Skip to main content
Practical AI

Agentic Coding and the Economics of Open Source

48 min episode · 2 min read
·
Miklos Skoren

Episode

48 min

Read time

2 min

Topics

Fundraising & VC, Design & UX, Sales & Revenue

AI-Generated Summary

Key Takeaways

  • Open Source Attention Economy: GitHub stars serve as a measurable proxy for human engagement with open source projects. Skoren's controlled experiment across 100 websites and seven AI models found that as AI model recommendations for a library increase, NPM downloads rise 3–5 million per week, but GitHub stars stagnate or decline — confirming machines consume code without sustaining developer communities.
  • Tailwind CSS as Canary: Tailwind CSS experienced simultaneous surges in AI-driven downloads and sharp drops in website visits, directly damaging its premium-product revenue model. Skoren's team is now testing whether this pattern holds across front-end JavaScript packages broadly, using weekly-frequency NPM and GitHub data to determine if Tailwind is the rule, not the exception.
  • Open Source Requires Scale to Survive: Harvard Business School research cited by Skoren estimates open source creates over 1,000 times more value than the labor invested in it, making direct monetization structurally impossible. This means open source projects require millions of active human users to justify maintenance — a user base that AI agents erode by consuming packages without any human-developer interaction.
  • Developer Role Shifts to Design and Requirements: With code generation effectively automated, the two remaining high-value developer functions are translating user needs into system requirements and designing component architecture. Skoren recommends treating AI as a fast coworker rather than a tool — using voice recordings or rough notes to convey ideas, then letting agents handle implementation while humans retain ownership of problem framing.
  • Comparative Advantage Applies to Human-AI Collaboration: Drawing on Ricardo's comparative advantage principle, Skoren argues humans retain a structural edge in thinking and problem framing even if AI outperforms on execution. He restructured his own scientific workflow to eliminate nearly all manual coding, replacing it with analog thinking — pen, paper, books — then converting ideas into working code via brief verbal descriptions to AI agents.

What It Covers

Economics professor Miklos Skoren presents research on how AI-assisted "vibe coding" disrupts open source software ecosystems. Using incentive theory and empirical data from NPM downloads and GitHub stars across 100 representative websites tested against seven AI models, the paper argues human attention — the lifeblood of open source — is being systematically redirected toward machines.

Key Questions Answered

  • Open Source Attention Economy: GitHub stars serve as a measurable proxy for human engagement with open source projects. Skoren's controlled experiment across 100 websites and seven AI models found that as AI model recommendations for a library increase, NPM downloads rise 3–5 million per week, but GitHub stars stagnate or decline — confirming machines consume code without sustaining developer communities.
  • Tailwind CSS as Canary: Tailwind CSS experienced simultaneous surges in AI-driven downloads and sharp drops in website visits, directly damaging its premium-product revenue model. Skoren's team is now testing whether this pattern holds across front-end JavaScript packages broadly, using weekly-frequency NPM and GitHub data to determine if Tailwind is the rule, not the exception.
  • Open Source Requires Scale to Survive: Harvard Business School research cited by Skoren estimates open source creates over 1,000 times more value than the labor invested in it, making direct monetization structurally impossible. This means open source projects require millions of active human users to justify maintenance — a user base that AI agents erode by consuming packages without any human-developer interaction.
  • Developer Role Shifts to Design and Requirements: With code generation effectively automated, the two remaining high-value developer functions are translating user needs into system requirements and designing component architecture. Skoren recommends treating AI as a fast coworker rather than a tool — using voice recordings or rough notes to convey ideas, then letting agents handle implementation while humans retain ownership of problem framing.
  • Comparative Advantage Applies to Human-AI Collaboration: Drawing on Ricardo's comparative advantage principle, Skoren argues humans retain a structural edge in thinking and problem framing even if AI outperforms on execution. He restructured his own scientific workflow to eliminate nearly all manual coding, replacing it with analog thinking — pen, paper, books — then converting ideas into working code via brief verbal descriptions to AI agents.

