Infrastructure, Common Goods, and the Future of Open Source Software

Over the past decades, open source software has evolved from a novelty—frightening to some, dazzling with promise for others—to the bedrock of nearly all software development. (Consider: Thirty years ago, almost no major programming language was free or open source. Now it is all-but-unthinkable to launch a new language which is closed source, still less to charge for a compiler license.) It is tempting to construe open source projects as commons: cultural and natural resources we all inherit. But they are not. They are infrastructure. It is time we started treating them accordingly. The future of open source software—indeed, the future of software—depends on it.

Transcript

1. Chris Krycho LambdaConf 2025 May 12, 2025 Infrastructure, Common Goods, and the Future of Open Source
2. The Log4j Incident (2021)
3. What happened—technically? The Log4j Incident (2021) • A foundationa l e l ement of the Java ecosystem for decades • Design a l l owed remote execution and data extraction
4. What happened—socially? The Log4j Incident (2021) • A l ibaba identi fi ed and reported the vu l nerabi l ity • “P l ease hurry up.” • The Apache Logging Services team con fi rmed it and then fi xed it • A team of 16 unpaid vo l unteers • Many of whom worked nights and a weekend to fi x it • …pure l y for professiona l pride. • The broader ecosystem scramb l ed to patch the vu l nerabi l ity • A ff ected 8% of packages on Maven
5. Log4j is a (digital) kind of infrastructure.
6. Log4j is a (digital) kind of common good.
7. Log4j was built by unfunded volunteers.
8. Open source software is infrastructure.
9. Open source software is weird infrastructure.
10. Infrastructure
11. What is Infrastructure?
12. What do these have in common? Examples of Infrastructure • Bridges, cana l s, roads • Freight—trains, boats, aircraft, the occasiona l horse and wagon • Power p l ants/e l ectrica l stations, p l umbing, sewage & water treatment • Increasing l y: computing
13. —Deb Chachra, How Infrastructure Works, p. 10 What makes infrastructure infrastructure? “All of the stu ff that you don’t think about,” turns out to be a surprisingly good starting point. For something to be considered infrastructure, its presence and characteristics are taken as a given.
14. —Deb Chachra, How Infrastructure Works, p. 10 What makes infrastructure infrastructure? “All of the stu ff that you don’t think about,” turns out to be a surprisingly good starting point. For something to be considered infrastructure, its presence and characteristics are taken as a given.
15. Software as Infrastructure
16. Everything I did not think about while building this talk Software as Infrastructure? • U l ysses for writing, Keynote for presenting • Many OSS l ibraries • Swift (OSS) and ObjC • C l ang (OSS) • LLVM (OSS) • macOS • Darwin • XNU (OSS) + FreeBSD (OSS)
17. LLVM and Clang Software as Infrastructure? • Origina l l y bui l t by fo l ks eventua l l y emp l oyed at App l e • Contributions from— • Indie hackers bui l ding their own l anguages • Goog l e (huge competitor to App l e!)
18. This contribution pattern is extremely common.
19. This contribution pattern is very strange.
20. Extremely common & very strange Contribution patterns • Not funded l ike infrastructure • Not funded by norma l market mechanisms
21. How Open-Source Works
22. How the sausage gets made How Open-Source Works • Who makes it? • Why do they make it? • Do they get paid for it?
23. How the sausage gets made How Open-Source Works • Who makes it? • Individua l s • Groups • Why do they make it? • Do they get paid for it?
24. How the sausage gets made How Open-Source Works • Who makes it? • Individua l s • Groups • Why do they make it? • Interest/passion/fun • Fixing a bug or sharing a so l ution • Pro fi t (?) • Do they get paid for it?
25. How companies produce open-source software How Open-Source Works • Not the main product • “Commoditizing a comp l ement” (undercutting competitors) • Part of the funne l • The main product • Dua l l icensing • Se l l ing support
26. The Infrastructure Assumption
27. —Deb Chachra, How Infrastructure Works, p. 10 Infrastructural systems are famously boring because the best possible outcome is nothing happening, or at least nothing unexpected or untoward.
28. Someone is responsible
29. Someone is on the hook
30. Is someone responsible or on the hook? The Infrastructure Assumption • No warranty c l ause • No shared responsibi l ity
31. Is someone responsible or on the hook? The Infrastructure Assumption • No warranty c l ause • No shared responsibi l ity
32. Is someone responsible or on the hook? The Infrastructure Assumption • No warranty c l ause • No shared responsibi l ity
33. Perceived responsibility vs. reality The Infrastructure Assumption Deve l opers have: • No idea who is using their code. • No l ega l re l ationship with those users. • No fi nancia l re l ationship with those users.
34. Fix it yourself
35. Build legal & fi nancial relationships with maintainers
36. Coordination problems The Infrastructure Assumption • How much is a fi x worth? • Which companies shou l d pay for the fi x? • And how much shou l d each company pay? • Who is responsib l e to co l l ect and distribute those funds?
37. Value
38. Infrastructure is a public good
39. —Deb Chachra, How Infrastructure Works, p. 105 [“Public good”] has a very speci fi c and distinct meaning: in contrast to a private good, a public good is nonrivalrous and nonexcludable. “Nonrivalrous” means that one person having access to or enjoying a good does not preclude other people from doing the same. “Nonexcludable” means that people can't be prevented from using and bene fi ting from it.…
