Procedures

An Idea Worth Exploring

A redlined recipe from my kitchen

What do procedures make you think of? For me, it’s skinny black ties, horn-rimmed glasses, and chain-smoking nerds. Basically, a scene out of Apollo 13. Procedures probably take up a similar spot in your imagination. Maybe you don’t see space nerds, maybe it’s biology nerds, or quality control at a factory, but I bet you associate procedures with pedantic and technical jobs.

I doubt you imagine procedures are something you would want to bring home into your family life.

I want to challenge that assumption. You have probably considered what a paper procedure is capable of, but digital procedures offer a wildly better value proposition. Paper procedures are invasive, we have to flip back and forth between reading, acting, recording again and again. With digital technology we can shrink that gap, making something that feels much more like pure action. In fact, I believe that digital procedures may offer us a language especially suited to addressing many contemporary social, technical, and environmental challenges. Not only that, digital procedures may help us renew our individual and collective sense of purpose, fulfillment, meaning, and even our joy along the way.

Crazy right?

Possibly. Buildonomy is here to test the idea. I want to share with you some of the reasons why digital procedures could be so powerful if given the chance.

First an announcement. There’s finally a home page for Buildonomy! Please head to https://buildonomy.com and check it out! Not only is feedback appreciated, it’s necessary for me to proceed. If you can, please leave a comment on this post, fill out the survey linked on the homepage, or schedule some time to talk. Please share this post, or https://buildonomy.com with any person or group who you think would be interested in the message. Buildonomy needs data to proceed. I need to ask you questions. In order to make Buildonomy real, I need to check my beliefs against the real world, and I hope you can help!

If successful, we can become a network of first-movers and begin to explore the possibilities that digital procedures unlock.

Anatomy of a Procedure

A Procedure maps an intention to a method for achieving it. Fundamentally, procedures describe how to achieve a goal. The method is always secondary, the intention or goal is primary. It is the reason why the procedure is written, remembered, or accessed.

Next, a procedure describes a method of achieving an intention. If the method is followed, and if all assumptions the author implicitly or explicitly made are true, then each execution of the procedure achieves the intended result.

Simple right? Yes, and no. The easiest way to show what’s special about procedures is to compare them to software. Unlike software, procedures assume the practitioner of the procedure has autonomy. While running a procedure, the practitioner always has the authority and responsibility to deviate from the script, or ‘redline’¹ the procedure. A computer processor has no such autonomy. If any assumption is not true between how software expects reality to behave and how reality actually behaves, the software will fail. With a procedure on the other hand, practitioners have the authority to augment the procedure’s content to their best ability, and can therefore achieve the intended outcome even if the procedure’s assumptions were invalid.

In addition, through repeated runs, practitioners gain proficiency with the procedure, deeply learning the content, not just memorizing the explicit material. This allows a skilled practitioner to perceive relevant contextual changes, enabling them to extend the procedure’s content into novel environments. This generalization cannot naturally grow out of explicit programming. This is learning from experience, and is obviously powerful. By creating structured ‘as-run’ records of each run of a procedure in order to record the “who, what, when, where, and why” of each run, digital procedures unlock a method to augment natural learning with machine learning.²

It is these three properties, 1) explicit intentions, 2) adaptive and interpretive execution, and 3) the process of learning-by-doing that make me so excited about digital procedures. I believe these properties enable is a medium for learning, skill acquisition, and most importantly for action that is simultaneously intimate and general. Intimacy comes because procedures create dialogue between the procedure and the practitioner, they do not prescribe or enforce their methods but encourage in-situ adaptation. This allows them to conform to virtually any situation where the procedure and practitioner share a common intention. Their generality emerges from intention driven execution and the emergent properties of information underwriting such a system.

Conclusion

Digital procedures have the potential to unlock positive-sum dynamics across a wide variety of fields. Because each run of a procedure is valuable not only to the practitioner running the procedure but also to the community that uses the same procedure, they offer a communication format for simultaneously enabling individual and group self-actualization. By leaning into this dynamic³, individual actions can improve group-sensemaking, and group-sensemaking can reciprocally improve individual actions. More begets more in a virtuous cycle of self-improvement. If this dynamic can be harnessed, then it can improve any number of measures of human and global well-being.

Errata: More about Intentions

In order to better understand why I’m excited about procedures as an under-explored medium, once again we need to talk about intentions. If you read The Importance of Inferring Intention, you already know I think intentions are important. But in this piece, I’m arguing that they tie the whole idea of digital-procedures-are-a-cool-and-under-explored-thing together. Why is that? As related in that earlier essay, intentions live in the boundary space between our internal states and the external world, which I believe is an essential space to understand if we want to yoke technology to meaning and purpose. Two particular properties of intentions are critical if we want to natively communicate meaning and purpose with one another and with our technology. These properties are that intentions are scale free, and intentions are partially orderable.