Notable Moment

Skoren describes running a controlled experiment where AI models were given functional requirements for 100 real websites — with all brand and technology names stripped out — then asked to build them. The resulting dependency choices revealed which libraries AI systematically favors, independent of any human developer recommendation or documentation visit.

Know someone who'd find this useful?

Episode Transcript

Welcome to the Practical AI podcast, where we break down the real world applications of artificial intelligence and how it's shaping the way we live, work, and create. Our goal is to help make AI technology practical, productive, and accessible to everyone. Whether you're a developer, business leader, or just curious about the tech behind the buzz, you're in the right place. Be sure to connect with us on LinkedIn, x, or Blue Sky to stay up to date with episode drops, behind the scenes content, and AI insights. You can learn more at practicalai.fm. Now onto the show. Welcome to another episode of the Practical AI podcast. I'm Chris Benson, principal AI research engineer. And, with me today, I have a a guest I've been looking forward to for some time now. I have, doctor Miklos Skoren, who is a professor of economics at Central European University in Vienna, and he has written a really interesting paper on, on the effect of Vibe coding, on open source. So Welcome to the show. Really excited to have you here today. Thank you, and thanks for having me. Yeah. So I think this is a a slightly different, you know, take for us. We tend to, in the on the show, you know, leap straight into models and all sorts of stuff. But I know you're a professor of economics and, and you study incentive systems. And so I'm really interested in understanding how you turn that particular lens, of economics onto open source. And so I was wondering if for listeners, if you could, you know, kinda talk a little bit about what drew you into the the the notion of exploring open source, through that lens of yours, upfront? You know, what was the what was the first thing that said this is something that that we need to go study? Yeah. Let me let me, give a little bit of background on that. So, as an economist, I'm my research is really focusing on, competitiveness. So, what does it take for a company to be competitive in the marketplace, or what does it take for a country to be to be competitive? And for a long time, I've been really interested in, whether it's technology that makes a business succeed or whether it's a talent that that they have or maybe both or maybe there's some interaction between technology and and talent. And, I would also call myself kind of an accidental software developer in the sense that, economics is a very quantitative science, and there's a lot of, computational research involved. And but I was never trained, as a software developer, but, you know, we have to be, effective at using your computer. So and and this part, I actually enjoy at least as much as as, thinking about the economic incentives that you mentioned or the other parts of of the science. And so, the the story of this paper is, we've actually been thinking with …

Get the full transcript (8,443 words) + summary by email — free

One-time email with the complete transcript and AI summary of this episode. No account needed.

One email, no spam. We’ll also show you what SignalCast does.

Browse all Practical AI transcripts →

You just read a 3-minute summary of a 45-minute episode.

Get Practical AI summarized like this every Monday — plus up to 2 more podcasts, free.

Pick Your Podcasts — Free

Keep Reading

Books, tools, and gear mentioned in this episode

SignalCast may earn commission on purchases via these links.

Tools

  • Using incentive theory and empirical data from NPM downloads and GitHub stars across 100 representative websites tested against seven AI models.
  • Tailwind CSS experienced simultaneous surges in AI-driven downloads and sharp drops in website visits, directly damaging its premium-product revenue model.
  • Using incentive theory and empirical data from NPM downloads and GitHub stars across 100 representative websites tested against seven AI models.
  • SPONSORS: Prediction Guard

company

  • Harvard Business School research cited by Skoren estimates open source creates over 1,000 times more value than the labor invested in it.

More from Practical AI

We summarize every new episode. Want them in your inbox?

Similar Episodes

Related episodes from other podcasts

Explore Related Topics

This podcast is featured in Best AI Podcasts (2026) — ranked and reviewed with AI summaries.

You're clearly into Practical AI.

Every Monday, we deliver AI summaries of the latest episodes from Practical AI and 192+ other podcasts. Free for one show.

Start My Monday Digest

No credit card · Unsubscribe anytime