40. —Deb Chachra, How Infrastructure Works, p. 105 Private goods have an obvious business model. Make the thing, sell the thing. There’s generally much less incentive for the private sector to produce nonexcludable goods, because if you can't prevent people from receiving a bene fi t and they no longer have a reason to pay for it, they presumably won't—the “free rider” problem. Too many free riders and too few paying customers means that the provider won't be able to make a pro fi t if their business model is based on selling access to that bene fi t.
41. We all bene fi t from infrastructure.
42. —Deb Chachra, How Infrastructure Works, p. 120 Infrastructural systems, by their nature, are more than just technical; they're inextricably social and political because they are intrinsically collective. … Because they incorporate nonmonetary externalities, both positive and negative, they can't easily be valued or assessed like a consumer good, where it’s “worth it” to buy something or not. So they don't lend themselves to decision-making that focuses solely on the costs or the returns on investment.
43. “value captured” & “value generated” are not value neutral terms.
44. metaphysical capitalism
45. How open-source software di ff ers from physical infrastructure What is striking and distinctive • Without exp l icit coordination—often without any coordination • Without pub l ic funding—often without any funding • How the free rider prob l em shows up
46. Open-Source Software is by nature a common good.
47. People’s time is not a common good.
48. Polity
49. —Webster’s Unabridged Dictionary, 1913 The form or constitution by which any institution is organized; the recognized principles which lie at the foundation of any human institution.
50. —Webster’s Unabridged Dictionary, 1913 The form or constitution by which any institution is organized; the recognized principles which lie at the foundation of any human institution.
51. There is no CE O of O pe n- S o u r c e S of t w ar e
52. “Show up and do things, if you want to.”
53. Anarchy and Anarchism
54. Power
55. Values
56. —Deb Chachra, How Infrastructure Works, p. 39 Like all technologies, these systems incorporate the values of their builders. Many infrastructural systems, like municipal water supplies, include ideals like serving the public good and universal provision, or meeting the basic needs of everyone in a community. But some of the values are far less laudable: the social and environmental costs of high-quality infrastructural systems for one group are often borne by others who are not in that group.
57. Open-source software as good infrastructure entails obligations.
58. The Future
59. —Deb Chachra, How Infrastructure Works, p. 39 Infrastructural systems are famously boring because the best possible outcome is nothing happening, or at least nothing unexpected or untoward.
60. —Deb Chachra, How Infrastructure Works, p. 10 Infrastructural systems are famously boring because the best possible outcome is nothing happening, or at least nothing unexpected or untoward. But nothing happens and nothing continues to happen as a result of su ffi cient attention, specialized care, and unceasing oversight. “Nothing happening” is usually the result of careful inspection schedules, preventive maintenance, and planned replacement, all of which require resources to be devoted to to what will be, at best, a null outcome.
61. —Deb Chachra, How Infrastructure Works, p. 10 Infrastructural systems are famously boring because the best possible outcome is nothing happening, or at least nothing unexpected or untoward. But nothing happens and nothing continues to happen as a result of su ffi cient attention, specialized care, and unceasing oversight. “Nothing happening” is usually the result of careful inspection schedules, preventive maintenance, and planned replacement, all of which require resources to be devoted to to what will be, at best, a null outcome.
62. —Alan Jacobs, Attending to Technology: Theses for Disputation, The New Atlantis, Winter 2016 If instead of thinking of the Internet in statist terms we apply the logic of subsidiarity, we might be able to imagine the digital equivalent of a Mondragon cooperative.
63. —Alan Jacobs, Attending to Technology: Theses for Disputation, The New Atlantis, Winter 2016 If instead of thinking of the Internet in statist terms we apply the logic of subsidiarity, we might be able to imagine the digital equivalent of a Mondragon cooperative.
64. —Alan Jacobs, Attending to Technology: Theses for Disputation, The New Atlantis, Winter 2016 If instead of thinking of the Internet in statist terms we apply the logic of subsidiarity, we might be able to imagine the digital equivalent of a Mondragon cooperative.
65. —Alan Jacobs, Attending to Technology: Theses for Disputation, The New Atlantis, Winter 2016 If instead of thinking of the Internet in statist terms we apply the logic of subsidiarity, we might be able to imagine the digital equivalent of a Mondragon cooperative.
66. —Alan Jacobs, Attending to Technology: Theses for Disputation, The New Atlantis, Winter 2016 If instead of thinking of the Internet in statist terms we apply the logic of subsidiarity, we might be able to imagine the digital equivalent of a Mondragon cooperative.
67. Foundations Collectives
68. Foundations Collectives
69. Foundations Collectives
70. Conclusion
71. Chris Krycho LambdaConf 2025 May 12, 2025 Infrastructure, Common Goods, and the Future of Open Source