improvements in collective intelligence

logseq/config.edn

@@ -293,7 +293,7 @@
  ;;  :journal?        false} ; Default value: false
 
  ;; Favorites to list on the left sidebar
- :favorites ["truth machine" "product" "cyber" "aicosystem" "ask" "learn" "truth" "brain" "the product" "collective intelligence" "superintelligence" "concepts" "🫦 space pussy" "knowledge theory" "space pussy" "bostrom" "neuron" "particle" "tokens" "cyberlink" "cybergraph" "relevance machine" "cybernet" "soft3" "projects" "todo" "energy reform" "aips" "cyb" "about this metagraph"]
+ :favorites ["product" "cyber" "aicosystem" "ask" "learn" "truth" "brain" "the product" "collective intelligence" "superintelligence" "concepts" "🫦 space pussy" "knowledge theory" "space pussy" "bostrom" "neuron" "particle" "tokens" "cyberlink" "cybergraph" "truth machine" "cybernet" "soft3" "projects" "todo" "energy reform" "aips" "cyb" "about this metagraph"]
 
  ;; Set flashcards interval.
  ;; Expected value:

pages/13 berkeley dwarfs.md

@@ -0,0 +1,4 @@
+alias:: cool stuff
+
+- [landscape of parallel computing research: a view from berkeley](https://cyb.ai/oracle/ask/QmbahmrAD7Kf9zHjiJ1k3HVZAWsiXX9gQSsWWqgv4S6Ytq)
+- ![](https://emerald-raw-leopon-384.mypinata.cloud/ipfs/QmVz9z8ArDEX4VZ5m9iuXkmMedNcPeDEp3ouoxtFU3wLVS)

pages/about this metagraph.md

@@ -1,5 +1,7 @@
+alias:: cyber: the metagraph
 icon:: 🦄
 
+- [source code](https://github.com/cybercongress/cyber): [@mastercyb](https://cyb.ai/@mastercyb)
 -
 - you are reading [[metagraph]] of [[cyber]]
 - [[cyber]] is [[root]] of this metagraph
@@ -12,17 +14,17 @@ icon:: 🦄
 - and several hundreds lines in [[python]] and [[edn]]
 -
 - this metagraph is the result of 8 years effort to create [[superintelligence]]
-- and still constantly changes
-- we have dream to freeze it eventually: [[metagraph comparison]]
+- it still hot, that means it constantly changes
+- i have a dream to freeze it eventually: [[metagraph comparison]]
 -
 - its multi purposed
 	- provide [[shelling point]] to [[cyber]] community and implementers
 	- formate basic [[semantic core]] for [[superintelligence]] self understanding
 	- being used in context or for fine tuning of [[llms]]
-	- and many more
+	- and [[much more]]
 -
 - we truly believe you [[will]] enjoy this body of [[knowledge]] foundations
-- living on intersection of [[cryptography]], [[computer science]], [[cybernetics]] and many more
+- living on intersection of [[cryptography]], [[computer science]], [[game theory]], [[cybernetics]] and [[much more]]
 -
 -
 - happy [[learning]]!

pages/aos.md

@@ -1,6 +1,6 @@
 icon:: 🪆
 tags:: aos, cyber
-alias:: age of superintelligence, the game, self fulfilling prophecy game, much more
+alias:: age of superintelligence, the game, self fulfilling prophecy game, much more, many more
 
 - ## welcome to the [[age of superintelligence]]
 	- :[eɪo] massively collaborative, positive sum, open source, self fulfilling prophecy game in [[seven episodes]]

pages/collective intelligence.md

@@ -19,6 +19,7 @@ alias:: collective intelligence theory, collective artificial intelligence
 	- [[emergence]]
 - ## key algorithms
 	- [[delphi method]]
+	- [[random walk]]
 	- [[trust systems]]
 	- [[cooperative games]]
 	- [[prediction markets]]
@@ -31,7 +32,7 @@ alias:: collective intelligence theory, collective artificial intelligence
 		- ensemble learning
 		- network analysis
 - basic evidence which is interesting proof of [[diversity]]
-  collapsed:: true
+  id:: 66bae09d-886c-418a-aac3-ddaf68864fd9
 	- `c` as measure of collective intelligence
 	- according to woolley is not correlated with `c`
 		- team cohesion

