Platform engineering is traditionally associated with the development of internal development platforms specific to the organization, however the Golden Path requires infrastructure that can support fully orchestrated packaged capabiltiies that are shared between organizations on a two-sided platform. Powerfully disruptive network effects are created from the Virtuoso no-code infrastructure with discoverable content that is fully orchestrated by Montage. To understand these network effects, and to leave nothing to the imagination, an economic deep dive is needed.
Software production may be considered a service economy that suffers from Baumol's cost disease, an economic disease that affects goods and services whose production are characteristically resistant to fundamental improvements in productivity. Like haircuts, medical care, or college tuition, certain services must be provided by skilled manual work. Characteristically, these production economies are resistant to productivity improvement through economic innovation, compared to other goods and service production economies like consumer electronics. Improvements in the means of production in other areas translates to higher standards of living, which translates to disposable income, which translates to more demand for goods or services affected by Baumol's cost disease. This "demand elasticity" results in even higher costs for services affected by Baumol's cost disease, like software production.
The advent of large language models has been perhaps the only meaningful technological innovation that can address Baumol's cost disease for software production. While there is still generally a low opinion that generative AI will replace software developers, improvements in software developer productivity of up to 20% (for those organizations that allow its use) is commonly cited. Some opinions hold that generative AI's productivity gains can double developer productivity, but even this is not a truly disruptive economic shift.
Nonetheless, Baumol’s cost disease, as it relates to software production, can be alleviated. The insights provided here are primarily those of economist Ronald Coase, boiled down and mapped to software production. Readers of this introduction are encouraged to read his article “The Nature of the Firm", to fill in gaps that can’t be filled in this intorduction. References to "the firm”, the “individual”, “entrepreneur”, “in the firm”, “economic coordination”, and others may otherwise be misunderstood. Coase’s interests lay in understanding why firms exist, and what determines their sizes. He determines that transaction costs are the key drivers to the optimization of forms of economic coordination that occur as either actions of individuals or actions of a firm. An important theoretical distinction is necessary to distinguish the firm and the individual. The “individual” coordinates production through market prices alone. The classic example is a quality shirt made in Malaysia using German machines from cloth woven in India using cotton grown in the United States. This chain of production is organized by market prices at each step of the supply chain. The supply chain participants of a quality shirt market then act in concert, performing the unconscious calculus that compares costs, features, and qualities, to select suppliers that have found the optimum chain of production.
While each supplier also obviously contributes to coordination, the primary coordination of this production is understood to be the unconscious market computation of prices as the aggregation of supply and demand. From these market calculations, individuals act. This process is akin to the power of quantum computing to arrive at the best solution from a nearly unlimited set of possibilities. Coase’s insight was that the production coordination of firms is different from individuals coordinating production through market prices. A firm sets out to produce economic value, certainly affected by market prices of inputs, however coordination of production here is something different. What is referred to as the firm’s “entrepreneur” or “undertaker” does not know exactly how much it will cost to, for example, take old rockets, add autonomous vertical take-off and landing, and completely revolutionize the means of producing space launch capability. This involves a conscious agency of a firm and its entrepreneur to coordinate the specialization of labor to produce something that market prices cannot.
The entire gestalt of economic production is thus described in “The Nature of the Firm” as a sea of unconscious market computation, with the entrepreneur’s coordination representing an “island of conscious power”. The firm entrepreneur “busies himself with the division of labour inside each firm and he plans and organizes consciously,” but “he is related to the much larger economic specialisation, of which he himself is merely one specialised unit. Here, he plays his part as a single cell in a larger organism, mainly unconscious of the wider role he fills. [sic]”
Coase’s Theorem includes interesting insights into factors of production, transaction costs, and how they shift the equilibrium of production coordination between individuals working with market prices, firms making conscious internal coordination, and firms coordinating with other firms. Each type of coordination involves transaction costs, and while unpacking them further is interesting, here only two improving shifts in production will be presented. In a strictly economic sense, a “combination” occurs “when transactions which were previously organised by two or more entrepreneurs become organised by one. [sic]" All things being equal, this is an improvement in production, as it only requires the transaction cost of one party. Beyond “combination” is an economic “integration”, which occurs “when it involves the organisation of transactions which were previously carried out between the entrepreneurs on a market. [sic]”.
“The Nature of the Firm” provides economic theory that makes software production more understandable and can avoid costly architectural decisions. The theory can also provide proof of value. We attempt to do both, first by reviewing the function of a software platform. Some platforms represent a “combination” in the economic coordination sense of Coase’s Theorem. Before the platforms were built, multiple entrepreneurs had to transact together to produce a capability. Once the platform is created, it increases software production by allowing users to build out capabilities leveraging the platform. The transaction costs involved are then incurred by only one entrepreneur, the user that uses the platform to produce a specific use case solution.
Other software platforms increase software production by replacing coordination (software development) that was previously performed “in the firm” and instead moving it outside the firm. Modern no-code and low-code platforms are an example of this. Here, the applications are developed by platform users that leverage no-code tooling to create capabilities. Some economic coordination (software development) is performed by the platform user, however most of the actual economic production is done by the no-code platform. This is an example of a no-code platform creator entrepreneur coordinating production to create a means of production mediated by market action: No-code users select a platform based on price and quality, sharing the cost of production. This technology, when applied successfully, results in a natural reduction in the size of the firm, as the transaction costs of using the platform are much lower than the internal transaction costs of coordinating the same production without the use of the platform.
In either example described, the combining coordination of platforms or the outsourcing coordination of no-code platforms, we see three missing processes of economic production described earlier. First, there is no economic integration. That is, there is no technology that supports a means of software production which previously required at least one entrepreneur, and now requires no entrepreneur coordination. Again, by “coordination” we refer to where actual economic production occurs. This missing first process results in the second missing process, which is a hierarchical orchestration of the means of production. Again citing Coase, “[the firm’s entrepreneur] is related to the much larger economic specialisation, of which he himself is merely one specialized unit. Here, he plays his part as a single cell in a larger organism, mainly unconscious of the wider role he fills. [sic]”. In the wider economy, means of production form “holons”, which are simultaneously a whole in and itself as well as a part of a larger whole. This process involves firms, in their conscious economic coordination, creating a higher-order emergence of coordination that is again unconscious. This second missing process results in the third missing process, which is a functioning marketplace of higher order composite capabilities, coordinated unconsciously by price, yielding enormous productive power to the firm’s entrepreneur.
The goal of this introduction section is to firmly establish these economic understanding in the minds of software infrastructure architects. The understanding clarifies the economic value that some platforms have, while also clarifying where their value might currently be limited due to a lack of infrastructure. We can also use this understanding to predict that modern no-code and low-code platforms will likely be extinct in 5 to at most 10 years. These business models require a platform service value stream, where they are compensated for providing capabilities that the firm internally coordinates/produces. These modern no-code and low-code platforms have nonetheless demonstrated their value. Competition among these firms is fierce, due to the enormous market size and the overall lack of key differentiators. And because all the productive coordination of capabilities must be done inside the respective no/low-code platform provider firms, the amount of capital that these firms require, and have attracted, is eye-popping. In our analysis, these investments in software production, where they can’t be properly re-architected to facilitate the three additional spheres of economic software production, will prove to be evolutionary dead ends. To understand why, we must consider infrastructure that enables these additional spheres of software production coordination.