Same Systems, Alternative Thinking

A few days ago, I accidentally stumbled upon a LinkedIn post about a brand-new newsletter on systems thinking “inside and outside technology.” Full of enthusiasm, I opened the link, and my spirit turned to disappointment just as I read: non-linear thinking, holistic views, causation loops, and archetypes. Every time the same: systems dynamics, a bit of systems theory with some cybernetics in different flavors.

Ironically it was mentioned in the context of technology. I mean, nothing wrong with system dynamics — it is a well-grounded and proven discipline that allows one to model and predict a multitude of relevant scenarios, very applicable to management and leadership. I can imagine it effective in product management practices (primarily strategic and visionary, though). But when it comes to technology, I have a much better alternative.

It was created long ago and earned deserved respect as “systems engineering,” but now it is more than just an engineering discipline. Much more.

one of the examples of a typical academic landscape with “Systems Thinking,” “Systems Approach,” and “Systems Engineering” placed separately as three different disciplines

History of “Systems Science”

The history of the 3rd generation of “our” systems thinking of engineering origin as a very brief and opinionated bulletpoints:

• Generation one: “general systems theory,” Bertalanffy, 1937. Ontology, epistemology, hierarchies, control, and feedback. Developed by cybernetics and the theory of dynamic systems. Widely adopted in biology, economics, sociology, and games theory. Perceived system as a ‘holon’ — an absolute and intrinsic object in a defined hierarchy. Direct relations types are “composition” (aka part-to-whole) and “feedback loop” (positive and negative). Now it’s known mainly through system dynamics, popularized by Peter Senge and others. Another exciting application is a STATIK method from lean kanban.
• Generation two: Based on ST1, but narrowed it down for engineering purposes. ~1990, the transdiscipline of engineering, developed by INCOSE, SEBoK. Has ISO 42010 as significant output. Adopted in the military and civil engineering, space and aviation, demonstrated optimal value in complex and costly projects. Mostly known for NASA, Airbus, Boeing, and SpaceX adoption. Spotted also in the IBM toolbox. Introduced role-based viewpoints, functional and physical architecture aspects duality, and a new type of relations “creator/enabler system”-to- “system-of-interest.” It usually works well when combined with TRIZ and Prince2, and sometimes SAFe. In the management field, it is represented by the (often misused) enterprise architecture discipline.
• Generation three: ~2017. Based on ST2 (INCOSE) but absorbed a few important SoTA principles: scaleless, evolutionary (memotype-phenotype, smart mutations), quantum-likeliness (multi-level quantum frustration, probabilistic inference). The system is not created, but develops/evolve. It is not a techno-organism, but a techno-species in the course of techno-evolution, competing with other species. Continuous delivery, continuous everything. Thoughtworks people, Eric Rees, Vanchurin-Katznelson, etc. (see all references here). Aggregated and harmonized by joint continuous work of the Russian chapter of INCOSE and a “the School of Systems Management” under the leadership of Anatoly Levenchuk and Tseren Tserenov. Their role is not to “invent” a new SoTA in systems thinking but enrich it by the edge of human knowledge in areas of fundamental and applicable knowledge: the grounding is available in the “foundational ontology” whitepaper — this is remarkable content worth reading.

Alternative Systems Thinking

The first thing you like about it is its “explanation power.” After learning this new way of thinking, I was able to retrospectively explain a lot of things that happened in my professional career:

• Why poorly-engineered software sometimes works better than well-engineered software?
• Why are there so many ideas, frameworks, and approaches, but there is no “silver bullet”?
• Why we cannot estimate a good User Story effort/duration, but can easily estimate a Task?
• Why do all that “strategy,” “digital transformation,” “lean,” and other buzzwords really work if made right?

And many others. It is funny, but the real power of the ST3 comes later. The ability to comprehend complex, long-running changes around us. The ability to frame your personal (and after a collaborative) attention on essential matters and control the focus for the necessary time. The ability to get to agreement numerous stakeholder parties on mutually beneficial rational commitments. The ability to scale up an organization, define clear boundaries, and promote self-development transparently and efficiently. All that is just a subset of ST3 application outcomes from the top of my mind.

Technology Sweet Spot

Even ST2 can be very efficient here, but we better talk about the most recent generation.

OMG Essence Kernel

In general, ST3 already has a multitude of applications. We can even see some exotic ones, like dance or fitness. But I still believe we can achieve the most impactful application in technological entrepreneurship. This is about the combination of 2 groups of systems (“Product” as system-of-interest with its entire systems hierarchy + “Organization” as a creator-system with its complete “chain of creation” systems).

ST3 framework can provide you with actionable insights for every element and every aspect of the “Product” and “Organization” combo, depending on what you need. Good news, you can also discover the “what you need” part with excellent add-ons: extended OMG Essence framework, augmented with Lean + TOC concepts and methods, nicely described in the TameFlow.

And last but not least, I’d like to mention another ST3 benefit for technologists and information system architecture practitioners like me: Domain-Driven Design become your best tool of choice. It is not the ambiguous guidance you read before adopting ST3. Whether you are dealing with complexity, maintaining the conceptual integrity of systems, or implementing a technology strategy — DDD will provide you with metamodels, communication, and oversight tools. It is now equipped with modeling techniques of domain-mapping (lifecycle model, capability model); it is now manageable and predictable over time, and it is now evolving together with your “Product” and “Organization.” My personal superpower, my best recommendation — is ST3 + DDD = SotA in the architecture design of complex dynamic systems.

