H1
Scale mismatch
Norms built for comparable actors. Scale broke the social contract.
Strongly validated
Prosperous Software Movement · PPL v1.0
Software success should fuel its infrastructure.
A third phase for the commons: a license that routes a fraction of commercial revenue back to the dependencies it was built on.
>99%
of companies run on open source
70%+
of production code comes from it
$100T+
economy built on that foundation
>30 years
when the MIT license was written (#1 on GitHub)
§ 01 · The Structural Mismatch
A $100 trillion economy, run on licenses written for a different era.
The software commons compounded for forty years. The legal instruments that govern it did not. MIT was written in 1988, the GPLv3 in 1989 — for a world of comparable actors, not trillion-dollar firms.
World GDP vs. license evolution · 1988–2026navy bars · world GDP
◦ MIT 1988
◦ GPLv3 1989
◦ Apache v2 2004
$100T
1988
1992
1998
2004
2010
2016
2022
2026
◦ license layer — effectively frozen since 1989
§ 02 · Why The Model Breaks
Five structural failures.
Not a culture problem. The funding gap is the predictable output of five conditions operating at once — each stands alone; together they are hard to dismiss.
H2
Maintainer crisis
The volunteer pool is exhausted. Donations can't hire their way out.
Validated
H3
Dependency graph
Donations reach the visible layer. The critical layer is invisible.
Validated
H4
VC lifecycle gap
Startups lean hardest on OSS exactly when they have the least resources to donate.
Hypothesis
H5
The AI economy
The entire corpus of code, ingested to train models worth hundreds of billions.
Emerging
§ 03 · The Dependency Graph
Donations reach what you can name.
Companies fund the frameworks they import — the visible layer. The hundreds of transitive packages beneath them, the ones that actually hold up the internet, receive functionally nothing.
Where the money lands · one codebase
Visible layer · funded
The few dependencies you can name.
Transitive layer · unfunded
~512
packages deep, with no visible face.
Case · curl
1Bdevices
47 car brands. 1 full-time maintainer. Brands that contribute back: the list is empty.
Case · core-js
40Mweekly downloads
Inside Apple, Amazon, Netflix. Effectively unfunded — almost no one pays for it.
§ 04 · Why Donations Fall Short
Goodwill isn't the bottleneck. Structure is.
Babel had Airbnb, Facebook, Salesforce and Discord as sponsors — and still couldn't pay three engineers. Donation is simply the wrong mechanism for the scale of the problem.
Babel · annual funding need vs. received
$333k
needed
~$166k
received
117M monthly downloads. Still couldn't sustain three engineers at market rate.
Who actually sponsors · GitHub Sponsors
4,200corporate
sponsors
sponsors
out of an estimated
300,000,000
companies using open source
≈ 0.0014% participation
§ 05 · The Fix · Profit-Left
If revenue > X, share Y% back.
One clause, pre-specified and universal. No bilateral deals, no permission. The obligation scales with success and flows through the dependency graph automatically.
01
License with PPL
The author adds a profit-left clause and picks a rate — PPL-1% or PPL-5%.
→02
Use it freely
Below $1M revenue: no obligation. Students, startups, hobbyists build unencumbered.
→03
Cross the threshold
Past $1M, the clause activates — one year to begin contributing.
→04
Funds flow & forks inherit
Routed algorithmically through the whole graph. Every fork carries the clause.
§ 06 · Three Phases Of The Commons
Three movements. One unfinished argument.
1980s
Phase I
Free Software
GNU · GPL
Improve software freedoms.
1990s
Phase II
Open Source
OSI · Apache · MIT
Improve collaboration.
2026
Phase III
Prosperous Software
PPL v1.0 · PGF
Improve sustainability.