The premise is flawed. What makes companies valuable is not having engineers with the ability to write good code.
Also
> Alas. Because, and I cannot stress this enough, I am not an engineer.
And then...
> Though it’s hard to benchmark how far I got in two days, this is my best guess: The app is roughly equivalent to what a designer and a couple professional engineers could build in a month or two.
"I have an idea for how to make a note-taking app that’s actually right"
I'm reminded of the bell-curve image that starts and ends with Apple Notes or a simple paper notebook. I always get the feeling that people who think that the technology is limiting their note taking abilities are looking for the wrong solution.
>Though it’s hard to benchmark how far I got in two days, this is my best guess: The app is roughly equivalent to what a designer and a couple professional engineers could build in a month or two.
As the author suggests, all kinds of domain experts are going to be freed to use a computer to solve their own problems, without having to first communicate them to a developer or team.
Plus, everything needed to deploy it and monetize has been turned into APIs which the LLMs know, so this can be added too.
A couple of decades ago a colleague had a great idea for a website called "negative knowledge" where people would dump all of their ideas that had failed, because there's generally more useful information from failure than there is from success given that it's easier to tell why something failed rather than why something succeeded.
> more useful information from failure than there is from success
This is massively true. It's why I'm much more interested in hearing stories of business failure than business success: it's trivially easy to replicate a failure mode, but success isn't so transitive.
> I have not failed. I've just found 10,000 ways that won't work.
As to the topic at hand, sounds a bit like hyperbole. AI tools are gonna make big changes, but there’s a hell of a lot more to running companies, and shipping products, than writing code. Some stuff will be enabled and multiplied by AI, some, not so much.
When you don't know what a profession does, you trivialize it.
This person isn't an engineer. They think that they made something reasonable in a day or two that an engineer would take a month to do.
This is exactly how AI people look at say radiologists. Oh, they'll be replaced by some computer vision, they just look at pictures and say "good or bad", right?
This is why all of those surveys where people are asked how is AI going to impact some profession are garbage. They mostly ask people who don't know what that profession is, to guess.
coding using llms, ai vibe coding, whatever you may call it, is the new "have fun staying poor" crypto bro wave. It's not new either. It should work well enough for generic stuff. Nothing new to come of it, though.
I’ve been looking for the perfect word for what AI allows me to do and I think manifest is it. I can conjure things and it feels like magic. Like poof here it just is. Like it was always there and I’m just noticing it.
It is not just that. He is making the same Twitter claims while showing nothing. He hypes you up about how AI did the app but then just close it with "oh that's not the idea i want it, hehe". Maybe show us what cursor produced?
I suspect AI does technically write 80% of code I currently generate (mostly writing Go probably inflates that number quite a bit). On a lot of lines I type a couple characters and the AI can complete that line, often times a few lines. But I have to write the strategically important characters, and I fix things when the autocomplete is wrong. That translates to maybe 20-30% increase in productivity because I’m not a coding monkey and my job isn’t just typing.
If I ask the AI to generate any 80% of my code without my experienced programmer handholding on every line/few lines, it’s more likely than not hot garbage that takes me longer to fix than writing from scratch. I’ve also generated entire small projects that appear to work, but the underlying code is so poor it’s gonna implode if you keep iterating and adding complexity.
The premise is flawed. What makes companies valuable is not having engineers with the ability to write good code.
Also
> Alas. Because, and I cannot stress this enough, I am not an engineer.
And then...
> Though it’s hard to benchmark how far I got in two days, this is my best guess: The app is roughly equivalent to what a designer and a couple professional engineers could build in a month or two.
"I have an idea for how to make a note-taking app that’s actually right"
I'm reminded of the bell-curve image that starts and ends with Apple Notes or a simple paper notebook. I always get the feeling that people who think that the technology is limiting their note taking abilities are looking for the wrong solution.
>Though it’s hard to benchmark how far I got in two days, this is my best guess: The app is roughly equivalent to what a designer and a couple professional engineers could build in a month or two.
Comedy
He did say it was the best guess. Not that it was accurate by any means.
As the author suggests, all kinds of domain experts are going to be freed to use a computer to solve their own problems, without having to first communicate them to a developer or team.
Plus, everything needed to deploy it and monetize has been turned into APIs which the LLMs know, so this can be added too.
Can’t wait to see what this new era gives us!
Too much hype x hyperbolic headline = lots of typing to explain why it’s wrong.
So much energy burnt to this.
“My long-loved thesis, when rendered on a screen, was catastrophically bad”
This is no small thing, you had an idea, you were able to try it out, and found it didn’t work.
But this does give you an understanding, and you no longer need to keep wondering if your idea is the answer.
A couple of decades ago a colleague had a great idea for a website called "negative knowledge" where people would dump all of their ideas that had failed, because there's generally more useful information from failure than there is from success given that it's easier to tell why something failed rather than why something succeeded.
> more useful information from failure than there is from success
This is massively true. It's why I'm much more interested in hearing stories of business failure than business success: it's trivially easy to replicate a failure mode, but success isn't so transitive.
There’s that old quote, attributed to Edison:
> I have not failed. I've just found 10,000 ways that won't work.
As to the topic at hand, sounds a bit like hyperbole. AI tools are gonna make big changes, but there’s a hell of a lot more to running companies, and shipping products, than writing code. Some stuff will be enabled and multiplied by AI, some, not so much.
We’ll see what happens. The die is cast.
When you don't know what a profession does, you trivialize it.
This person isn't an engineer. They think that they made something reasonable in a day or two that an engineer would take a month to do.
This is exactly how AI people look at say radiologists. Oh, they'll be replaced by some computer vision, they just look at pictures and say "good or bad", right?
This is why all of those surveys where people are asked how is AI going to impact some profession are garbage. They mostly ask people who don't know what that profession is, to guess.
coding using llms, ai vibe coding, whatever you may call it, is the new "have fun staying poor" crypto bro wave. It's not new either. It should work well enough for generic stuff. Nothing new to come of it, though.
Manifest.
I’ve been looking for the perfect word for what AI allows me to do and I think manifest is it. I can conjure things and it feels like magic. Like poof here it just is. Like it was always there and I’m just noticing it.
I haven't been able to vibe code Rust. It must be just a React / JavaScript thing.
It is not just that. He is making the same Twitter claims while showing nothing. He hypes you up about how AI did the app but then just close it with "oh that's not the idea i want it, hehe". Maybe show us what cursor produced?
I write tonnes of React / JS. Vibe coding seems to be a cult I am not a part of.
I snicker when I hear "AI writes 85% of the code in our codebase". It has never worked for me.
I suspect AI does technically write 80% of code I currently generate (mostly writing Go probably inflates that number quite a bit). On a lot of lines I type a couple characters and the AI can complete that line, often times a few lines. But I have to write the strategically important characters, and I fix things when the autocomplete is wrong. That translates to maybe 20-30% increase in productivity because I’m not a coding monkey and my job isn’t just typing.
If I ask the AI to generate any 80% of my code without my experienced programmer handholding on every line/few lines, it’s more likely than not hot garbage that takes me longer to fix than writing from scratch. I’ve also generated entire small projects that appear to work, but the underlying code is so poor it’s gonna implode if you keep iterating and adding complexity.
[dead]