🔖 Glossary
This page is for terminology & concepts that are often referenced in & around the Open Memetics Institute's networks
Dark memetics - any research or engineering work that is not auditable by the people that are being operated on
- Paying an influencer to say something without disclosing that they are paid is "dark memetics". So are more benign things like Amazon A/B testing a million people to see which color button triggers more purchases
- "dark memetics" doesn't mean "evil", it just means it's done without the awareness of the recipients. It could be for good reasons, or bad reasons
egregore - also known as "collective mind", or generalized "agent" in the cybernetics sense.
- One way to explain this at a normie level is to say "group chat". When you work at a company, you're part of the company wide group chat, but also your team's group chat. You have an implicit "group chat" for every information network you're in, your friend circle(s), family, etc
- egregores can be "nested" inside a parent egregore (like a team within a company)
- This is one of the basic "units" that we study in open memetics. For any human we can ask which egregores are you a part of? In what ways do they influence you, and how do you influence it? Are you part of a read-only or read-write egregore (it broadcasts information to you, vs also taking information from you)?
Coverage - this is like how journalists have a "beat", every human has a part of the memetic landscape that they occupy & have visibility into.
- The OMI network aims to have maximum coverage by getting enough volunteers from as many networks as possible
- This is technically just "what egregores are you a part of" plus "which egregores are you aware of"?
epistemology - how we know what we know. An accurate description of the epistemology of the agent would allow you to predict how they will react to new information.
- A great way to map epistemology is using ORI's A/B/U system.
normie - a category of agent that only updates their beliefs when a critical mass of people around them have updated.
- a normie is NOT necessarily a passive consumer. Their purpose is holding the floor of culture steady. They are gates that only allow changes if they have been tested by enough others first.
- secret coordination / dark memetics breaks this by making it appear as though a change has already happened, in order to get over this filter. In open memetics we use this filter, we don't override it
epistemic noise - any information that makes it difficult to surface the truths the observer seeks. Usually because it's false, irrelevant, or worse: is true but guides you towards false conclusions.
Examples
Example: the Streisand Deflect is an example of epistemic noise because it injects true information into the system to hide a specific truth.
Example 2: an engineer writing a design doc pointing out all the flaws in a proposed system, only because it threatens their team's headcount. Everything in the doc is true, but it fails to convince the discerning reader. It's designed to convince those with less discernment, like decision makers up the management chain. This is sometimes referred to as Gish gallop
Open / closed research gap - this is the gap between what is being studied openly, and what is "not allowed" to be studied (and thus happens freely in private)
- See University of Zurich scandal, their attempt at closing this gap: Researchers Secretly Ran a Massive, Unauthorized AI Persuasion Experiment on Reddit Users
- The OMI recognizes that it is impossible to regulate what you cannot see - so we aim to close this gap through active monitoring
Self checking statement - For an announcement like "the person who discovered X is so & so", this statement is either (1) true or (2) false. If it's false, it tells you something about their (lack of) awareness. Either way, you gain truth.
- When you go & correct it, if they are good faith, then they will update. If they do not update, they signal themselves as bad faith. Either way, you gain truth
incompetent OR malicious = bad faith actor - A bad actor can hide forever perfectly under the guise of incompetence.
- The way we short circuit this is to create a community where we assume ALL bad things happen due to incompetence.
- Thus, an actor who is incompetent AND doesn't respond to feedback is considered "bad faith", and must be removed from their position of power, because they will hurt themselves/others
quokka - a node that cooperates 100% of the time.
They are useful in surfacing bad actors or hostile environments. A quokka generates value but never tries to capture it. If you insert a quokka in an environment and watch what happens to it, it essentially sacrifices itself to surface an unfakeable signal (who is willing to steal from it & lie about it).
A real example of a quokka is the "autistic engineer" archetype. They function as "NPCs" in the game of "company politics".
(side note: this is why "normie" != "NPC". See definition of normie above)