The article the youtube video is talking about: Vibe Coding Is Killing Open Source Software, Researchers Argue
The paper the 404 media article is talking about: https://arxiv.org/pdf/2601.15494
The paragraph that 404 media wants to highlight:
“Our main result is that under traditional OSS business models, where maintainers primarily monetize direct user engagement…higher adoption of vibe coding reduces OSS provision and lowers welfare,” the study said. “In the long-run equilibrium, mediated usage erodes the revenue base that sustains OSS, raises the quality threshold for sharing, and reduces the mass of shared packages…the decline can be rapid because the same magnification mechanism that amplifies positive shocks to software demand also amplifies negative shocks to monetizable engagement. In other words, feedback loops that once accelerated growth now accelerate contraction.”
My key takeaway: If you see a “buy me a coffee” button, click on it and make it a $1 subscription.
Probably not the case tbh.
Open source has a lot of flaws and I am not sure that AI/vibe-coding will make things worse out of the box.
Can do both greater and worse, really depends on how it’s used.
But they are specifically tied to go bad together.
Sure, more scrutiny will now be needed by the end user but developers already do the checkups before using a tool.
Did you just discount 4 academics and their research paper using vibes?
Wouldn’t dare doing that.
I mostly saying that a white paper doesn’t mean it’s the holy truth.
I haven’t checked who they are, maybe they do FOSS but I haven’t be able to quickly find out but if they’re outside of the industry, I am not sure how valuable their opinion is.
Mine is subjective for sure but I also been in the thick of it, seen friends around me and experienced it myself over the years. So yeah, just my personal 2cts.
At the same time, I do know amazing people.
They are even more productive as of today and still continue to be self-sustained thanks to their FOSS work, while using AI daily to speed up their workflow.
Without any drop in quality or anything alike.
FOSS was anyway, hard to study in a lab.
And most people succeeding in the fields are abnormalities, not the norm.
But sure, AI is also adding a lot of noise/slope and nonsense to the daily work of tons of people. Nothing new here, things are just going as they always have been, just faster it being in the wrong or correct direction.
But anyway, FOSS is a huge topic and I might agree/disagree with specific parts of the statement given a specific scope.
Even more, I wouldn’t call every Twitter vibe-coder hype bro to be a FOSS contributor of any sort so we even need to agree on the definition of open source here: no, having a public repo on Github doesn’t make you a FOSS maintainer out of thin air.
What are your disagreements with the 404 article?
I don’t know the content of it, it’s paywalled. ![]()
Also, I’m pretty not willing to write more about that topic on here because it will pretty much lead me to write an article for 5 hours and nobody will agree with it.
FOSS is probably one of the biggest and deepest topic of disagreement ever.
I mean, now we also do have AI.
So yeah, mixing both topics and bikeshedding about all of that with random people in a written form is definitely not on my list of things I do wish to spend time on.
I already need to push back hard enough on a daily basis and it gets quite tiresome to slow people down…with AI adoption.
Quickly glancy over the video’s screenshots:
Yes, AI doesn’t give anything back. Nothing new.
I do have a few thousands of thorough answers on StackOverflow that nobody truly gave me anything in return[1]. And some of my answers today are probably coming from data trained on my own answers posted there publicly.
As for Adam Wathan, I’ve been an early supporter (bought the book), saw the rise and downfall of Tailwind Labs. I listened to his interview on Primeagen’s channel too and it was very saddening to see that happen.
The business model is also just not sustainable because no Twitter hype-bro will ever pay the project in any form or shape. Even less when it comes down to actual businesses.
Again, begging for donations is usually not a way to go either for those projects either: this is just not how software works, especially nowadays (value is pretty much down to zero).
Been there myself (trying to support some of the people I do/was rely(ing) on).
It’s quite a captain obvious situation here overall.
TLDR: nothing new for people in the field and shallow of a research as a whole.
nvm I’m lying, I got 60€ over my 5 years of helping there fulltime ↩︎
Just quickly throwing in the non-paywalled version of the 404media link using archive:
https://archive.ph/sgl5M
curl stopped its bug-bounty program because it received too many AI-generated reports.
The never-ending slop submissions take a serious mental toll to manage and sometimes also a long time to debunk. Time and energy that is completely wasted while also hampering our will to live.
I saw months ago yes.
Nothing specific tho.
Again, AI just makes things move faster.
Beg-bounties[1] were already a thing, now you just have more of them.
You can still have a few nice developers doing nice stuff with AI IMO.
This one hurts particularly
AI is the ultimate rent seeker, a middle-man that inserts itself between a creator and a user and it often consumes the very thing that’s giving it life. The OSS/vibe-coding dynamic is playing out in other places. In October, Wikipedia said it had seen an explosion in traffic but that most of it was from AI scraping the site. Users who experience Wikipedia through an AI intermediary don’t update the site and don’t donate during its frequent fund-raising drives.
But again, nothing new here either.
It’s already happening with books, music and art hence surprised Wikipedia wasn’t a victim sooner if we’re honest.
as one of my colleague calls them ↩︎