Where is all the new, good software?
natemeyvis.com
natemeyvis.com
Writing code is not the bottleneck anymore, making decisions is. Making the right decisions needs context and thinking. A coffee break might actually help productivity. Driving a Ferrari full speed in a crowded city - only because you can - doesn't make the outcome better.
Even in the 90s, the internet and the web dramatically lowered the costs of distribution, and YC itself was built on top of that (ViaWeb).
Then there has been so many frameworks, libraries, tools, services and APIs to pick and choose from that most of the coding work has been about organization, selling an idea and doing a lot of testing. Which are still unsolved problems.
The only new idea here is the idea that Software Developers are disposable.
This was brought by the same people who did RTO and layoffs post-covid, but now with support of non-tech workers who want a piece of that (supposedly) sweet money.
Always has done. Writing code was never the bottleneck.
Hammock-Driven Development: https://youtu.be/f84n5oFoZBc
Some motherfucker today suggested we should use Haiku for all coding tasks, and this guy is a Sr Director of real humans.
I played with our AI and it found about 100 real but subtle issues in the codebase. Took less than 30 minutes to find what would have taken a sharp senior months.
I started reviewing the changes manually. They look pretty good but I want to understand each one. Some are design issues and our product needs weren’t spelled out to the AI.
I could blindly put all 100 up for review in half an hour or take a full week to scrutinize them.
So yeah, the process with AI is going to take 100x longer what it could have been.
1. A frontend to MacOS Quicklook that gives me many additional playback controls.
2. A daemon that listens for messages on my LAN to add scheduled notifications to my preferred app launcher without cluttering my reminders.
3. An improved version of a tool I built to watch for specific streamers and capture the content compared to the one I created years ago. I wouldn’t have enhanced it beyond the version I had that was good enough.
4. A metadata inspector that I can invoke in the GUI with a hotkey focusing only on the stuff I’m interested in seeing.
5. A hotkey terminal launcher based on the color themes I have configured on my device so I can arrow through and choose in the moment.
etc, etc, etc…
I saw many comments here saying SaaS is dead, for example. That any service can now be vibecoded and replicated internally. And I can't help but think if they've ever tried running, supporting and maintaining a piece of software. The code was never the hardest part. Not even close. In fact, cheap "coding" has been a thing for over 25 years now if that's what you wanted.
To me this feels like saying everyone can be a carpenter now because it's so cheap and easy to just buy a hammer and some nails and go at it.
I do think it's great to hear that people can now vibecode their own custom fit tools that work exactly the way they want them to. But that process is not free.
> We are in fact seeing software improvement, in the form of more and better features in existing products.
After reading these two points, I had to scroll down to see if it was the opening of some elaborate joke, because I'm seeing more bugs, worse performance, and less and worse features.
1. Marketing (and distribution) of said software is a chore, and doesn't really give me any benefits 2. It's tailored to me rather than to average Joe.
Wife: Have you got anything without slop?
Waitress: Well, there's slop egg sausage and slop, that's not got much slop in it.
Wife: I don't want ANY slop!
Man: Why can't she have egg bacon slop and sausage?
Wife: THAT'S got slop in it!
Man: Hasn't got as much slop in it as slop egg sausage and slop, has it?
I've been able to create my own tooling that beats out the preexisting best across every metric. It's incredible.
So maybe the big revolution is that we won't have "big software" that tries to cater to all kinds of users but "small software" that is tailored perfectly to the needs of a single user.
There's a bunch of reasons for this, but the obvious one to me is just that I haven't seen any backed up claims that actually break down how a 10x (or similar) speed up would get delivered in practice.
For the sake of an example, let's imagine an engineer who works an 8 hour day split up like this:
- 2 hours: investigating and debugging potential issues
- 1 hour: reviewing code
- 2 hours: in various team meetings, catch ups or otherwise
- 1 hour: designing (including at a level of writing code for prototyping etc)
- 2 hours: actually writing code
I think that's a semi-reasonable day breakdown. Realistically, I think AI helps debug faster, and generate code faster. Possibly design work too, but I think this isn't really the case since AI workflows necessitate a lot more design work in order to make the AI generation deterministic.