Intentions are scale-free in the sense that I can have intentions that apply to any scale relevant to my existence. For example, I intend to help humanity be healthy, wealthy, and wise. Intentions extend down and can be as specific as the intention’s I express while manipulating my keyboard to write these sentences. They are scale-free in the sense that nothing about the breadth or duration of the intention makes it more or less of an intention. In fact, we use many different words to point to different scales of intention. A possible ordering of some of these terms might be: aspiration, purpose, goal, aim, and action. I believe that all these terms live on a spectrum that ranges in generality from true across the breadth of my consciousness down to only true within the most specific moments of perception.

Intentions are partially orderable because of how they relate to one another. Intentions don’t directly relate as peers, but only through ordering along this continuum of intention. A general intention may relate to more-specific intentions, but will have no direct relationship with any equally general intention. For example, If I intend to help make the world a better place by increasing health, wealth, and wisdom in the world, my personal methods for achieving that may be tied to a series of sub-intentions, such as intending my family and I to fulfill that aspiration, and also intending to direct my remaining resources towards meaningful contemporary issues that I am most suited to engage with. Those two sub-intentions are peers, but only because they are both related to how I engage with their more general parent intention. In fact, specific intentions may fulfill multiple general aspirations. For example, as a sexual being I also want my family to be healthy, wealthy, and wise in order to help ensure the continuing presence of my hereditary line, yet this general aspiration is not otherwise related to my aspiration to make the world a better place.

A network of intentions creates a Directed Acyclic Graph (DAG), which imbues the network with order and structure without any externally driven constraint. It simply happens because of the information properties of intentions. We don’t see many human-made systems that innately have this property. One that comes to mind is how open-source software starts to generate ecosystems. A software application commonly depends on numerous external libraries, each of which may have its own dependencies. In this way software can build-upon general abstractions and extend those building blocks to many different applications.

Buildonomy shares many of the same goals as both Personal Knowledge Management Systems, and Project Management Software. The scale-free and partial order properties of intentions serve as a useful point of comparison on what differentiates Buildonomy from either product category.

Personal Knowledge Management Systems understand the primacy of references for knowledge retention and acquisition. These systems also understand that links are scale-free, because any material can link to any other. But they offer no innate sense of structure or order. Structure and order are arbitrarily enforced by the users of such systems. There is no way to tell purely based on a link network whether the structure of the network has general or extensible properties. This prevents knowledge from organically extending its relevance into unanticipated territories. One person’s knowledge graph cannot be naturally aligned and compared with another’s. One person’s intention graph can, because there is always an intersection between either their high-level intentions or their shared reality that can serve to communicate the similarities and differences between the two structures.

Both Buildonomy and Project Management Software ensure task relationships are ordered and structured, but Buildonomy allows this structure to organically evolve, whereas Project Management Software requires such relationships to be explicitly enforced. Additionally, Project Management Software is not structured around scale free relationships. For example milestones relate to tasks, tasks relate to personnel, or assemblies relate to parts. Project references have meaning, but each type of reference comes with assumptions that prevent any general information from emerging from the specifics. In my experience, an endemic issue with Project Management Software is how much effort it takes to curate the contents into something useful. Without continuously examining and pruning the system, it quickly becomes a junkyard of half completed ideas and dated information. Any such system quickly loses the big picture in the face of a million microscopic metrics.

Even if Project Management Software is obsessively pruned for relevance, it invariably does not organically keep information in sync with either the big picture or the current particulars relevant to each project. Why embark on a project in the first place? Why are we developing gizmo X? Is it just because the project manager said so? No, it’s because we as a company are really good at developing gizmo’s like X, and to continue to exist we need to keep up our reputation and relationships with our business partners who count on our gizmos. Such reminders may seem trivial and redundant, but without explicit relationships between the high level aspirations that lead to the project, it constrains an organization’s ability to modify their intentions when the environment changes. If the market for gizmos dries up and business partners now demand doodads instead, the project data doesn’t ‘know’ this at any structural level and so its ability to augment the organization’s relevance realization is limited.

1. A term from paper procedure days, where changes were marked with a red pen

2. I believe the type of machine learning most suitable to this system is based on Active Inference, not the standard back propagation methods that currently dominate the technological hype cycle. I will discuss the reasons why in a later essay.

3. While ensuring individual data sovereignty and autonomy.