Today, if you want to build a new internet services, very often you will need to incorporate user data, user relationships or functionality contained and corralled within monopolistic established services owned and run by Big Tech. Essentially, this means that to function correctly, new services you or others build often depend heavily upon access to APIs (Application Programming Interfaces) provided by companies such as Microsoft, Google, Amazon and Facebook. Sadly, the history of the past decade shows that building on these APIs is like building on sand, which is why we are seeing fewer and fewer interesting new startup services being launched and succeeding.
The Internet Computer provides a means to create internet services in a totally new way using “autonomous software”. These internet services can provide guarantees to users about how their data is processed, but even more importantly, can provide guarantees about the availability of the APIs they provide to other services that wish to incorporate the data or functionality they share. Ultimately, this will enable entrepreneurs and communities of developers to create a new breed of service that benefits from better network effects. These network effects will be mutualized and allow vast numbers of services to be built out collaboratively because they can build on and extend each other without trust, creating a far more diverse, dynamic, fertile and eventually dominant ecosystem.
The essential problem has been demonstrated many times. A harbinger was provided by Zynga, the successfully social games company. Zynga built on Facebook and managed to IPO and gain a valuation of almost $15 billion until Facebook changed the APIs it depended upon causing its share price to plummet 85% in just three months. Another example is LinkedIn, which provided APIs that allowed thousands of startups and other services to incorporate professional profiles created by people around the world. However, once Microsoft purchased LinkedIn, it revoked the APIs being used by startups and other smaller services, leaving only other Big Tech corporations such as Salesforce to continue business as normal. With lessons like these it is no wonder that venture capitalists don’t want to fund new startups building on Big Tech. In fact, at the time of writing, 17 of the last 22 technology startup IPOs in the USA specified “platform risk” as existential threats to their business in their S-1 filings. Meanwhile, the monopolization of the Internet and the power of Big Tech continues to grow apace.
These big monopolies are squeezing out opportunity and creativity. To create a more vibrant ecosystem we have to create a new environment in which it can grow. The Internet Computer provides such a place, by supporting the creation of “open internet services” using “autonomous software”, which runs without an owner. This new kind of internet service addresses the problem with platform risk head on, because it allows internet services to guarantee the availability of APIs to other services, allowing entrepreneurs and developer communities to build out synergistic new services without risk that the rug will be pulled from under them. Today, at the time of writing, there are billions of dollars waiting to fund this new more dynamic and fairer internet economy.
But how will one of these new open internet services be created? For example, let’s imagine some entrepreneur wishes to create a replacement for LinkedIn on the Internet Computer. The first step they would probably take, is to create a basic prototype, since LinkedIn is no marvel of engineering, and then upload the software “canisters” involved to the Internet Computer. At this point, the entrepreneur retains control over the service because they have a cryptographic key that allows them to configure and upgrade its systems. However, they would then upload an autonomous tokenized governance canister, and then irrevocably transfer control of the new service to this (so that it becomes part of that autonomous system). Of course, at this stage they would hold all the governance tokens, but their aim would be to divest themselves of tokens over time to fund development and incentivize contributors, such that eventually they hold only a minority stake and the service becomes truly “open”.
We think that the larger the number of holders of voting tokens in the governance systems of internet services, the more attractive the internet service will be and the greater the number of supporters and promoters they shall have. Founders and key participants will still drive key decisions and day-to-day operations as most governance canisters, such as the standard ones offered by DFINITY, will support variations on “liquid democracy” algorithms that allow token holders to assign their voting power to others (where after the assignments of voting power will generally mostly be updated only when systemic problems emerge or emergencies occur) providing for strong dynamic leadership even though there may be enormous numbers of stakeholders. However, even if the founders of an internet service fail to divest themselves of sufficient governance tokens to make it truly open, their project can still guarantee that APIs they publicly share will not be revoked.
In order that our imagined open replacement for LinkedIn benefit from increased network effects through enabling other entrepreneurs and communities of developers to safely extend its functionality and reach by the construction of new services, it must configure those APIs it wishes to share as “permanent”. For example, it might wish to provide an API that allows other services to pull professional profiles for their own usage by username. If it marks this API as permanent, then other services will be able to safely incorporate the professional profiles it hosts. For example, an innovative open CRM service might be created using the profiles. Not only might this make the profiles more useful to their owners, but the CRM service would gladly direct new users it engages to enter their profiles into the open LinkedIn service, since it would be secure in the knowledge it always has access, creating mutualized network effects.
The Internet Computer protects permanent APIs in two ways. Firstly, once a canister contains a permanent API, the Internet Computer will not allow its controller (whether that is an individual, an organization, or an autonomous governance system) to perform software upgrades that would remove the API. However, this only helps prevent permanent API end points being removed by accident, and a software upgrade might deliberately degrade the functionality supplied, for example modifying the behavior of our hypothetical open LinkedIn service so that instead of returning real professional profiles, dummy profiles are returned. Here the Internet Computer provides a second level of protection for autonomous services controlled by its governance systems. In the event such a degradation occurs, a proposal can be submitted to the Internet Computer’s own governance system, the Network Nervous System (or NNS) to restore the APIs. If the proposal is adopted, the NNS will then inflate the governance tokens of the service, for example by creating and then publicly auctioning new tokens, and continue doing this at regular intervals until the API is restored. Of course, the game economics involved will ensure that such actions hardly ever become necessary.
Autonomous software, governance systems and permanent APIs, might seem esoteric, but in fact they are powerful tools that will enable the creation of a new, better internet, which provides more opportunity for innovation, creativity, entrepreneurs and investors. What part of the new ecosystem will you build?