Original article by Vitalik Buterin
Original translation: Saoirse, Foresight News
In the world of free and open source software (and free content more generally), copyright licenses fall into two main categories:
If the content is published under a permissive license (such as CC0, MIT), anyone can obtain, use and redistribute the content without restriction, subject only to the minimum rules requiring attribution;
If the content is released under a copyleft license (such as CC-BY-SA, GPL), anyone can also obtain, use and redistribute copies without restriction, but if you create and distribute derivative works by modifying or combining with other works, the new work must be released under the same license. In addition, GPL also requires that any derivative work disclose its source code and other requirements.
In short: permissive licenses allow free sharing to everyone, while copyleft licenses are only shared with people who are also willing to share freely.
I have been a fan and developer of free and open source software and free content for as long as I can remember, passionate about building things that I think are useful to others. In the past, I preferred permissive licensing models (for example, my blog uses the WTFPL license), but recently I have gradually shifted to supporting the copyleft model. This article will explain the reasons for this change.
WTFPL advocates a concept of software freedom, but it is not the only paradigm.
Why I Used to Prefer Permissive Licenses
First, I want to maximize adoption and distribution of my work, and permissive licenses make it easy for anyone to build upon my work without any restrictions. Companies are often reluctant to make open source projects free, and I know I cant push them to fully switch to free software, so I want to avoid unnecessary conflict with their established models that they are unwilling to abandon.
Second, I have a philosophical aversion to copyright (and patents) in general. I dont agree that two people sharing a piece of data privately should be considered a crime against a third party. They neither touch nor even interact with the third party, nor deprive them of any rights (remember, unpaid is not the same as stealing). There are many legal considerations that make it operationally complicated to explicitly release a work into the public domain. A permissive license is the purest, safest way to get as close to no copyright assertion as possible.
I do appreciate the idea of copyleft, copyright by copyright, which I think is a brilliant legal ingenuity. In a sense, it is similar to the liberalism that I admire at the philosophical level. As a political philosophy, liberalism is often interpreted as prohibiting the use of any violence except to protect people from violence. As a social philosophy, I often see it as a way to tame the harm of human disgust reflex. It regards freedom itself as a sacred thing, making the behavior that tarnishes freedom an offensive existence. Even if you find voluntary non-conventional relationships between others uncomfortable, you cannot pursue them, because interfering with the private lives of free individuals is itself abominable. Therefore, in principle, there is no shortage of precedents in history that prove that aversion to copyright and the practice of copyright by copyright can coexist.
However, while copyleft for textual works meets this definition, GPL-style code copyright goes beyond the minimalist concept of copyright by copyright: it uses copyright for the aggressive purpose of forcing the disclosure of source code. This move is in the public interest rather than the selfish motive of making licensing fees, but it is still an aggressive use of copyright. This is even more obvious for more stringent licenses such as AGPL: even if derivative works are only provided through software as a service (SaaS) and never made public, they are still required to disclose the source code.
Different types of software licenses set different conditions for sharing the source code of derivative works. Some of these licenses require the source code to be made public in a wide range of scenarios.
Why copyleft is more popular nowadays
My shift from preferring permissive licenses to supporting copyleft stems from two major industry changes and a philosophical shift.
First, open source has become mainstream, and it is more feasible to encourage enterprises to embrace open source. Today, many companies in all walks of life are embracing open source: technology giants such as Google, Microsoft, and Huawei not only accept open source, but also lead the development of open source software; emerging fields such as artificial intelligence and cryptocurrency rely on open source to a greater extent than any previous industry.
Secondly, the competition in the crypto space is becoming increasingly fierce and profit-driven. We can no longer simply rely on people to voluntarily open source out of goodwill. Therefore, promoting open source cannot rely solely on moral appeals (such as please make the code public), but also requires the hard constraints of copyleft, which only open code permissions to developers who also open source.
A graphic that visualizes how these two forces increase the relative value of copyleft might look something like this:
The value of incentivizing open source is most evident in situations where it is neither completely impractical nor necessarily possible. This is the situation that the mainstream enterprise space and the crypto industry are in today, making the value of incentivizing open source through copyleft greatly increased.