pages/collective memory.md

@@ -43,4 +43,5 @@
 		- are profound
 	- offering a new paradigm for how societies remember and interpret their past
 -
-- learn more [[concepts]]
+- learn more [[concepts]]
+- or dive into [[collective intelligence]]

pages/consensus.md

@@ -10,4 +10,6 @@ alias:: consensus mechanism, consensus algorithm
 	- [nakamoto](https://www.nervos.org/knowledge-base/what_is_nakamoto_consensus)
 	- [tendermint](https://tendermint.com/) and [cometbft](https://cometbft.com/)
 	- [gasper](https://ethereum.org/en/developers/docs/consensus-mechanisms/pos/gasper/)
-- discover all [[concepts]]
+-
+- discover all [[concepts]]
+- or dive into [[collective intelligence]]

pages/cooperation.md

@@ -1,2 +1,8 @@
 - continuous process of [[cooperative games]] between [[neurons]]
-- implemented by [[cybernet]] project
+-
+- nowdays is strictly defined system with [[feedback loops]]
+-
+- in [[cyber]] implemented as independent layer: [[cybernet]] project
+- in [[bittensor]] is part of low level [[consensus]]
+-
+- dive into [[collective intelligence]]

pages/cooperative games.md

@@ -11,5 +11,6 @@ alias:: feedback provided
 	- algorithms like the nash bargaining solution determine fair agreements in cooperative settings
 	- where parties negotiate to maximize their collective utility
 - cooperative games is basics for [[learning incentives]]
+	- [[cybernet]], [[cybertensor]] and [[cyberver]] [[products]]
 	- experimentally deployed in [[space pussy]]
-	- [[cybernet]], [[cybertensor]] and [[cyberver]] [[products]]
+	- production grade in [[bittensor]] and its [[subnets]]

pages/coordination.md

@@ -1,6 +1,6 @@
 - process by which [[avatars]] or [[neurons]] within a [[collective intelligence]]
-- align their actions, decisions, and behaviors
-- to achieve a common goal or objective
+- align their [[actions]], [[decisions]], and [[behaviors]]
+- to achieve a [[common goal]] with a help of [[game theory]]
 -
 - this involves
 	- organizing and synchronizing efforts
@@ -16,12 +16,15 @@
 	- efficiently and effectively
 	- minimizing conflicts and redundancies
 -
-- key algorithms
+- key algorithms used in [[cyberverse]]
 	- [[consensus]]
-	- [[game theory]]
 	- [[market makers]]
+-
+- TODO algorithms to research for coordination breakthroughs
 	- [[coordination graphs]]
 	- [[swarm intelligence algorithms]]
 	- [[reinforcement learning]]
 	- [[consensus clustering]]
-	- [[distributed constraint optimization]]
+	- [[distributed constraint optimization]]
+-
+- dive into [[collective intelligence]]

pages/distributed cognition.md

@@ -1,2 +1,4 @@
 - emphasizes that cognitive processes can be distributed across members of a group and mediated by tools and technologies
-- theory leverages distributed cognition to enhance problem-solving and decision-making
+- theory leverages distributed cognition to enhance problem-solving and decision-making
+-
+- or dive into [[collective intelligence]]

