People can dismiss AI coding but some of us are using it for actual products and making money from it. I have heard for two years now that I will pay the price of vibe coding as I will not understand the code when it breaks. What? I just ask the AI what the code does to understand it and vibe code to fix it. What the heck is everyone on about? Knowing syntax is about as beneficial as knowing the machine code. Who cares in 2026 other than really mission critical apps that AI is not ready for.
You’re absolutely right!
Now time for some pen-testing if you don’t mind sharing the links to your products.
/s
What you are doing is the coding equivalent of not reading the manual on a new appliance. It will work, but you will get surprised. Probably not in a good way.
That is the argument many make but it is assuming that any issues that come up AI will not be able to help which is just not true. I have a top rated, multiplayer VR app and I have not written code for it for 2 years now as AI does it all. Sure issues have come up, but AI can fix as long as you guide it correctly.
But you wrote the bulk of the code 2 years ago? If so you know 90% of the manual.
Most of the code was written in the last 2 years as I have expanded the app a lot, added multi player and reworked all the original code. For my simple app, AI has allowed me to grow much faster than before. I am very grateful for the tech even with its may issues.
I keep seeing the “it’s good for prototyping” argument they post here, in real life.
For non-coders it holds up if you ignore the security risk of someone running literally random code they have no idea what does.
But seeing it from developers, it smells of bullshit. The thing they show are always a week of vibing gave them some stuff I could hack up in a weekend. And they could too if they invested a few days of learning e.g. html5, basic css and read the http fetch doc. And the learning cost is a one-time cost - later prototypes they can just bang out. And then they also also have the understanding needed to turn it into a proper product if the prototype pans out.
I disagree. What I could hack over a weekend starting a project, I can do in a couple hours with AI, because starting a project is where the typing bottleneck is, due to all of the boilerplate. I can’t type faster than an LLM.
Also, because there are hundreds of similar projects out there and I won’t get to the parts that make mine unique in a weekend, that’s the perfect use case for “vibe coding”.
I would agree with that.
Especially, “being 70%” finished does not mean you will get a working product at all. If the fundamentale understanding is not there, you will not getting a working product without fundamental rewrites.
I have seen code from such bullshit developers myself. Vibe-coded device drivers where people do not understand the fundamentals of multi-threading. Why and when you need locks in C++. No clear API descriptions. Messaging architectures that look like a rats nest. Wild mix of synchronous and async code. Insistence that their code is self-documenting and needs neither comments nor doc. And: Agressivity when confronted with all that. Because the bullshit taints any working relationship.
It can pop out a pojo based on copy paste of an API document faster than I can.
I wouldn’t trust it for logic though. That’s just asking for trouble.
I say this as someone who’s not particularly a fan of AI and tries to use it very sparingly.
For me AI is not so much about productivity gains. Where I find it useful instead is to push me past the initial block of starting something from scratch. It’s that initial dopamine rush that the article mentions, from seeing an idea starting to take shape.
In that sense, if I compare projects by time spent on them with or without AI after they are completed, I too would probably find there were no productivity gains. But some of these things I would never get started at all by myself.
If you are a senior developer in a corporation, you know what you have to do, you are an expert in your domain, you rarely start something really new (and when you do, it is only after endless discussions and studies on tools, language, tech stack, architecture). AI is probably not a great help for you.
But even in corporate life, there are a lot of things that are inportant but that you constantly set aside: from planning your career, to honing your communication skills or whatever it is that you could certainly learn to do (with time and dedication) but for some reason you keep postponing because you are not already an expert at them and it takes motivation to learn. That’s where AI found its niche in my life.
There is no paradox. vibe coding = not writing code = not being a programmer.
so what is it then? i used llms to write the code for a feature complete desktop screen dimming application. did i produce it, if not develop it? am i just a… logic guide? legitimately asking because the program works better than any available alternative
I see myself as a project manager and executive producer. I know I don’t write that code, and I couldn’t even if I tried.
But I am skilled at directing and verifying, and this has allowed me to create (not code) a fairly complex WordPress plugin for course bookings with online card payment processing etc.
I have made a few manual tweaks here and there but that code is 98% Claude. And you know what? It works. That is good enough for me.
You’re dictating to and micromanaging an algorithm. Cringe individuals would refer to themselves as “Prompt Engineers” or something like that.
Instead of actually writing the code and understanding what each function actually does, you’re basically just skimming whatever the output is. You’re not gaining experience in programming by doing that and you may not be aware of what exactly everything does. Could be harmless until something unexpected starts causing issues.
For your specific case, an LLM seems completely overkill. I’ve also setup my desktop monitors to change brightness via a couple keyboard shortcuts using ControlMyMonitor and AutoHotkey. It’s like 10 lines of code.
my program is significantly better than either of those options, including fully functional hotkeys, hybrid gamma/overlay brightness logic, hdr detection, quicktray shortcut slider, a complete scheduling system, window behavior/startup options, and an attractive dark interface all in a tiny window.
and i did it in only 3 days. overkill? good, cuz i got the software i needed without knowing how to code in only 3 days. also, I’ve learned a ton about programming logic and code structure. it’s hard to tell an llm to do something you yourself don’t understand. at least if you want to build anything complex.
Is it though? Sure he could hire a programmer, but Claude is far less expensive. I agree with his position, he’s a project/product manager, not a programmer. And that’s okay sometimes.
You’re right, I apologize, it’s really 3 lines of code repeated over and over. It’s repeated for each monitor/action (brighten/dimming) being performed. The script file it technically 47 lines because I have 3 monitors and 10 different shortcuts.
Here’s the first action for the first monitor. Just edit the name and the brightness amount and voilà. The comment inline code formatting might put it on one line.
^!+PgUp:: Run, D:\Programs\ControlMyMonitor\ControlMyMonitor.exe /ChangeValue "MSI G274QPF" 10 20 returnTo be honest, I hesitate to even consider this example programming. The only thing it’s doing is executing a single command when a key combination is pressed. It doesn’t require really any programming experience. No loops, variables, scope, no time complexity, nothing.
Using an AI robbed them of learning even the smallest of concepts and they have not grown as a result.
It’s the same thing with any other concept; I don’t need to dedicate my life being an auto mechanic, but I should at least be able to know how to change a tire when I need to.
Is this the same fast to ship but hard to maintain argument we’ve seen a thousand times already?
It’s not a paradox, but a very typical result of using “fast” solutions.The main paradox here, seems to be: the 70% boilerplate head-start being perceived faster, but the remaining 30% of fixing the AI-introduced mess, negating the marketed time-savings; or even leading to outright counterproductivity. At least in more demanding environments, not cherry picked by the industry, shoveling the tools.
I’ll take that as a “Yes”.
