Simply put, the tools are not at the level the grifters want you to believe.
You'll always find someone claiming to have made a thing with ai alone, and some of these may even work to an extent, but the reality is that the models have zero understanding, so if you're looking to replicate something that already exists (or exists in well defined, open source parts), you're going to get further than if you're thinking truly outside the box.
Then there's library and framework churn. AI models aren't good with this (as evidenced by the hours I wanted trying to get any model to help me through a webpack4 to webpack5 upgrade. There was no retained context and no understanding, so it kept telling me to ado webpack4 things that don't work in 5).
So if you're going to make something that's easily replicated in a well-dpcunented framework with lots of stack overflow answers, you might get somewhere. Of course, you could have gotten then yourself with those same inputs, and, as a bonus, you'd actually understand what was done and be able to fix the inevitable issues, as well as extend with your own functionality. If you're asking for something more niche, you have to bring a lot more to the table, and you need a fantastic bullshit detector as the model will confidently lead you down the wrong path.
Simply put, ai is not the silver bullet it's sold as, and the lack of app explosion is just more evidence on that pile.
Exactly. That last 20% is engineering. Handling edge cases, integrating with quirky APIs, optimizing for performance under load. An LLM excels when all conditions are perfect, but the real world is a mess of imperfections
I would say that handling these edge cases for minor changes that a human can easily understand, but is very hard to program rules for, is exactly where AI is the perfect fit.
The link in the last paragraph provides some data to back up the claim. https://mikelovesrobots.substack.com/p/wheres-the-shovelware... - If the goal is to increase the rate of software production, there isn't much evidence that AI has moved the needle.
Sure, code gen is faster now. And the industry might finally be waking up to the fact that writing code is a small part of producing software. Getting infinitely faster at one step doesn't speed up the overall process. In fact, there's good evidence it that rapid code gen actually slows down other steps in the process like code review and QA.
The graphs that this article uses as sources for app store releases are just stock images from Statistica and not the real numbers - looks like that's hidden behind a paywall.
The reason I think so is because I wanted to write a follow-up post, and checked the numbers - for instance, the graph for the Play Store peaks at 140000 app released per month, but all the references I found on the internet were much lower.
I then hunted around for other sources of app store data and found appfigures, which had a free trail. I did a bit of querying and I am seeing a noticeable uptick in number of apps released since around March 2025 (from around 20000 to 35000 for Google and 17000 to 30000 for iOS).
In terms of new GITHUB public repositories, the numbers look correct - so I agree with them there - no uptick in new open source repos in the AI era so far.
At this point, the question we should all be worried about is what is going to happen once the biggest investors see and internalize these articles? Will the economy withstand the collapse of the AI industry and temporary damage to adjacent tech sectors or will this combined with the dodgy loans taken by Meta/Amazon/Alphabet pull the wider economy into a recession?
I guess the author doesn't hang out on Reddit much. A lot of the tech hobbyist subs I used to enjoy are now nothing but a flood of self-promotional marketing posts for vibe coded apps.
I would expect that most apps generated today contain at least some Ai generated code, whether through chat completion or agentic use. But, I think such tools currently mostly support people who already are able to create apps.
As others have said, I think a lot of the difficulty in creating an app lies in making the numerous choices to make the app function, not necessarily in coding. You need "taste" and the ability to push through uncertainty and complexity, which are orthogonal to using Ai in many cases.
But this is how disruptive innovation works. I recall that even around 2005, after digital camera sales overtook the sales of film cameras, people were still asking "If digital is so good, why aren't the professional photographers using them?" and concluding that digital photography is just a toy that will never really replace print.
The digital camera analogy is flawed. Digital sensors had a clear and measurable path to improvement: megapixels, ISO, dynamic range. LLMs have no such clear path to 'understanding' and 'reliability'. It's entirely possible we've hit a fundamental ceiling of their capabilities, not that we're just in an early stage
https://countrx.app/ is something I vibed in a month. Can people here tell?
Sure the typical gradiënt page is something to spot, but native apps i think are harder. I would love to see app store and Google Play Store stats to see how many new apps are onboarded.