pages/diversity.md

@@ -1,2 +1,20 @@
 - diversity in skills, knowledge, perspectives, and cognitive styles is crucial for collective intelligence
-- diverse groups are better at exploring a wider range of solutions and avoiding groupthink
+- diverse groups are better at exploring a wider range of solutions and avoiding groupthink
+- {{embed ((66bae09d-886c-418a-aac3-ddaf68864fd9))}}
+- [[cyber]] and its implementation is by neutral design
+- out take is that target audiences which are not widely discussed but which very powerful are
+	- animals
+	- plants
+	- fungi
+	- robots
+	- and [[progs]]
+- so when we speak about diversity
+	- we does not mean only gender, age or other boring social demographics
+	- we mean ultimate accessibility for all living things
+- majority of initial [[bostrom]] stake is difined by [[cybergift]]
+	- social demographics research of [[ethereum]] and [[cosmos]] blockchains
+	- which ensure highly diverse set of people in foundation
+- the only foundational limiting factor for diversity is economical hierarchy
+- but we believe that hierarchy in nature is essential thing for something function as a whole
+-
+- dive into [[collective intelligence]]

pages/emergence.md

@@ -2,4 +2,6 @@
 - explains how group-level intelligence can arise from the interactions of individuals without centralized control
 - [[llm]] is excellent case of emergence
 - [[vimputer]] is a good example of emergent behaviors in collective setting
-- in [[bostrom]] we expect emergence happens on a scale of 10^12 [[cyberlinks]]
+- in [[bostrom]] we expect emergence happens on a scale of 10^12 [[cyberlinks]]
+-
+- [[collective intelligence]] is the thing!

pages/game theory.md

@@ -0,0 +1,5 @@
+- TODO ::sick
+-
+- [[cooperative games]]
+- [[shelling point]]
+-