So looking at that day, if I was able to fully automate writing code and debugging issues so that it took 0 time at all (which I think is a very optimistic assessment of the actual speed ups AI provides), we're looking at a 2x speed up of an engineers productivity.
If we think AI makes those things twice as fast (so 2 hours, instead of 4) the engineer is now at a 1.25x speed up. That's great, but it's hardly revolutionary. It'll change your output something like the way going from a team of 4 to a team of 5 might do.
This is obviously complete conjecture and I'm sure the average time engineers spend on things is different, but I think there's an obvious take away. If you want to make claims about the performance impact of code, you need to know where the code spends its time, and what impact you're making on those sections of code.
Similarly, if you want to know what performance gains are being made, you need to know what engineers spend their time doing and what impact AI is having of those tasks.
I've seen very little attempt to actually make realistic claims on that basis, and a lot of vibes-based "wooah, AI makes me so much faster" type claims. Those claims are obviously not nothing, but I wouldn't trust people's gut reactions to make serious claims about the overall productivity changes in our entire industry.
You need to be an an org who is able to deliver directly to your consumer, who needs to be defect-tolerant with the expectation of quick iteration. That customer will have access to advanced tools that its competition is unwilling to adopt. Many will fail.
There are many verticals for which this does not work at all - but this is unlikely to be the permanent state. The new model will work its way through the market, and companies will fall as fast-moving competitors find a willing customer segment they can target, make that segment out-compete the others using their access to tools with superior abilities but higher defect rates. The customer base will eventually adopt, having become accustomed to this, and so slowly expectations will change, and so will our field.
I for one welcome our slop overlords.
This is not what I am hearing from people who work in LLM heavy environments at all. The exact opposite seems to be the case.
I would like to see some of this software improvement in the form of fewer bugs. If anything it's adding more bugs.
I'll give an example from a game I like called ARC Raiders. For the first ~9 months they did not fix any of the bugs in the main menu. The damn main menu! The first thing you see when you load the game! They also failed to fix, in that time, the bug where simply pressing the Xbox button on the controller would break in-game voice chat.
They only just solved that, as of a few days ago. The main menu bugs still exist.
Just because you can throw tens of thousands of dollars at frontier LLMs doesn't mean what they write would be good, you still need to have some oversight for their work.
It works relatively well in native mode as a standalone app (tho not being an actual photoshop expert, i dont know if the features compete properly, i simply hear that those trying it are impressed).
The excel clone works well enough for me to do taxes on, even tho the look and feel is a bit janky. I would stick with google sheets, but it's a usable replacement if i didnt already have sheets.
I'm seeing people vibe-code inference engines for custom hardware/models too. Are they good? That remains to be seen, but I didn't used to see this at all.
It does feel like something has changed to me lately. I expect it's only going to accelerate, and the quality only going to increase.
AI written software doing the exact thing it was supposed to but breaking apart on any further improvements or changes matches what I've seen from AI generated output. The code is often bloated and fragile.
In my professional life however, we are NOT producing more software ($bigco). We are on the edge of it maybe. My experience is we've been on the razors edge - just enough staff to meet the business needs - and so every AI iteration sped up some process, which exposed some limit in a dependent process, which we tried to free up with a reorg, which cut capacity somewhere else, and the snake eats its tail. We've had enough iterations to get to a place where an increase in developer output probably won't cause the whole thing to fall over. I hope.
But anyway, I spend a lot of time exploring new dev tools, it's sort of a hobby of mine; I like to understand the state of "what's out there" and this has helped me professionally. This has recently become impossible do the sheer mass of llm productivity, while at the same time more fun than ever; the amount of crazy interesting and seemingly mature brand-new things (in the dev space) has astounded me. I found Swamp Club recently, like wtf where did this come from? It's like five dudes with LLMs.
My own company has shipped a slew of features that were going to die in the backlog, tracked down almost every bug that comes through Sentry, and given me time to focus on bigger things (it's a 1-person company).
For myself, home automations that I was never going to spend the time getting "just right", I've replaced MyFitnessPal with an app I built, I've got more self-hosted projects running than ever before, and I've built websites and apps to scratch mine or my families needs that were never going to get built in the past.