How it Works

Modeling. Focus on “what’s important.” This is achievable by using a meta-model, like a “schema,” for creating new models. ST3 gives you a set of meta-models for organizational and product engineering meta-models — like meta-meta-model, allowing you to build a relevant and fine-tuned meta-model for your activity domain and use it to scale knowledge in a clear, repeatable, and governed way. Not a disruptive innovation by itself, this approach is different because it is grounded in a more universal “theory of intellect,” which is modeled in ST3 as a stack structure:

• “Conceptization” defines how to highlight an object from the background/noise to make it the subject of consideration.
• “Attentiveness” defines how to keep in focus the “objects” that have already been highlighted in Conceptization.
• “Semantics” defines how to separate physical objects from mathematical/abstract/mental/ideal, thereby separating objects and their more or less formal descriptions.
• “Mathematics” … Set, axiom, calculus, space, field, countable and uncountable infinity. The subjective theory of probability (non-physicality of probability), Kolmogorov and non-Kolmogorov probabilities.
• “Physics” … Thermodynamics (action, energy, free energy, phase space), information theory (including superinformation and including “replication invariance”, including thermodynamic semantics), physical principles and laws. Quantum, ergodic. Stochastics: emergence and frustrations, renormalization. Constructor theory.
• “The theory of concepts” works like a “typification engine”: it is able to cluster together some objects that are similar to each other in some sense, and this is defined by types. We can analyze objects by their relationships with others. Examples of common types of relationships are classification, specialization, and composition. And the “Information theory” has already defined how descriptions manifest themselves in physical reality.
• “The systems approach” (aka ST3) defines that the physical world has a multi-level nature in relation to composition (mereology), and each level has new properties, so descriptions of some levels are irreducible to descriptions of other levels (emergence). The theory of concepts has already given composition as one of the basic relations, and we can continue to discuss mereology.
• “Ontology” defines the ways we can describe/model the world: how do we define what is important and what is not important (modeling), and how do we use models to answer questions (interpretations).
• “Algorithmics” … Calculations (including the concept of inference and update) and theoretical and practical computability, intelligence, physicality and universality of calculators (including a universal approximator — a neural network), calculation complexity, algorithms, ways of describing calculations/algorithms (programming paradigms, including probabilistic and quantum). No free lunch theorem, P and NP equality problem.
• “Logic” … Reasoning by rules (with given types according to given algorithms), reasoning rules (Boolean, Bayesian, quantum-like), arguments, errors in reasoning, and attitude to them (ignore or take seriously). Formal incompatibility of different ontologies, microtheories/ontics.
• “Rationality” … Descriptions of the reality (models) and rationality versus empiricism, the sufficiency of RL and evolution with “free energy” as the absence of a global “reinforcement,” the functions of rewards are temporary and can be learned). Decision theory as a causal quantum-like active inference (excess Bayesian inference). Three levels of the ladder of understanding - explanation - causality: associations, “what if” — interventions, possible worlds, and counterfactuals. Action as a reaction to changes in the environment (active/embodied, not any inference).
• “Research” … critical rationalism (refutability without evidence), Deutsch’s additions: hard-to-variate explanation, universality, compactness, optimism, the infinity of knowledge, the structure of reality (fundamental explanatory theories). Creativity and novelty. Freedom of criticism as a precondition for research.
• “Aesthetics” … beauty, elegance.
• “Ethics” … Ethical principles, evaluation of ends and means. Whose ethics (the problem of the individual and conflicts of system levels).
• “Rhetoric” … Convincing explanation. Model of agent’s knowledge, emotions etc. Conceptual distance, narrative creation.
• “Methodology” … Creator practice, the role of creator agent, life cycle, and continuous development. Agent, pragmatism (ends and means), benefits, strategy, alternative, preference, resource, subjective value, a negative value of work, intertemporal preferences, opportunity cost, marginal cost, marginal choice.
• “Systems Engineering” … Practices and roles: development, architecture, DevOps/platform engineering. Concept of operations, the concept of the system, architecture, modularity, engineering justifications and forks, and continuous delivery. A “genotype” and “phenotype” in techno-evolution. Doyle’s architectural ideas.

When we apply this stack model to our day-to-day engineering/architecture/management work, we can clearly see that almost every practice requires our minds to be active on multiple (or even all) stack levels. The bad news is that activity usually goes subconsciously, which can cause degradation of quality, loss of precision, and increase of complexity. The good news is that it can be refleсed, modeled, learned, and improved — this is what ST3 is focused on.

scales/levels

What to Do and Where to Go

ST3 school is about to “go global” with a new Aisystant brand, represented by https://aisystant.com/.

The bad news is that Aisystant is taking its first steps to globalize its knowledge and scale to the world level from a local Ru-speakers community. This means that most of the institute curriculum is currently unavailable in English. If you are lucky enough to understand Russian, the best way to start this journey is to subscribe to Open-endedness Program. After getting the basics, apply to any of the instructor-led training there.

For my English-speaking friends: there is a translated foundational course (Systems Thinking) and its base (Ontologics); more content is arriving. Translation quality is questionable, but the struggle from it will be rewarded with SoTA world-class knowledge and skill. An additional option is the “Study Group,” a collaborative discussion club with Aisystant-accredited instructors or volunteers like me :)

I hope one day we can have a more intelligent world with smarter us equipped with systems thinking; wondering what ST4 will bring?

Robert McCall, “The Prologue and the Promise”