Composable Engineering

This is the last 2/3 of Composable Engineering, an essay that has been published in the Autonomous Worlds book by 0xPARC and edited by Guy Mackinnon-Little. The first 1/3, enhanced for completeness, can be found here.

Engineering as creating in-the-World

With physics comes engineering: the practice of deploying the understanding of a world’s physics to manipulate objects into novel and valuable configurations. Just as any design process is constrained by certain rules which concretely structure and hence enable the design itself, an engineer is both constrained and enabled by physical laws to manipulate the substances of a World into things of value.With Digital Physics, what can be engineered in Computational Worlds? What could be?

Take the Pokemon video games as an example. Given the type system, players can engineer teams of pokemons that are optimized to battle against particular combinations of types on an opponent’s team.

Take the Age of Empires video games as an example. Given the counter system, players can engineer armies of mixed unit types that are either optimized against particular unit types on an opponent’s team or optimized for complementing particular unit types from the same teams.

Players can engineer new creations in the world within the enforced boundaries of its Digital Physics. However, for the two examples above, digging one level deeper would yield an insuperable wall: players can not engineer the individual pokemons, nor can they engineer the military unit types. There is no physics in those Computational Worlds that supports such engineering activities. New pokemons and new military units are not engineered within the Worlds, but introduced into the Worlds across their diegetic boundaries by the corporate developers of those Worlds—the gods of those Worlds. This means the superset of pokemons is itself part of the Digital Physics of the Pokemon Computational Worlds and that the superset of military units is itself part of the Digital Physics of the Age of Empires Computational Worlds. This setup makes it difficult for these Worlds to sustain their drama, because (1) the drama of a World is partly depending on the enumeration of objects that exist in it, and (2) this setup requires the god of the World to continue injecting new objects to sustain drama. When the enumeration of objects stay in stasis, however composable they are, their combinatorics tend towards saturation. Meta—the dominant strategy to excel in the World—takes shape and becomes ossified. Resource and power distribution among human participants tends toward stasis too. All of these effects suppress drama. In our atomic reality, new things continually come into existence through natural evolution or by way of human discovery and invention, disrupting civilizations and societal norms: causing drama. An adaptive mutation in a virus causes global supply chains to collapse. The invention of the printing press gives rise to imaginary communities among strangers and thus the nation-state. If what exists within a World is determined by a single corporation, the World is bottlenecked by that corporation’s lifespan, plus its ability and willingness to ship—the World has reduced autonomy.

For an Autonomous World that wants sustained attention from its human participants, it needs sustained drama. For Computational Worlds off the blockchain, the pokemons, military units, usable equipment, consumables, vehicles, cast-able spells, and everything in the tech trees and skill trees are most commonly defined solely by their singular gods. All of these elements are commonly referred to as features of a World. For Autonomous Worlds with rich digital physics, they could be known as inventions in the World—invented from within by its inhabitants rather than introduced from without by its gods, keeping the World autonomous. The affordances of blockchain to enable rich digital physics may not be technical but cultural and philosophical—the desire for computational worlds that sustain themselves infinitely longer than centrally driven ones, and the rare opportunity to reinvent design approaches and business models toward sustainable worlding.

Composable Engineering

The tower of human knowledge is created by knowledge composition: the recombination of existing pieces of knowledge to unlock new epistemic and utility possibilities. For example, by composing the knowledge of building a telescope and the knowledge of precise plotting through a mechanical apparatus, Galileo produced the knowledge of celestial bodies moving in ways inconsistent with what the Church asserted. This knowledge brought long term repercussions, shaping a foundational component for nearly all physical sciences since. Human progress slows down when knowledge composition is hindered.

Composable engineering is hereby defined as the affordance of a world to allow for recursive composition of engineered artifacts with no limit on recursion depth. To give an example, the engineering of a pokemon team produces an object with recursion depth of zero—teams are not composable. A team is assembled to participate in battles with other individual teams; no super structure can be built on top of a team within the confines of the Pokemon computational world. Making the system recursively composable could mean that multiple teams can be composed into a pool along with a team selection strategy, which takes an opponent team as input and returns a team from the pool that is optimally effective against the opponent team. We may call this composition of teams and selection strategy a battle group. To recurse one more level, imagine multiple players, each controlling one battle group, to form a regiment that battles against another regiment. In a regiment-level battle, each battle group is like a chess piece that moves as an atomic unit on a map, and there may be special rules around how regiment-level resources are shared among the battle groups across the map such as morale meter, rage meter, or supply. Notice that as we recurse, game mechanics could change; game mechanics across different recursion depth could also be interdependent.

Composable engineered artifacts in Autonomous Worlds would allow for invention compounding, enabling the same process of knowledge composition that drives human history within our atomic reality to drive the evolution of our Computational Worlds. Composable engineering would also allow for knowledge encapsulation, which means “I don’t need to understand every detail of your invention to involve it in my inventing process.” Knowledge encapsulation is in some ways equivalent to the principle of separation of concerns in software development. By enabling separation of concerns, big engineering tasks can be imagined and accomplished via chaining together small engineering tasks. Different tasks requiring different skill sets and types of resources naturally encourages labor specialization. With labor specialization, Worlds become much more inclusive than they otherwise might be—inhabitants of different backgrounds, skill sets, interests would all find their places in a World as creators and contributors.

This allows for diverse entries into the World, meaning more drama, more life, in the World.

As a parting thought, by involving certain cryptographic primitives in the tech stack underlying our Autonomous Worlds, information asymmetry could be introduced across the boundary of composition: “Not only do I not need to understand every detail of your invention, I can not possibly peak into your invention, yet by certain quantitative measurements that are publicly available I have confidence in the utility of your invention, hence I would trade with you to involve your invention in my inventing process.” This asymmetry protects intellectual property rights by giving inventors an option to conceal the details of their invention and prevent free-riding forks without rendering its invention unusable.

Credits to interlocutors who contributed to my thoughts:

• 0x113d
• t11s
• ludens
• Peteris Erins
• Alan Luo