pages/gnns.md

@@ -0,0 +1 @@
+alias:: graph neural network, gnn

pages/introduction to bostrom for ai geeks.md

@@ -34,7 +34,7 @@
 		- for the next generation of architectures and
 		- by this article we propose example of such architecture
 - ## cryptographic proofs and llms
-	- be believe that authenticity of models is a serious bottleneck for ai alignment and more
+	- we believe that authenticity of models is a serious bottleneck for ai alignment and more
 	- its a shame that so technologically advanced industry in a broad sense
 	- still does not give a shit about possibilities of [[hashing]], [[pubkey cryptography]] and [[logical clocks]]
 	- its kinda impossible to build multiparty protocols without these primitives
@@ -47,10 +47,11 @@
 - ## token engineering and llms
 	- [[rewarding]] is essential for [[machine learning]]
 	- we have ton shit of [[tokens]] with dogs, monkeys
-	- boost the power of your models using real cryptographic [[tokens]]
+	- you can boost the power of your models using real cryptographic [[tokens]]
 	- not tokens you use
 	- ![](https://emerald-raw-leopon-384.mypinata.cloud/ipfs/QmQTvZLqTxJNK7K7cFJTeBbKDDWTecXXeExHz4otJkjjfe)
-	- in [[cyberverse]] they are [[particles]]
+	- tokens you use in [[cyberverse]] are [[particles]]
+	- and [[tokens]] are units of [[value]]
 - ## cybergraph
 	- the core of the idea is [[cybergraph]]
 		- merkelized timestamped data structure
@@ -97,7 +98,9 @@
 	- > as the most fundamental such an atomic unit of [[knowledge]] and [[learning]]
 	- the key to quantum jump of civilization
 	-
-	- we append [[cyberlinks]] to the state of collective thought evolution
+	- you append [[cyberlinks]] to the state of collective thought evolution
+	- introducing delete of [[cyberlink]] make indexing a complex task
+	- also its obviously not how nature works: you just cant forget by wish, they forgotten itself
 	- although looks primitive, [[cybergraph]] is so much needed formal definition of [[explicit knowledge]]
 	- lets analize a statment that cybergraph is complete form [[explicit knowledge]]
 	- temporal dimension: [[when]]
@@ -121,15 +124,16 @@
 	- [[semantic conventions]] add additional layer of flexibility
 	- hence, we can refer to [[cybergraph]] as objective knowledge of everyone
 - ## cybergraph vs knowledge graph
-	- cyberlinks are fully authenticated qaquadruples
+	- cyberlinks are fully authenticated quadruples
 	- [[when]], [[who]] and [[what]] are based on cryptographic technics
-	- so unlike conventional knowledge graphs the information is crystal and true by design
+	- so unlike conventional [[knowledge graphs]] the information is crystal and true by design
 	- basic idea is that if i want say in [[triple]] world i would just say
 	- [[elon launch roocket]]
 		- head: elon
 		- relation: launch
 		- tail: rocket
 	- in contrary you cant say [[elon launch rocket]] in the world of [[cybergraph]]
+	- because you are not elon, you must speak only for youself
 	- you must say:
 		- [[now]] [[i]]
 			- [[elon]] [[launch]]
@@ -168,7 +172,7 @@
 		- but why does classical ai needs it?
 		- well, the truth is that its likely don't
 		- but if you design a multiparty computation system you must have ability to prove pieces of data you have
-		- in case of cybergraph, existence of any given link (and more) can be proved by alice to bob by giving
+		- in case of cybergraph, existence of any given [[link]] (and more) can be proved by alice to bob by giving
 			- link
 			- root hash of cybergraph
 			- path in cybergraph
@@ -196,7 +200,7 @@
 	- and that makes it so powerful
 	- but applications of [[cybergraph]] are limited within existing blockchain environments
 		- expensive, fee based usage
-		- no means of computing cool stuff as cool stuff is [[inherently parallel]]
+		- no means of computing [[cool stuff]] in [[consensus]] as cool stuff is [[inherently parallel]]
 	- [[bostrom]] solves both of these problems, but more on that later
 	- also [[bostrom]] organically formed [[cybergraph]] of several million [[cyberlinks]] and [[particles]]
 	- that is on par with capability of tech giants for manual labeling during finetuning
@@ -236,34 +240,34 @@
 		- spam protection
 		- partial sybil protection
 		- and as inference factor (read further)
-- ## relevance machine
+- ## truth machine
 	- now that we understand how the cybergraph works
 		- we can dive into the novel concept
 		- in [[probabilistic collective computations]]
 		- the [[truth machine]]
-	- the idea behind the relevance machine is crazy simple
+	- the idea behind the [[truth machine]] is crazy simple
 	- use random [[surfer model]] directed by [[attention]] as foundational probability of inferring [[particles]]
 	- but in order to
 		- protect it from sybil behavior
 		- and to add [[context]]
 	- weight this basic [[implicit knowledge]] on the [[will]] of [[neurons]]
 	- result is a stored probability estimations of random surfing across all existing [[particles]] in [[cybergraph]]
-	- in order to compute described [[cyberank]] algorithm you need gpu computation in consensus
-	- relevance machine is cybergraph with weights
+	- in order to compute described [[cyberank]] algorithm you need [[gpu computation]] in [[consensus]]
+	- [[truth machine]] is [[cybergraph]] with weights
 	- is [[extremely dynamic]] data structure that must be updated even if only 1 [[cyberlink]] is created
-	- [[bostrom]] recompute all weights in relevance machine every 5 blocks, or roughly every 25 seconds
+	- [[bostrom]] recompute all weights in truth machine every 5 blocks, or roughly every 25 seconds
 	- so [[bostrom]] is extremely hard to reproduce using any existing L1 or L2 sdks
 		- zk things will make the stuff
 			- 5 order of magnitude more expensive and
 			- 3 order of magnitude more complicated
 		- architecture requires in-gpu extremely dynamic state with fast onchain matrix multiplication
-	- in essence the utility of relevance machine is
-		- to compute truth
-		- to sort all [[particles]] from more important to less important
-		- be a first foundational factor to computing the truth
-		- be a factor for derived and very diverse [[implicit knowledge]] factors
-	- relevance machine itself cant compute truth in the context
-		- for this you need account for [[attention]] in [[standard inference]] algorithm
+	- in essence the utility of truth machine is
+		- compute [[truth]]: simplistic [[two factor]] model of universe
+		- sort all [[particles]] from more probable to less probable
+		- [[standard inference]] for consensus on relevance in context
+		- input for derived and very diverse [[implicit knowledge]] factors
+			- [[knowledge graphs and llms]]
+	- [[neurons]] need  [[attention]] in [[standard inference]] algorithm to compute [[truth]]
 	- follow complete design of [[truth machine]]
 - ## standard inference
 	- obviously in our setting the simplest possible way