When I run into an annoyance on my computer(s) now I don't just put up with it, I send an agent after it to figure out why it happened, how we can stop it from happening again, and/or if we can write some custom software to make the problem better or go away. It feels magical.
Pedantic much? It's new to me, it's new to the people using it. What's ever truly "new" then?
> Slop clones
This is getting boring. Many of us are creating custom software that scratch our particular (and sometimes others') itch. You can call it slop or whatever derisively term you want, it doesn't change the reality. Before I didn't have an app that did what I wanted, the way I wanted, now I do.
Even if some bloated, expensive software that may have contained those functions already exists, bespoke software that does exactly what I want, and only what I want, in the way I want, is still "new" in my opinion.
For example, it's extraordinarily good at applying the searched-for information (once found) to a specific problem at hand—at least at the first-order level of putting existing pieces together in a correct way. It can do that much faster than I can, with far greater command of detail.
It can also find problems and correct them in my own work in a way that still amazes me. Again, not deep problems—they're first-order in the sense of being obvious deviations from correct practice - but to call this 'search' feels like a stretch. It's not searching for bugs in the code I just wrote and then returning a corrected version, as if one had been sitting somewhere waiting to be found.
Plenty of people insist it goes far beyond these types of things, and they're probably right—I'm not going to say they aren't, since I've been wrong enough times already. My point is that even these first-order things are enough, for me, to refute the it's-just-search view. It does things with search results that in the past, I would have had to do.
I am not sure what the current state of models is, but as far as I know lifting a limit on the model output length tends to noticeably degrade the quality of the result so that it starts to show the same delirious tones.
Fundamentally it is recalling of past experience, a reaction. In many ways it is an opposite of acquiring true new knowledge. We keep that mechanism in check using other human qualities models do not have.
Working with LLM is basically like hiring a madman. It can do useful work, but you cannot trust it. To be economically viable the verification cycle has to be cheap or the cost of an error must be small. Searching a library or doing a home project are fine. But these aren't the lucrative markets investors are dreaming of.
Straight up untrue anecdotally and by general sentiment in the industry.
Most new tech is just shit.
This might be the biggest factor in my opinion. Corporate incentives usually don't reward someone delivering 2x-10x, so an employee who suddenly found a way to multiply their productivity might be keeping that method to themselves and freeing up some time instead.
I have multiple open-source projects and 2 commercial offerings (see: https://race-term.com, https://inboxctl.com and https://github.com/exlee).
With RACE (which is great I use it daily) I have maybe 30 users. With Inboxctl there are (I think) 2 except me.
Recently I figured that there's no reason to work on open source because it's going to be a vaporware even if I set up the whole architecture. With paid software I might (at least) get paid for my time.
But long story short: there are no channels to share/market/show people. HN is hit or miss (most often miss).
Is it good software? Yes. I'm in this industry for more than 2 decades and it ticks all the boxes for "good software". But all the channels where you could meet it are dead waters.
I am yet to see the advancements in AI manifest in more leisure-time. It is simply that the amount of work we're expected to be able to pump out has increased, and the actual programming work is no longer what slows us down.
It’s a lose/lose proposition for every software company with more than three management layers. We're frozen in place.
Thanks to AI, I was able to finish my bookkeeping and accounting app I had been working on years ago. https://financica.app
It’s not just “vibe coded”, it’s really damn good. Because thanks to AI, whenever I spot an issue, a quality of life increase etc I can just immediately implement it rather than have it seat in a perpetually growing backlog of “maybe one day”.
And there’s SO MANY standards to comply with especially in Europe and I just … implemented all of them. It’s more full featured than any of its competitors.
It runs the bookkeeping and accounting for all of my own companies plus plenty of customers. And the funny thing is for me, this is pretty much a side project. I just have people helping on sales, but all the development is me.
My backlog for my side-project/business had a number of things in it that were frankly never going to happen. They only existed in the backlog so I could say they were in the backlog and add a note about who expressed interest in it so that, if it reached a critical mass, I could justify dedicating the time. That backlog is largely empty now. Things I thought were a pipedream or a "well, maybe one day if this gets bigger" became a reality.
