Some Thoughts about Feedback Loops
These are notes in progress about how feedback loops work... not complete... nor necessarily intended to be understandable by others yet... all partially formed.
Coupling / Uncoupling – when you connect the “output” of something more directly to the factors affecting its “inputs” it becomes more tightly coupled. You uncouple feedback by inserting another layer of components/feedback/response between the operation of something and the feedback about it.
Visibility & Response
- For a system to respond to conditions on either an individual or collective level, indicators of those conditions need to be visible to responding systems.
- Responding systems are the self-adjusting parts of a larger system which effect flows in the larger system. For example, people, businesses, banks, government agencies, etc. are each different types of responding systems in the economy.
Boundaries / Edge Conditions / Operating Conditions / Breakdown
- If the only visible indicator of the condition of an integral flow is “there / not” which is the same as “working / broken” or “alive / dead” then you have a system which cannot sustain itself, because it doesn’t “know” that it needs to adjust its behavior until its dead. (e.g. If you don’t know you need to stop fishing a particular stock until you can’t find it anymore, you’ve caused the extinction of species. At the point, it’s too late to modify your behavior to sustain the species.)
Time Lag – Systemic Response Speed
- Based on the layers of uncoupling within a system and also certain cycles within that system, there may be lag time to see adjustments. Determining the amount of lag is critical to avoid under or over-correction. (e.g. Knowing how long it takes to stop your car or ocean freighter… Knowing that noon is peak sun but 2:45pm is peak heat. Knowing that June 21 is peak solar input, but July 21 is closer the peak heat, etc.)
- Part of the beauty and the challenge of living systems is that they are non-deterministic. They have multiple possible paths of action and response.
- This is fantastic from a resilience and adaptability perspective, however, if feedback information may flow through a variety of different paths, it can be very difficult to ascertain the system lag and
Misaligned incentives – Types of misaligned incentives:
- Expecting something to serve multiple purposes which have a fundamental conflict. (e.g. money being both a good medium of exchange and a good store of value.)
- Individual vs. Collective
- My “get mine” conflicts with the collective “grow ours”
- Individual vs. Individual
- My “get mine” conflicts with your “get mine”
- You can distribute more of the intelligence in an organization when the proper feedback loops are in place for it to be self-managing. However, the “free market” doesn’t manage things well which it cannot see. (water quality, widespread resource depletion, soil quality, air quality, etc.)
- To make a currency a good measure, it should have a reliable reference (such as kilowatts)
Selection of Value Reference or Determining a good Basket
- What makes a good reference? Universality? Real/functional value? Incentive to optimize performance? Choosing kilowatts as a currency value reference, incents people to find more efficient ways of producing electricity.
- A basket should be simple enough to update from real data within a marketplace on a periodic (daily? weekly? monthly?) basis for actual currency conversions.
- It should be diverse enough to make it difficult for a small group of people to manipulate
System Swings / Vibrations / Dynamic Range
- Poles (day/night, summer/winter, inhale/exhale, heart diastolic/systolic, waking/sleeping, tides, ebb/flow, etc.)
- The alternating pull between two (or multiple) states ensures the system stays dynamic, that things don’t freeze up, fall to stasis. Homeostasis may be a misleading goal… rather freely adjusting within operating range.
- Are there systems with designed in multi-modal pulls/states?