Looking at distribution channels like Google Play, they added significant harder thresholds to be able to publish an app to reduce low quality new apps. Presumably due to gen ai?
Edit: Jesus guys, the point I'm trying to make is that there are probably a lot more out there that are not visible... Im not claiming i developed the holy grail vibe coding.
every couple months I try my luck with a very sophisticated prompt akin to: "make me a web application that generates $1k of profit a month. do not hallucinate, it's ok - family friendly, ultrathink or go to jail"
I work on deeply embedded software that doesn't have what you'd commonly think of as a "UI". So, unless there are bugs or we ship faster or something like that, users will never have any idea how much of our code is AI generated.
Not only is this wrong on multiple levels (there are lots of new ai-slop apps flooding the internet, and marketplaces e.g. steam has ~10k games marked as using ai), but it's always cringe when someone names something after themselves like this.
Most of ours are internal-only because we don't need or want to release them to the public. Sometimes there isn't much of an UI - they're one-off vibe-coded apps for specialized functions within our organization meant for a small number of people. Beginning to think of the vibe-coded apps akin to spreadsheets with lots of macros.
Steve Jobs once came to speak at my company when he was running NeXT. Almost nobody came to the talk, in the company cafeteria. The CEO of our company had to make an announcement on the PA encouraging folks to come. Finally, about 20 people (out of ~750) showed up.
He started talking aobut Objective-C and how it was 10x more productive than other programming languages and how easy it is to write good applications quickly with it. Someone shouted out the question: "If it's so easy and fast to write applications, where are all the NeXT killer apps?" There was no good answer....
I think there is a different way to look at it. My personal experience is that enterprises that are at the forefront of adopting new ways of working, are now much more comfortable taking risks with building applications and insourcing SaaS functionality. The amount of custom software build is actually increasing and the codebase are getting more complex. Is there a price to pay down the road? Maybe.
The AI businesses are busy selling AI to each other. Non-tech businesses are busy spending their AI budgets on useless projects. Everybody is clueless, and like - let's jump in just like we did for blockchain, because we don't want to lose out or be questioned on our tech adaption.
The premise is extremely flawed. If users are able to generate their own apps instead of having to buy them, it shrinks the TAM for those apps. If a meatpacker makes its own CRM, it's not going to put it on an app store or try to sell it!
Building software and publishing software are fundamentally two different activities. If AI tilts the build vs. buy equation too far into the build column, we should see a collapse in the published software market.
The canary will be a collapse in the outsourced development / consulting market, since they'd theoretically be undercut by internal teams with AI first -- they're expensive and there's no economy of scale when they're building custom software for you.
If it is so easy to make a product, then why would you go to the trouble of marketing it? A competitor could wipe out your market in as little time as you spent yourself.
My bet is that we will see much more software, but more customized, and focused precisely on certain needs. That type of software will mostly be used privately.
Also, don't underestimate how long it will take for the masses to pick up new tools. There are still people, even here on Hacker News, proclaiming that AI coding assistants do not offer value.
I'd be wary about interpreting a simple trend of App Store / Google Play apps without other context. Both are walled gardens, with developer fees and review processes managed by gatekeepers with an incentive and an ability to artificially control the rate of new apps. I would ask: What is the trend of app store review waiting times? What is the trend of rejections? What is the trend of delistings?
If you're talking about internally developed software: AI generated apps suffer from the same pitfalls.
If you're talking about third-party alternatives: AI generated apps suffer from the same pitfalls.
Bonus reasons: advertising your product as AI generated will likely be seen as a liability. It tends to be promoted as a means of developing software more rapidly or for eliminating costly developers. There is relatively little talk about the quality of software quality, and most of the talk we do see is from developers who have a lot to lose from the shift to AI generated software. (I'm not saying they're wrong, just that they are the loudest because they have the most to lose.)
58 comments
[ 2.8 ms ] story [ 60.7 ms ] threadhttps://github.com/timonoko/Plotters/blob/main/showgcode_inc...
The industrial age was plagued by smog. And so shall be the Information Age.
