Vikash Agrawal

If you visit Europe, cultural surplus is everywhere: small towns that built towering churches, merchants who funded fanciful art they just installed in the market square. It is a form of what I call ‘cognitive surplus’. Surplus takes the following meaning: you produce more than you need today, so you attempt to store it. We are all familiar with ‘financial surplus’: you store part of what you earn so that you can consume it later when you are sick or can’t work. Open source is a form of cognitive surplus. To twist @esrtweet ’s words around, open source is akin to a cathedral. Instead of acquiring financial or military capital, we build this tremendously expensive collection of software. And it is expensive: according to Santiago Calderón et al. (2022), U.S. investment in open source in 2019 was about $37–38 billion. On the short term, it makes little sense on paper. Why would anyone work hard to build fancy software that everyone can just enjoy without paying? Yet, if you look around, you see plenty of engineers working in corporations who are in silly jobs. Or you have successful entrepreneurs who have plenty of time for side projects. You have young people who are underemployed, incapable of finding work matching their skills. Open source offers a way to channel this productivity and make something useful out of it. Ryan Fleury says that it makes software financially unsustainable. He is right but only in the sense that building cathedrals in small towns was unsustainable. Chromium is unsustainable. It makes no sense for a company like Google to spend a billion dollars a year (reportedly) to fund a web engine you can just fork and use for free. And yet, it is being sustained (for now). Fleury is likely correct that open source drastically limits the price of software: Hoffmann et al. (2024) estimate that it cuts the price of software by a factor of about 3.5. But we should be careful: it does not follow that software programmers would earn 3.5 more money without open source. Software has been eating the world in part thanks to open source software... without it, it might be a much smaller industry. I cannot predict the future but I have proposed that AI coding changes the equation somewhat and might redirect the software cognitive surplus. It is possible that we might stop investing so much in the open source cathedrals. It is becoming increasingly visible that the code crafted with love by some poor engineer in Nebraska is getting milked by large corporations without credit, through AI. Furthermore, it is becoming easier to just automate the boilerplate code, arguably easier to do so than to download an open source dependency.