Why are some disciplines open and others not?
Georg Cantor, a 19th century mathematician, was driven insane by contemplating the mathematical theory of infinity. It seems we find it excruciating to fit infinity into our finite minds.
I sometimes feel a tinge of insanity when contemplating open-source business models. If a piece of open-source software's (OSS) value is defined as the labor cost required to write it divided by the money people pay to use it, then the value of OSS always tends to infinity for its users, and towards zero for the producers. When we think of OSS, we often imagine a lone university professor sitting in his office, frantically writing code. He loves his software and he wants other people to love it as well. There is no expectation of payment.
It seems, however, that the vast majority of open-source software being developed today is by well-funded organizations with commercial intent. PyTorch, the AI tensor calculation library, was hatched at Facebook. Pulumi, a company that creates open-source Infrastructure as Code (IaC) solutions, raised $37.5 million in 2020. And Taichi, an open source library used for speeding up Python, raised $50 million last year.
The trend of OSS being developed on the back of enormous pools of capital seems to be the norm in AI and cloud computing. Many other disciplines don't seem to be trending in this direction. Examples of such fields are actuarial pricing software, and anything related to the hard engineering disciplines like mechanical and civil engineering. The OSS available to these fields adhere more to the "mad professor" origin story.
Why are the two fields so different in this regard? Why are some companies funding software and giving it away for free, while others only invest in paid software? One explanation is that some fields are more difficult "solve" than others. Some software is therefore worth paying for and others not. It is difficult to believe this explanation in a post GPT world. Artificial General Intelligence has long been punted as one of the most difficult challenges for humanity to solve. We are on the cusp of solving it, mostly using OSS.
I believe the hard disciplines are simply fortunate that VC funding has not been focusing on them in recent years. It’s almost like a police investigation deciding to focus on other, sexier suspects, and the real perpetrators are getting away with it. They can therefore continue charging money for software because good open-source alternatives don't exist.
Wikipedia was founded in 2001, and the last hard copy of Encyclopedia Brittanica was printed approximately one decade later in 2012. Maybe all software is going to be free one day, but most "Wikipedia" type disruptors have not reared their heads yet.
We are in the age of last movers.
Wikipedia is the last mover in the encyclopedia industry. There is no incentive for anyone to compete with them. I believe every information-based industry is going to have a last mover. A general prediction is that the larger the industry, the sooner this is going to happen.
