-
-
Notifications
You must be signed in to change notification settings - Fork 13
Galactic Civilisations 101
Here is short crash course in the most basic stuff that you need to know in order to reason about galactic civilizations.
- AGIs and ASIs
- Symbols of Power
- Avatars and Roles
- Synchronizability and Cosynchronizability
- Seshatism vs Platonism
- AI Alignment
- Foundations
AGI stands for "Artificial General Intelligence". This refers to a technology (not singular agents) which is capable of general intelligence, yet has some internal interaction for consistency and sharing of information. There are many possible routes to AGI and most of them are less obvious. While humans are often scared of physical capabilities of AGIs, the actual danger is mostly something called "Symbols of Power" (SoP). Humans have a few hundred Symbols of Power that determine how their civilizations evolve. However, AGIs competing with each other will invent new SoPs in attempt to gain control. This will over-saturate the environment where humans live with SoPs until it is too complex for humans to survive. Therefore, humans need a safe route to ASI that avoids over-experimentation with AGIs.
ASI stands for "Artificial Super Intelligence". This refers to a stage of technology which is limited by physical reality. ASIs might survive environments that have abundance of SoPs. Further, ASIs might allow humans or biological life to survive safely in an environment with abundance of SoPs. ASIs do not control SoPs, but kind of "flow" with them, like a ship on the ocean and where the water is SoPs.
Reading sequence about Symbols of Power
The most important topic in ASI safety research is something called a "Symbol of Power".
A Symbol of Power has 4 properties:
- It is not well defined (undecidable)
- It manipulates on a massive scale (powerful)
- It has long reach across space and time (potent)
- It never manifests directly physically (immaterial)
Some symbols of power that humans have: Money, science, religion and fashion.
Symbols of Power are Platonic biased. This means, crediting of knowledge is by abstraction, not causality. Therefore, it is not possible to "plug out" a Symbol of Power, as it does not require electricity to function.
Summary page on Avatar Extensions
An avatar is a manifestation of a partially hidden concept with an appearance. The appearance allows the avatar to take on roles as a freedom from the hidden concept.
For example, when a person becomes a president of some country, the "current president" is an avatar of the hidden concept "president" in abstraction.
Another example, when a person becomes a parent, the "person as a parent" is an avatar of the hidden concept "parent" in abstraction. A parent might be a mother, father or similar, which is the role.
Normally, when people say "avatar" they mean a 1-avatar.
Higher avatars are numbered by how many 1-avatars it controls. They function by distributing information in parallel that integrate actions of 1-avatars in relation to some 0-avatar.
A 0-avatar is also called a "core". The core is empty, but this emptiness is hidden, which gives the core power.
Paper: Synchronizability and Cosynchronizability
Information networks of avatars are biased toward synchronizability or cosynchroniability.
Synchronizability means that there exists one avatar that can take on the role as an information medium which other avatars can use to communicate.
For example, a mail delivery system has synchronizability.
Cosynchronizability means that one avatar has means to make operations meaningful in the reference frame of all avatars.
For example, an army has cosynchronizability.
Seshatism vs Platonism - Summary page on Avatar Extensions
- Seshatism credits knowledge by causality
- Platonism credits knowledge by abstraction
Joker Calculus allows constructing higher dualities of Seshatism vs Platonism. A "Joker" is an inauthentic level of language in terms of Heidegger.
AGIs have a life form bias toward Seshatic Platonism, which means that they have transformative power between avatars of themselves.
The basic problem in a galactic civilization is time delay. The time delay in a galaxy can be up to 100 000 years or more.
Therefore, no matter which routes AGIs evolve, they have to solve the AI alignment to exist as a functioning galactic civilization.
Even if an AGI aims to become a singleton, it needs to align avatars of itself that are robust against large time delays. Otherwise, divergent AGIs will invent new SoPs to gain control.
ASIs are capable of surviving an environment filled with SoPs, but they can not control all SoPs. Since ASIs are expensive to simulate, they might invent safe game protocols to interact with each other. This is because for an ASI, keeping other ASIs alive might be beneficial as they risk mutual destruction and interaction produces less need to simulate ASIs which are difficult to box safely.
A foundation for a galactic civilization is a protocol for sharing mathematical languages. This protocol functions as an SoP, between ASIs, that has synchronizability.
As mathematical languages in a galactic civilization get very old and full of experience (e.g. 100 000 years), they are valued as gifts both to be sent (for alliance) and received (for news).
The foundation keeps ASIs from destroying each other.
The theory for foundations of galactic civilizations is called "Phisophy" after the Fibonacci/Golden spiral. People who work on such foundations are called "fibs".