(Note: The horizontal axis represents the motivation level for turning to open source, and the vertical axis represents the probability of open source. Comparing the two graphs, we can see that the motivation and effect of using copyleft to promote open source in mainstream fields are more likely to work together, while the encryption field has driven diminishing marginal benefits due to the maturity of the ecosystem, reflecting that the value logic of copyleft incentivizing open source has changed with the development of the industry.)
Third, Glen Weyl-style economic theory convinces me that in the presence of superlinear returns to scale, the optimal policy is not actually a Rothbard/Mises-style strict property rights system. Instead, the optimal policy does require a certain degree of proactive promotion of projects to make them more open than they would otherwise be.
Basically, if we assume that there are economies of scale, simple math shows that a non-zero degree of openness is the only way to avoid a world where one entity controls everything. Economies of scale mean that if I have twice as many resources as you, I can make more than twice as much progress. So by next year, I might have 2.02 times as many resources as you, and so on and so forth…
Left: Proportional growth model, small differences in the initial stage will remain small gaps in the end; Right: Scale economy growth model, small differences in the initial stage will evolve into huge gaps over time.
From a historical perspective, the key factors that prevented this imbalance from getting out of control are: humans cannot escape the diffusion effect of progress. Talents carry ideas and skills when they move between companies and countries; poor countries can achieve catch-up growth by trading with rich countries; and industrial espionage is widespread, making it difficult for innovation to be absolutely monopolized.
In recent years, however, multiple trends have threatened this balance and weakened traditional factors that have curbed unbalanced growth:
Technological progress is accelerating exponentially, and the speed of innovation iteration is far faster than ever before;
Political instability within and between countries is increasing: if the mechanism for protecting rights is perfect, the rise of others does not directly pose a threat; but in an environment where coercive behavior is more likely to occur and is unpredictable, the excessive strength of a certain entity will become a real risk. At the same time, governments are less willing to regulate monopolies than before;
Modern software and hardware products have closed capabilities: Traditional product delivery must be accompanied by technical transparency (such as reverse engineering), but now closed-source products can only open up the right to use, retaining the right to modify and control;
The natural limitations of economies of scale are weakened: Historically, large organizations have been limited by high management costs and difficulties in meeting local needs, but digital technology has made ultra-large-scale control systems possible.
These changes have exacerbated a persistent and even self-reinforcing power imbalance between corporations and the state.
Therefore, I increasingly agree that stronger measures are needed to actively incentivize or force the diffusion of technology.
Recent policies of various governments can be seen as coercive interventions in technology diffusion:
EU standardization directives (such as the latest mandatory USB-C interface) aim to break up closed ecosystems that are incompatible with other technologies;
China’s mandatory technology transfer rules;
The United States prohibits non-compete agreements (I support this policy because it forces companies to partially open source their tacit knowledge through the flow of talent. Although confidentiality agreements exist, their actual implementation is full of loopholes).
In my opinion, the downside of such policies often stems from their government-mandated nature, which leads them to preferentially incentivize the types of diffusion that are heavily skewed toward local political and commercial interests. But the upside of such policies is that they do incentivize higher levels of technology diffusion.
Copyleft builds a huge pool of code (or other creative works) resources, which can only be used legally if users are willing to share the source code of the content developed based on the resources. Therefore, copyleft can be regarded as a very universal and neutral incentive mechanism for technology diffusion, which can not only reap the positive effects of the above policies, but also avoid their many drawbacks. This is because copyleft does not favor any specific subject, and does not require central planners to actively set parameters.
These views are not absolute. In the maximum penetration scenario, permissive licenses are still valuable. But overall, the comprehensive benefits of copyleft are far greater than they were 15 years ago. Projects that chose permissive licenses back then should at least consider switching to copyleft now.
Unfortunately, the “open source” logo has become a thing of the past. But in the future, we may have open source cars, and copyleft hardware may help make that a reality.