But more importantly, why should the market accept your software? Since plenty of people are "making more software" cause they are trying to "make software better" while creating similar product with the use of AI, its difficult for the market to choose which software actually solves their problem.
On the other hand, many don't accept or adopt new software as they don't see the problem they are facing, or they don't mind the problem itself. So why bother adopting new software that isn't going to pluck the problem by the roots all together with its' use?
Human coders usually ask this question in a malicious way to imply that this technology doesn't just have to be faster than humans (which it is), it also has to produce better code (which it usually does) as cohesive product (so fixing features or refactoring in projects/products do not count) in record time (this is the trap because these models are not even a year old in their current advanced format). It's hopeless.
Anecdotally, completely disagree. Maybe this is nostalgia, but I perceive a distinct and massive drop in software quality over the past few years.
Worst offenders for me are the poster children for LLM-based development. I run into bugs and performance issues with both Claude and Codex literally every day. I don't think I'm an outlier here, either.
- more bugs becoming visible that were hidden within the software we use daily
- more bugs (in absolute terms) as more people create more software in a very short amount of time (regardless of their professional training)
- less bugs in existing software as the developers and dev orgs of existing projects is capable of automating bug hunting, test case creation, ... in essence: They get a multiplying effect out of AI regarding their capabilities.
In short: Actual developers with actual experience (the more, the merrier) can do better work while the barrier to enter "the market" for untrained individuals is a lot lower resulting in a huge amount of new software whose "success" is to be determined.
I'm not convinced there are more than a handful of orgs in the world who are disciplined enough to achieve this. Or organisationally willing to increase their QA Eng headcount sufficiently to achieve it.
[1] https://itbranschen.com/wp-content/uploads/2026/05/the-hidde...
AI firm code are likely a different category though.
If the leadership (C-suite) is afraid to take the full responsibility over AI output, then the only rational choice for developers is to "pace the frontier" themselves.
If there are skills, design MCPs, automatic testing and evaluation, feature bug monitoring, security monitoring, crash handling and automatic fixes, then we could see the productivity in teams. But now it is still our responsibility and not responsibility of the system.
Move fast and don't break things.
And then the C-suite fires them and replaces with ones who are not pacing it. It must come from the high level.
https://news.ycombinator.com/item?id=50065892
Nice one...
What we've seen very little of is fundamentally new ideas, usage paradigms, user experiences or similar.
What I thought all along: the fast-food era of software has come.
Its so weird I spend every day AI coding, thinking I'm building software, and at the end of the day, strangely, nothing was built at all.
The next day the same thing happens.
Here's the part nobody tells you: I think this guy is right - AI is building nothing at all.
BUT if you want to see what people are building, visit https://www.reddit.com/r/SideProject/ where there is roughly 355 project announcements per day.
This is just plain wrong. I don't think the people asking for new software are in any way averse to adopting new software. I for one am actively waiting for new software to be released so I can adopt it, the Rad debugger for example. The first Linux beta-builds have been shipped and I can't wait for it to be in a more usable state.
I try new software all the time, not to mention I usually try the same software multiple times before reaching a final verdict. In the text-editor space I've been around Vim, Neovim, Doom Emacs, Spacemacs, Zed, Vs Code, Helix, Kakuone, all of them several times over before settling on Neovim.
I would love for more actual CAD-alternatives, more Photoshop alternatives, more video editors (shoutout to Blick) that all cover some niche. They don't even need to be full replacements!
But no. AI has not helped on this front in the slightest. If anything the bleeding edge (if you wish to call vibe-coded, half-assed clones that) is in its most broken state ever.
Point Jev at them, then!
All you need to do is become a UX designer and complain about every small problem you see. That way any bad patterns leaves the repo fast and your agents will repeat the good ones. Pick frameworks with minimal boilerplate and good performance so you don't get lost in syntax noise or end up with Electron slop.
And always read code before you commit, that's the only thing that separates you from a stupid vibe coder.
I think AI is reshaping the industry but it's still far from the promised benefits. But the Silicon Valley hype machine isn't happy with "reshaping" because that's not enough to attract trillions. It needs to be "redefining the meaning of being a human" and garbage like that.