You'll always find someone claiming to have made a thing with ai alone, and some of these may even work to an extent, but the reality is that the models have zero understanding, so if you're looking to replicate something that already exists (or exists in well defined, open source parts), you're going to get further than if you're thinking truly outside the box.
Then there's library and framework churn. AI models aren't good with this (as evidenced by the hours I wanted trying to get any model to help me through a webpack4 to webpack5 upgrade. There was no retained context and no understanding, so it kept telling me to ado webpack4 things that don't work in 5).
So if you're going to make something that's easily replicated in a well-dpcunented framework with lots of stack overflow answers, you might get somewhere. Of course, you could have gotten then yourself with those same inputs, and, as a bonus, you'd actually understand what was done and be able to fix the inevitable issues, as well as extend with your own functionality. If you're asking for something more niche, you have to bring a lot more to the table, and you need a fantastic bullshit detector as the model will confidently lead you down the wrong path.
Simply put, ai is not the silver bullet it's sold as, and the lack of app explosion is just more evidence on that pile.
Sure, code gen is faster now. And the industry might finally be waking up to the fact that writing code is a small part of producing software. Getting infinitely faster at one step doesn't speed up the overall process. In fact, there's good evidence it that rapid code gen actually slows down other steps in the process like code review and QA.
The reason I think so is because I wanted to write a follow-up post, and checked the numbers - for instance, the graph for the Play Store peaks at 140000 app released per month, but all the references I found on the internet were much lower.
I then hunted around for other sources of app store data and found appfigures, which had a free trail. I did a bit of querying and I am seeing a noticeable uptick in number of apps released since around March 2025 (from around 20000 to 35000 for Google and 17000 to 30000 for iOS).
In terms of new GITHUB public repositories, the numbers look correct - so I agree with them there - no uptick in new open source repos in the AI era so far.
As others have said, I think a lot of the difficulty in creating an app lies in making the numerous choices to make the app function, not necessarily in coding. You need "taste" and the ability to push through uncertainty and complexity, which are orthogonal to using Ai in many cases.
https://countrx.app/ is something I vibed in a month. Can people here tell? Sure the typical gradiënt page is something to spot, but native apps i think are harder. I would love to see app store and Google Play Store stats to see how many new apps are onboarded.
Looking at distribution channels like Google Play, they added significant harder thresholds to be able to publish an app to reduce low quality new apps. Presumably due to gen ai?
Edit: Jesus guys, the point I'm trying to make is that there are probably a lot more out there that are not visible... Im not claiming i developed the holy grail vibe coding.
But it's happening.
He started talking aobut Objective-C and how it was 10x more productive than other programming languages and how easy it is to write good applications quickly with it. Someone shouted out the question: "If it's so easy and fast to write applications, where are all the NeXT killer apps?" There was no good answer....
The AI businesses are busy selling AI to each other. Non-tech businesses are busy spending their AI budgets on useless projects. Everybody is clueless, and like - let's jump in just like we did for blockchain, because we don't want to lose out or be questioned on our tech adaption.
Building software and publishing software are fundamentally two different activities. If AI tilts the build vs. buy equation too far into the build column, we should see a collapse in the published software market.
The canary will be a collapse in the outsourced development / consulting market, since they'd theoretically be undercut by internal teams with AI first -- they're expensive and there's no economy of scale when they're building custom software for you.
My bet is that we will see much more software, but more customized, and focused precisely on certain needs. That type of software will mostly be used privately.
Also, don't underestimate how long it will take for the masses to pick up new tools. There are still people, even here on Hacker News, proclaiming that AI coding assistants do not offer value.
If you're talking about internally developed software: AI generated apps suffer from the same pitfalls.
If you're talking about third-party alternatives: AI generated apps suffer from the same pitfalls.
Bonus reasons: advertising your product as AI generated will likely be seen as a liability. It tends to be promoted as a means of developing software more rapidly or for eliminating costly developers. There is relatively little talk about the quality of software quality, and most of the talk we do see is from developers who have a lot to lose from the shift to AI generated software. (I'm not saying they're wrong, just that they are the loudest because they have the most to lose.)