I set the foundation for a Telegram assistant, a web app, and a desktop app, while ditching Figma, Notion, Slack and Pipedrive. Not bad for a fistful of tokens
The amount of bugs and tech debt boggles the mind here
I don't understand why so many VCs fall for "not invented here". Vibe-coded or not, this is just another in-house solution, inferior and more expensive than most out-of-the-box products already out there.
> Overall, LLMs aren’t yet at the point where they can replace all engineers. But I don’t doubt they will be soon enough, and in the meantime, they just can’t be ignored.
The moment LLM's can replace engineers do we need VCs? Because any one at any time can will any application they want into existence.
> Overall, LLMs aren’t yet at the point where they can replace all engineers. But I don’t doubt they will be soon enough, and in the meantime, they just can’t be ignored
Someone has to maintain that code and there is not a single mention of that caveat after the software is built in the article and 99.9999% of the most widely used software is maintained by humans - even if parts of it was vibe-coded, a human has to maintain it so that it functions correctly, especially a must if it is mission critical software.
It is like we are celebrating mediocrity under the guise of AI and rebranding 'prototyping' as 'vibe-coding' but worse - software with fast accumulating technical debt, slapping on third-party risks and close to no tests at all.
Eventually, someone has to maintain that software and surely 9 times out of 10, a typical senior software engineer will look at that vibe-coded slop and will either throw it all away or reduce these third party services with existing robust open source versions.
Vibe-coding gets you faster to maintain the same negatives from traditional software engineering without you understanding what you are doing.
> Overall, LLMs aren’t yet at the point where they can replace all engineers. But I don’t doubt they will be soon enough.
All engineers? This doesn't match my hands-on experience at all.
If you give a chainsaw to everyone, it doesn't make everyone a lumberjack. And a chainsaw itself certainly isn't a lumberjack.
If you give Claude Code or the like to everyone, it's doesn't make everyone a highly skilled software engineer. And Claude code itself isn't a highly skilled software engineer.
I've come around to this view. When I first began using these things for building software (the moment ChatGPT dropped), I was immediately skeptical of the view that these things are merely glorified autocomplete. They felt so different than that. These computers would do what I _meant_, not what I _said_. That was a first and very unlike any autocomplete I'd ever seen.
Now, with experience using them to build software and feeling how they are improving, I believe they are nothing more or less than fantastically good auto complete. So good that it was previously unimaginable outside of science fiction.
Why autocomplete and not highly skilled software engineer? They have no taste. And, at best, they only pretend to know the big picture, sometimes.
They do generate lots of code, of course. And you need something / someone that you can trust to review all that code. And THAT thing needs to have good taste and to know the big picture, so that it can reliably make good judgement calls.
LLMs can't. So, using LLMs to write loads of code just shifts the bottleneck from writing code to reviewing code. Moreover, LLMs and their trajectory of improvements do not, to this experienced software engineer, feel like they are kind of solution and kind of improvements needed to get to an automated code review system so reliable that a company would bet its life on it.
Are we going to need fewer software engineers per line of code written? Yes. Are lines of code going to go way up? Yes. Is the need for human code review going to go way up? Yes, until something other than mere LLM improvements arrive.
Even if you assume 100% of code is written by LLMs, all engineers aren't going to be replaced by LLMs.
tbh, it could be a breakthrough in model design, an smart optimization (like the recent DeepConf paper), a more brute force (like the recent CodeMonkeys paper), or a completely new paradigm (at which point we won't even call it an LLM anymore). either way, I believe it's hard to claim this will never happen.
(author here) in a way, yes.
you can see how I start with a god-mode euphoria, and slowly realise the vibe is mostly for boilerplates, and very (very) basic tasks.
"I freestyled with absolutely zero planification, and I'm paying the cost now as the product inches toward a production-ready form."
I'd say it helped me transition back into building, and even tho it can be (extremely) painful at times, I still enjoy ‘coding’ this way. These days I probably decide on and closely monitor 95% of the backend (and I'd probably be much faster coding some parts myself), but I still let Claude handle about 80% of the React bullshit I hate.
12 comments
[ 3.3 ms ] story [ 36.9 ms ] threadThe amount of bugs and tech debt boggles the mind here
The moment LLM's can replace engineers do we need VCs? Because any one at any time can will any application they want into existence.
Someone has to maintain that code and there is not a single mention of that caveat after the software is built in the article and 99.9999% of the most widely used software is maintained by humans - even if parts of it was vibe-coded, a human has to maintain it so that it functions correctly, especially a must if it is mission critical software.
It is like we are celebrating mediocrity under the guise of AI and rebranding 'prototyping' as 'vibe-coding' but worse - software with fast accumulating technical debt, slapping on third-party risks and close to no tests at all.
Eventually, someone has to maintain that software and surely 9 times out of 10, a typical senior software engineer will look at that vibe-coded slop and will either throw it all away or reduce these third party services with existing robust open source versions.
Vibe-coding gets you faster to maintain the same negatives from traditional software engineering without you understanding what you are doing.
All engineers? This doesn't match my hands-on experience at all.
If you give a chainsaw to everyone, it doesn't make everyone a lumberjack. And a chainsaw itself certainly isn't a lumberjack.
If you give Claude Code or the like to everyone, it's doesn't make everyone a highly skilled software engineer. And Claude code itself isn't a highly skilled software engineer.
I've come around to this view. When I first began using these things for building software (the moment ChatGPT dropped), I was immediately skeptical of the view that these things are merely glorified autocomplete. They felt so different than that. These computers would do what I _meant_, not what I _said_. That was a first and very unlike any autocomplete I'd ever seen.
Now, with experience using them to build software and feeling how they are improving, I believe they are nothing more or less than fantastically good auto complete. So good that it was previously unimaginable outside of science fiction.
Why autocomplete and not highly skilled software engineer? They have no taste. And, at best, they only pretend to know the big picture, sometimes.
They do generate lots of code, of course. And you need something / someone that you can trust to review all that code. And THAT thing needs to have good taste and to know the big picture, so that it can reliably make good judgement calls.
LLMs can't. So, using LLMs to write loads of code just shifts the bottleneck from writing code to reviewing code. Moreover, LLMs and their trajectory of improvements do not, to this experienced software engineer, feel like they are kind of solution and kind of improvements needed to get to an automated code review system so reliable that a company would bet its life on it.
Are we going to need fewer software engineers per line of code written? Yes. Are lines of code going to go way up? Yes. Is the need for human code review going to go way up? Yes, until something other than mere LLM improvements arrive.
Even if you assume 100% of code is written by LLMs, all engineers aren't going to be replaced by LLMs.
Is this your firm's investment thesis?
"I freestyled with absolutely zero planification, and I'm paying the cost now as the product inches toward a production-ready form."
I'd say it helped me transition back into building, and even tho it can be (extremely) painful at times, I still enjoy ‘coding’ this way. These days I probably decide on and closely monitor 95% of the backend (and I'd probably be much faster coding some parts myself), but I still let Claude handle about 80% of the React bullshit I hate.