Video Playlist: The Age of Networks & the Rebirth of Cybernetics
Presented at the 2019 Web3 Summit, in this nine (9) part video playlist Blockscience Chief Engineer Dr. Michael Zargham shares insights on engineering complex systems, cybernetic steering, and the roles of humans and infrastructure within adaptive networks.
Clipped from the full presentation above, this post shares nine (9) short videos on the following topics:
- Engineering & Adaption: Patterns of Interaction
- Thinking About Networks: Nodes & Edges
- Cybernetics: The Science of Steering
- Safety Critical Systems
- Applied Mathematics of Coordination
- Blockchain-Enabled Complex Systems
- Automating Methods for Decision-Making
- Steering Purpose-Driven Systems
- Autonomous Systems & Adaptive Maintenance
1. Engineering & Adaptation: Patterns of Interaction
We have always co-evolved our social and economic systems with our technological and physical systems. Whether learning to build bridges and roads, skyscrapers, or digital infrastructure, we have adapted our patterns of interaction
Infrastructure is the key word here, unlike building products - infrastructure is successful when it's invisible. We can turn on a light, browse the internet, and water the garden, without going out of our way to access the infrastructure that enables these activities. Economics may help us understand the limitations of what we can do, it is important to recognize that in a complex adaptive system, where we are a part of the system, we must reason about our participation in the system.
2. Thinking About Networks: Nodes & Edges
Networks are nodes and edges in which we are the links and nodes. Whether peer-to-peer, economic, or social networks, these are systems comprised of humans and the infrastructure that allows us to interact with each other.
Networks are about people and the technologies that allow us to coordinate. Any particular edge in a social and economic graph can be an exchange of wants and needs. But there's a difference between the transactions that happen and the transactions that are possible, and it is often infrastructure that defines what we can do with each other.
3. Cybernetics: The Science of Steering
Used to understand the way systems interact with each other Cybernetics is a science of steering systems that have their own objectives, wants, and needs and are in some way co-steering each other. With cybernetics, we can design decentralized systems, or at least make inferences about how a decentralized system might adapt.
More importantly, cybernetics helps us make inferences about how we impact adapting systems, and recognize that we are not just steering the system, but that we are a part of the system. At the same time, it is hard to reason about systems with feedback loops. Fortunately, we have modern progression to cyber-physical systems and tools to improve complex system modeling. In the case of truly decentralized systems, this also includes feedback loops between the design and modeling tools and the decision-makers.
4. Safety Critical Systems
Engineers need to ensure that safety-critical systems have protections that help them avoid breaking out of their models. If a system starts to move in a direction that could violate the safety-critical assumptions of the system design, then it should be programmed to change modes, through some sequence of contingencies.
This does not necessarily require escalation to humans. With automated systems, we can escalate potential risks as something approaches the known boundaries of the systems models. While all models are wrong, they are useful. Even when we start small, learning how to build social and economic infrastructure and letting systems grow in scale as we understand them better, we can still build upon the knowledge of large-scale model-based systems engineering.
5. Applied Mathematics of Coordination
Significant advances in applied mathematics related to coordination include game theory and multi-agent dynamic control problems.
These are the same concepts that we use to model human behavior and other rational actors, with the caveat that we don't know what the human actors' incentives are, and we probably do not want to impose incentives on them.
6. Blockchain Enabled Complex Systems
Describing complex systems as dynamic, networked, adaptive, multiscale, and stochastic helps us understand the behavior of systems on a network, even when we do not know exactly what is going to happen because we never really know what people will do yet we can know what people can do.
As engineers, we need to be purposeful, to understand what it is that we hope to accomplish and how the systems we make might behave so that we can govern them. One of the biggest challenges with complex systems is that you can not have good-quality participation by the governed without informed participants, and so we can not have an arbitrarily high attention cost for participating.
7. Automating Systematic Methods for Decision Making
Highly technical systematic methods for making decisions are hard to apply. Yet when we start automating these systems, we can push stronger methods higher up into the domain of applicability because we can embed them and abstract them away from the end users, to provide people the benefit of higher validity methods, without the need to understand that they are there.
For example, a market is an emergent estimator that attempts to discover price. If you properly design and embed a bonding curve into an interface - as an adapter between two economic systems - you can turn it into a better estimator. You could argue that the Uniswap liquidity pool, because of its statefulness can serve as a stronger estimator of price than the market would without it.
8. Steering Purpose Driven Systems
We steer purpose-driven systems not by telling them what to do, but by establishing goals and embedding epistemological (strong scientific) goals, and automated methods.
Steering systems help guide or coordinate us towards shared goals - even when they are in tension with our individual goals - and as a whole, those systems co-steer each other.
9. Autonomous Systems, Feedback Loops & Adaptive Maintenace
Engineering, large-scale infrastructure for social and economic systems, can not be perfect, but it can be adaptive, so we must pay attention to what our systems are, how they interact with each other, and governance.
It is important to recognize that while what one actually does and what one can do are closely related, they are inherently different; what we can do is infrastructure, and what we actually do is behavior. If our goals change as a community, our steering system needs to change with those goals. A truly adaptive system requires us to participate in the decisions about adapting the rules and the feedback systems so that they continue to meet our needs. Ultimately, informed participants steer the system in the direction they want it to go; this is adaptive maintenance.
About BlockScience
BlockScience® is a complex systems engineering, R&D, and analytics firm. We analyze and design safe and resilient socio-technical systems by integrating ethnography, applied mathematics, and computational science. With deep expertise in Market Design, Distributed Systems, and AI, we provide engineering, design, and analytics services to many clients, including for-profit, non-profit, academic, and government organizations.
