New Comment
11 comments, sorted by Click to highlight new comments since: Today at 2:17 AM

To learn, we must be exposed to sources of knowledge. Sources of knowledge can take the form of the environment, ourselves or other people. 

We rely on other people as sources of knowledge and learning because the people are amazing at learning and can synthesize what they have learnt into forms that make it easier for others to learn the same thing. 

However, it is important to know what we don't know. Given a source of knowledge, think about the set/space of knowledge that would be impossible / unlikely to learn from that source, even though said source has that knowledge, and the reasons for this.

What might teachers at common institutions of learning know of but be unwilling to teach their students? Why?

What might parents know of but be unwilling to explain to their children?  Why?

What kind of common knowledge might be hard/difficult to find on common repositories of knowledge on the internet? Why?

What knowledge is currently unreachable from your current position, not for lack the understanding of others, but merely for the fact that it is inconvenient and inelegant to have it explained to you?

Yes, Markets are Efficient, but only when they conform to my biases. If not, they are clearly fraudulent and incorrectly valued.

Hashing out my incomplete understanding of Simulacra.  :

Level 0 - Reality, let's call this R.

Level 1 - Agents map/point out reality to each other. For simplicity let's say there are two agents, A1 and A2. A1->R , A2->R. A1 and A2 can attempt to come to consensus on reality. Value is assigned to Truth, Power is over Reality.

Level 2 - Recursion. Agents can point out agents pointing out reality to each other, thereby potentially distorting reality. Agents realize they can "point pointing", and influence each other thus. Let R' be a incorrect reality. The following all belong as this level : A1->(A1->R) , A1->(A1->R'), A1->(A2->R), A1->(A2->R'). Note : A2->R might be distorted as well if that never happened in the first place! Value is assigned to Perception, Power is over Agents.

Level 3 - Emergence and clinging to reality.  Realizing that all agents can distort the perceived realities of other agents, agents start pointing to "shadows" of reality. This is level 2 recursed on itself, NOT another level-1 recursion on level 2. (level-2-squared instead of level-2-plus-1, in terms of recursions). Level 2 folding on itself. Agents point at things that seem (at lower levels) to concern agents and reality, but those agents are trying to convey information that is independent of those agents and reality. Thus, the agents and their pointing at reality are necessary but NOT sufficient to convey the desired information. The desired information emerges implicitly out of the distorted realities that hang around level 2. For example, we can let D1 = A1->(A2->R'), a where A1 points that A2 points at an incorrect reality. Then a level 3 example of an agent pointing is A1->(A1->D1), A1 is signaling (outer pointing) by talking (inner pointing) about D1. Value is assigned to Knowledge (understanding the meaning of these "shadows" of reality, which is distinct from Truth!), Power is over Groups of Agents.

Level 4 - Dissipation and Void. Agents realize that all agents are pointing at shadows of reality instead of at reality itself, thus the shadows lose their capability to convey the implicit information of level 3. Agents and their pointing at reality are not necessary to convey information, because no meaningful information can be conveyed. Nothing can be Valued, and no Power can be gained over other entities or realities. This is a new and separate "Reality" Level 0.

Anonymity reduces iterative prisoner's dilemma (staked on reputation) to one-shot versions.

The externalities are offloaded onto the platform providing said anonymity.

Anonymous and "Anonymous" tend to be different: Death Note: L, Anonymity & Eluding Entropy

Technically, the externalities are on the one experiencing them, positive or negative. What are the positive externalities of anonymity?

Just discovered and read about Conflict vs Mistake Theory, in my own mind my summary would be : Mistake Theory is about the "mind", Conflict Theory is about the "heart". 

I was also tickled by the meta-level problem.

We can have objects of a given type in a set, and we can have an order defined on those objects in that set.

Some people seem to hold values that positively value increasing the types of object in that set, while negatively valuing an order / large distances between those objects.

Others seem to negative value the increase of object types, favoring a smaller number of types while holding that an ordering between objects in a  set cannot be avoided.

While I admire the attempt to avoid politics, this (the topic of what different people value, and how you value that variation) is something that doesn't generalize very well without specific examples.  

I want to join/create a society of people who do not judge others at all, but how will they decide who to let in?

More thoughts on Simulacrum.

Assume that the setting is such that Agents can make statements about Reality.

Level 0 : Reality

Level 1 : Agents are concerned about Reality and making statements about Reality that are True / Honest. Agents in Level 1 seek to understand and exploit Level 0 - Reality. All Agents in level 1 trust each other. As Level-0 Reality asserts its constraints and agents face scarcity, some thus shift to... 

Level 2 : Agents are concerned about perceptions (theirs and others) of Reality, and making statements about Reality that induce perceptions about Reality that are beneficial to them. By making potentially False / Dishonest statements, Value in Level 1 is destroyed. All Agents in level 2 are parasitic on agents in level 1. As enough Level-1 agents wise up from being exploited and become Level-2 agents, some thus shift to...

Level 3 : Agents are concerned about statements of Reality. Yes, the statements themselves. Agents are concerned in making statements about Reality which are implicitly valued by having made the statement. This value cannot be derived from explicit statements because all agents distrust each other's statements about Reality due to Level-2 actions. Thus, the value of statements about Reality do not lie in what they state about Reality, but merely in that they are stated. Note that even though Agents focus on the statements themselves, agents cannot simply state any random statement about reality, as the implicit value from making the statement has to at least partially be derived from the substance of the statement as it would be interpreted in Level 1 (and 2?). As enough Level-3 agents gradually start focusing on the statements themselves, some thus shift to...

Level 4 : Agents are concerned. with everything and nothing. As every statement can be and is interpreted in a potentially infinite number of implicit subtexts, agents make statements without bearing on what they might or might not imply. No practical meaning and value can be derived from the object of statements or the statements themselves.


In the context of agents in a resource-constrained setting, one could see it as follows :

Abundant resources : Level 0 - Agents cooperate to map and exploit resources.

Constrained resources : Level 1 - Due to constraints, Agents trick each other to get a comparative advantage.

Scarce resources : Level 2 - Agents have to band together to fight other bands for resources. Agents make statements to signify their Tribe as an implicit focusing point.

Dying resources : Level 3 - Resources are too scarce to sustain a band, so it is every agent for themselves. 

Shortform on "Hedonic Collapse"

Assumptions :

  1. One is subject to hedonic adaptation.
  2. In the absence of external hedonic input, "entropy" renders one's hedonic stablepoint to be negative.

Desires :

  1. It is desired that one's response to events are temporally invariant. (All else being equal, my reaction to an event should not depend on whether I experience it today or tomorrow)
  2. It is desired to be able to forecast the (probability and impact of the) occurrence of future events as well as possible.


Given the above desires and assumptions, an all-knowing, time-invariant oracle that perfectly fulfills desires 1 and 2 will be perpetually unhappy (?).

New to LessWrong?