The reason we're not seeing a lot of good software is because the current AI tech can't (yet?) deliver that by itself as advertised.
> Software teams are capturing a lot of the speedup in the form of more leisure or quasi-leisure time. (To greatly simplify: software teams are offering 2x speedups to their companies. They're then using tools that allow them to work 10x as efficiently, and taking nice long coffee breaks.)
That is just an insult. With AI we are working more than ever, either because of "you should be faster now" or because half the people are getting layed-off ...
We've lost 75% of the team over the last year.
We write code faster. It's not getting merged faster because we usually find something janky.
Decisions aren't made faster.
Understanding doesn't come faster.
Code doesn't have value. A well working system has value, the bottleneck is now between having code and having a working system, rather than producing the code.
With AI you can write a new Harry Potter novel in under an hour. But if I ask you about the story you can only answer “It involves wizards”.
In pre-AI era you could solve a bug in one week. But you would end up having a detailed mental picture of the system, which requires time to be created. That mental investment helps you make better decisions on the system or solve related problems instantly.
Clear your glasses. For example - https://getartcraft.com/apps Thanks to those people we may soon have great alternatives to monopolies.
Adobe is cheap compared to that miserable experience.
What it really shows is ego, arrogance and disdain for the users while running an AI cargo cult.
> What it really shows is ego, arrogance and disdain for the users while running an AI cargo cult.
It seems you have unreasonable expectations of a new piece of software because you don't like people who use AI.
Sounds exactly like every Adobe product ever. I guess they went too far on feature parity.
And that is because they've had 30 years of effort thrown into them. It's not the big picture that's the issue, it's the thousands of small things. Stuff hidden in the menus that millions of people use. The corpus of educational material out there to help you get stuff done. And the flexibility which is really difficult to match. A good example is the ability for Lightroom to pull reasonable detail out of shadows from RAW images. No other product can touch that, not DxO/affinity/Photomator/Darktable/RAWtherapy.
LightCraft's exposure slider doesn't even work well. Blows everything out. No interpolation curve. It's just BAM!
And remeber Flash? The software that Adobe bought and murdered becuse of their software development inaptitude? If it didn't end up in their hands it would be part of web standards by now and all of the human creativity that this software spawned wouldn't end up smashed into smitherines.
Adobe is cancer of user facing software. Always has been.
Adobe is a company that firmly squats in few very large niches of human activity and despite its mediocrity and awefulness starves any other software that dares to venture into them. It should have been broken by anti-monopoly laws 3 times by now.
I don't know how somebody can look at how much and in what manner they are charging their customers with cancelation fees and such and think that it's an evidence of providing great value and not that's it's a sign of clearly patgological grasp on the market that they have.
I will never forgive them killing Fireworks and Flash and developing Acrobat, Photoshop and Illustrator the way they did.
I haven't touched any of their stuff since content aware resize. I'm sure that they learned to polish their turds, because it's 2020's, everyone did. I'll be celebrating their ride down as more and more people start banging out simple tools with AI agents to make the edits they need instead of digging through tons of documentation and "educational" materials to find the one almost correct option at the bottom of nested menus of Adobe crapware.
> A good example is the ability for Lightroom to pull reasonable detail out of shadows from RAW images. No other product can touch that, not DxO/affinity/Photomator/Darktable/RAWtherapy.
> LightCraft's exposure slider doesn't even work well. Blows everything out. No interpolation curve. It's just BAM!
Do you know that an agent can build a simple Rust app for you that's going to be small, blazing fast and will implement for you the best 3 algorithm for that, straight from their respective papers and give you knobs for adjusting every parameter they have with any manner of precision?
Unfortunately corporate IT doesn't care - they are not using it. On my own systems I've long switched to alternatives.
I don't like the Creative Cloud suite that much (FWIW I prefer Affinity, A.R.T., DxO PhotoLab, CaptureOne, any real alternative).
But Adobe's apps are pretty solidly reliable, if uninspiring in places (Affinity's layer model is better for example).
All of them cost money. They are ripping off Adobe to create the new Adobe.
