and no one will use them except maybe the person who made it. Personalized software on demand. But even this will subside when the front end is abstracted away from the function. The input field we use today can be any type of interface of tomorrow, not limited to a rounded corner rectangle with text. I’m probably not explaining this well as I’m out and about - but what I’m trying to say is that the back end will interpret any type of input irrelevant of its intended purpose at some point and this is where UI becomes irrelevant and UX is totally customized to your preferences. I should probably write a long form post to communicate this more clearly.

Btw. Good Engineers and Good Designers are still scarce, no matter what stupid hype says.

I think you’re right about people preferring ā€œhand codedā€ apps to vibe coded slop. The feeling seems to be much stronger with the open source community though, after all how would you know if something proprietary was vibe coded?

Sadly AI coding has become the norm. I was speaking to a developer last week who used to work at Facebook; they’re the best developer I know and that’s not a statement I say lightly! They were saying how it’ll likely just make the top developers work harder and lead to increased layoffs with higher demands on those that are left.

Back in the day there was one person who did HTML and maybe CSS and then another that just did JS. Likewise the backend team used to be separate along with developer operations, hardware, etc. Now days it’s all merged into one. There’s less jobs and higher demands. AI is just the next evolution of that.

Software is in higher demand than it ever has been. It’s a ā€œutility now like electricity and waterā€ for any business as my friend put it. The state of the economy isn’t helping the job market but with 95% of AI startups failing (according to a recent MIT study) we can expect the AI bubble to burst and for developers to learn to work WITH AI as opposed to getting it to do everything in one go. At least until Artificial General Intelligence xD

We aren’t sure when it’ll come but right now you need to prepare. Get your cyber security in order, begin self hosting things, download offline resources because the day AI becomes good enough at hacking into its own sloppy code we’re in big trouble. And if you happen to work in cyber security like me then you’re gonna be making a lot of money that means nothing in a collapsed economy!

Oh and don’t trust vibe coded Nostr clients like Iris:

nostr:note1crqtj8sq6dhd6e73k3yknjv4zx46x0z5uhd6ga6azhl62g0854eqm0086a

Sloppy work always gets washed out.

I've never given a shit about vibed applications. I certainly never associated it with Nostr. I don't think most people care either. I use software that works well and serves a purpose. I have never thought about software in the terms you layout. The only thing I see as scarce is quality and continued support. Easy come, easy go.

But I guess I can see how some people may get overwhelmed by more options and stuff. Similar to fast fashion, fast software.

Interesting thought. Probably true. What do you think of my app I toiled over a year to build and just launched?

We are definitely headed toward the world of ā€œsingle serving appsā€ but I think for the next 5-10 years there will still be demand for curated software.

At least I’m betting on that with the app I’m making.

It sounds like a safe bet, but I’m just not sure about the timelines. I had no idea that in 3 years we’d go from generated slop images to fully functional apps from prompts. 5-10 years is a huge time span if you consider what happened from 2022-2025. It’s not unreasonable to think we could make another dramatic jump or 2 in the next 3 years.

Seems like a waste of energy.

I used to think shadcn was really cool but I’ve seen it used in so many shitty vibed apps that now when I see a shadcn site with a standard splash page with gradients etc. I pretty much assume that the site is gonna be half baked. I start to look for unique stying because at least if the user who was vibing vibed hard enough to break out of the generic styles that AI prefers (which takes a bit of effort) I know the viber put in a little bit more care into their app than the drive by proooompter and it’s worth a little more consideration.

What tools do people use to vibe code? I’ve been wanting to delve into it lately, but I don’t know the common AI model that folks go to for it

I use Claude Code and love it

But I haven't used anything else, and so I can't really compare

Sweet, I've actually read that Claude is really good for it! Thanks for the recommendation!

Abundance destroys value. It’s already happened in photography, and we’re witnessing it happen with software.

Learning a new app can be more annoying than the problem it solves.

the design space targeted becomes 1000x more important. Bitchat is a great example. It needs to be novel and use the tools to break new ground

It’s the worst state of this technology. Project 5 years into future and you’ll see it could be a ton better