AI generated code is effectively plagiarized; it ground beef produced from stolen cattle.
An investor would be foolish to dump a lot of money into a startup after doing due diligence and discovering that they cannot legally put their copyright header on any of their shit, which means it's not actually an IP asset that they properly own.
All output of LLM's and diffusion models is clearly a derived work of the training data. It is radioactive.
In if you work in any sane software organization, by now you should have gotten a memo about not uploading your company IP into any AI dialog box in the cloud, or pasting generated code into company repos.
The last thing I'm interested in is what a table of multimillionaires has to say about the latest crappy social media tech fad.
They're saying it because they want it to be true; they want to make it true. They cannot resist the prospect of getting paid for every line of code produced by amateurs who regenerate their toy apps every time there's a bug.
Might as well be titled "Expertise is a thing of the past".
I watched the video, and am not super sure how to feel about it.
They talked about how great it is to go from 0 to 1, and how it changes the kind of person your startup hires, etc. They talk about how you don’t even bother debugging errors; you just throw away the code and have the AI give it another shot.
But then they say that these are useless for the 1 to n stage of your startup. So… does that mean you throw everything away, get rid of your employees, hire new ones and then get to work making something scalable???
Yeah, you don't have to put up with all the shade cast on someone who "just needs a technical co-founder". So negative. Just keep "re-rolling" long enough, and you can hire a lowly first employee instead and have them deal with the mess!
i watched 20 seconds of this and stopped it, then just ran a transcription of the video and read that instead.
When you don't have the sun-dappled conference room and professional sheen, just their words, these YC founders come across as quite shallow. They're only interested in making production cheaper. No fucks given about code quality. Look how disappointed they are that humans still need to debug.
"Oh, and interestingly, I guess one thing the survey did indicate is that this stuff is terrible at debugging. [...] There doesn't seem to be a way to just tell it debug. You were saying that you have to be very explicit, like as if giving instructions to a first time software engineer"
"And it's about, are you really, do you have taste and you want to solve product problems? Or are you an architect and you want to solve systems problems?"
"if you can just like rewrite a thousand lines of code in like six seconds, like why not?"
It wasn't too long ago much of the public was horrified to realize maybe everyone will need to "learn to code" in order to keep their jobs.
I think in many ways we're still there and "vibe coding" is the unproductive temper tantrum about it.
Software is only accelerating at eating the world. Sure AI is kind of useful, but I don't understand what all this fear is about. At once, code is both the requirements and implementation. Yes, everyone needs to at least understand their organization's code to not run a dysfunctional business. We've been here for a long time, actually. What's the problem with that? Finance is not the only language of business and never really was.
14 comments
[ 2.9 ms ] story [ 46.6 ms ] threadAn investor would be foolish to dump a lot of money into a startup after doing due diligence and discovering that they cannot legally put their copyright header on any of their shit, which means it's not actually an IP asset that they properly own.
In if you work in any sane software organization, by now you should have gotten a memo about not uploading your company IP into any AI dialog box in the cloud, or pasting generated code into company repos.
They're saying it because they want it to be true; they want to make it true. They cannot resist the prospect of getting paid for every line of code produced by amateurs who regenerate their toy apps every time there's a bug.
Might as well be titled "Expertise is a thing of the past".
They talked about how great it is to go from 0 to 1, and how it changes the kind of person your startup hires, etc. They talk about how you don’t even bother debugging errors; you just throw away the code and have the AI give it another shot.
But then they say that these are useless for the 1 to n stage of your startup. So… does that mean you throw everything away, get rid of your employees, hire new ones and then get to work making something scalable???
When you don't have the sun-dappled conference room and professional sheen, just their words, these YC founders come across as quite shallow. They're only interested in making production cheaper. No fucks given about code quality. Look how disappointed they are that humans still need to debug.
"Oh, and interestingly, I guess one thing the survey did indicate is that this stuff is terrible at debugging. [...] There doesn't seem to be a way to just tell it debug. You were saying that you have to be very explicit, like as if giving instructions to a first time software engineer"
"And it's about, are you really, do you have taste and you want to solve product problems? Or are you an architect and you want to solve systems problems?"
"if you can just like rewrite a thousand lines of code in like six seconds, like why not?"
I think in many ways we're still there and "vibe coding" is the unproductive temper tantrum about it.
Software is only accelerating at eating the world. Sure AI is kind of useful, but I don't understand what all this fear is about. At once, code is both the requirements and implementation. Yes, everyone needs to at least understand their organization's code to not run a dysfunctional business. We've been here for a long time, actually. What's the problem with that? Finance is not the only language of business and never really was.