The worst thing is that there are genuine legit open source projects that already offered an alternative but they did not need AI or ripping off Adobe as a marketing gimmick.
Vibecoded should not be an endorsement in 2026, it should be a warning.
As it stands right now, it proofs little more than it can replicate the post-boot screen of the original products.
Just because every tech news outlet was dumb enough to report it without trying it doesn't mean you have to join the echo chamber.
---
There are definitely examples out there that do use a similar relentlessly agentic development style that have produced great results. e.g. xberg[0] (formerly kreuzberg), is a great document content extractor, and in a relatively short amount of time has become one of the most robust and performant. I've reported about a dozen edge cases of documents of various types that were unextractable, and the maintainer (or their agents) were able to add a fix in maximum a day turnaround every time.
Moral compass/integrity stuff. Can't explain why, it just feels like absolutely the wrong direction to be heading in.
If I posted it you would call it slop.
I made a cute little ESP32 flight radar for my desk. Two, actually, one for my wife. It shows the planes that fly over my house and has a little web interface I can use to set my position and a few other settings. It runs 24/7 and brings my joy to look at.
If I posted it you would call it slop.
I'm designing my first synthesizer / groovebox. It has 36 hall-effect keys and buttons, 7 rotary encoders, a beautiful monochrome blue OLED, audio and midi in and out, stereo microphones, SD sample streaming, and an audio engine that scales across multiple ESP32-S3 modules. I started the firmware by hand but quickly realized I could get everything I needed much faster with gentle prodding at Claude.
I've been working on it since January and have put hundreds and hundreds of hours into it. I'm finishing up rev 2 of the PCB and I hope to one day sell a few.
When I post it, you will call it slop.
People are learning to keep their projects to themselves, because the Internet is full of assholes. Look at the comments already on this post. A large group of people have decided that AI does not work, whether it works or not.
..
> Again, some of these are speculative and some should be given little weight.
You can drop this right off the top of HN when the OP blatantly acknowledges that their entire post is BS.
“Oh, you can’t critique me, I said these were just speculative…”
I would not say that software on average is getting less buggy and more performant. I think the only big exception is being able to bruteforce-find security vulnerabilities and similar defects, which LLMs are well-suited to. But when you look at the remaining 95%+ of the software market, where everyone is seemingly all in on AI, all I see is software becoming more janky and less stable. Everyone just ignores it because it's faster for now.
> We are in fact seeing software improvement, in the form of more and better features in existing products.
Maybe 'features' in terms of measure-turned-target metrics that management at software companies loves to use, but overall I'm not seeing an explosion of features that's inconsistent with the rate of improvement until today. And what new is here is more janky than ever before.
> We are in fact seeing software improvement, but the "we" is very limited.
Surely the people on a tech news website would know about this very useful AI-only software, even if they didn't start using it. It's funny how no one can even give some names.
> Many businesses' software-development processes have subsystems, especially political subsystems, that have not been accelerated with AI.
The entire industry is obsessed with AI, almost everyone is stripping out everything that's not AI and replacing it with AI. Doesn't seem plausible.
> Software teams are capturing a lot of the speedup in the form of more leisure or quasi-leisure time.
How many people did you talk to recently that said they're spending less time working?
> AI improvements in software sometimes push us to use and make software less.
In what ways? AI, by allowing you not to understand the solution, can carry a doomed project much further than a regular programmer would before realizing their mistake. If instructed to, the AI will just build and build and build until the thing implodes under its own weight. It gives you less awareness of what you're doing.
All of these are unconvincing. I think only one of the points has merit:
> We're making more software instead of making software better.
I think people try to avoid acknowledging that AI may be much better at improving the speed side of the equation than the quality side. Look around - all you see is demands for more things faster, even if it's total garbage in the end. They didn't want to bother with quality already, now they found an excuse to try to forget about it altogether. If it's wrong, we'll just regenerate the same garbage 50 more times, they say. Focus on shipping something now, they say.
- or make a microsoft office suit that does the same in performance
- or take any game's exe file and add coop to it . I drag and drop COD black ops and I get 2 person coop mode added to the exe instantly
- wake me up when one of your frontier models are capable of pulling this kinda feat