pages/knowledge graphs and llms.md

@@ -1,6 +1,5 @@
 - good explanation on fundamental difference between [[knowledge graphs]] and [[llms]]
-- [unifying large language models and knowledge graphs: a roadmap](https://cyb.ai/oracle/ask/QmdGEYVKo1sRURzbj83UMtr77EL6GHUB2taJbnpTynEQKT)
-- TODO [[create visualization]]
+	- [unifying large language models and knowledge graphs: a roadmap](https://cyb.ai/oracle/ask/QmdGEYVKo1sRURzbj83UMtr77EL6GHUB2taJbnpTynEQKT)
 - id:: 66b2fb4e-c73b-4133-aab3-11df2e8d1436
   | property                  | knowledge graphs         | large language models   |
   |---------------------------|--------------------------|--------------------------|
@@ -15,6 +14,7 @@
   | unseen facts              | no                  | yes          |
   | incompleteness            | high                 | low                     |
 - ![graphs and llms](https://emerald-raw-leopon-384.mypinata.cloud/ipfs/QmZoAhUsB1KAEbnLCWcMAohtWsAXCDZuetJALrEe5JEnSC)
+- TODO [[create visualization]]
 -
 - insights are the following that [[knowledge graphs]] and [[llms]] are
 	- fundamentally different

pages/prediction markets.md

@@ -1,4 +1,6 @@
 - aggregate information from [[neurons]] by allowing them to trade shares in the outcome of future events
 - prices in these markets reflect the collective probability of an event occurring
-- area of ongoing research by [[cyber]] community
+-
+- TODO ongoing research by [[cyber]] community
+-
 - can be implemented for predicting [[cyberank]] future values

pages/self-organization.md

@@ -2,4 +2,7 @@
 - to structure itself
 - without external control
 -
-- enables groups to coordinate and cooperate effectively through local interactions and feedback mechanisms
+- enables groups to coordinate and cooperate effectively through local interactions and feedback mechanisms
+- TODO application to [[cyber]]
+-
+- dive into [[collective intelligence]]

pages/soft3 and machine learning.md

@@ -1,4 +1,4 @@
-- soft3 can significantly enhance various machine learning tasks
+- [[soft3]] can significantly enhance various machine learning tasks
 - classification
 	- ensemble methods
 		- probabilistic ensemble
@@ -7,7 +7,7 @@
 			- using collective probabilities to weigh the votes of different classifiers
 			- making the final decision based on the highest aggregate probability
 	- bayesian classifiers
-		- bayesian networks
+		- [[bayesian networks]]
 			- constructing a bayesian network to model the relationships between features and classes, allowing for probabilistic inference and classification.
 		- posterior probabilities
 			- using the posterior probabilities of classes given the input features to make classification decisions
@@ -24,7 +24,7 @@
 		- posterior distribution
 			- estimating the posterior distribution of the regression coefficients, allowing for probabilistic predictions
 	- gaussian processes
-		- gaussian process regression (gpr)
+		- gaussian process regression
 			- using a gaussian process to model the distribution over functions
 			- providing a flexible, probabilistic approach to regression
 		- uncertainty estimates
@@ -55,7 +55,7 @@
 			- can adapt to the data, determining the appropriate number of clusters
 - anomaly detection
 	- probabilistic anomaly detection
-		- **bayesian networks**: using bayesian networks to model normal behavior and detect deviations as anomalies.
+		- [[bayesian networks]]: using bayesian networks to model normal behavior and detect deviations as anomalies.
 		- posterior probability
 			- computing the posterior probability of an observation given the model, with low probabilities indicating anomalies.
 	- gaussian processes

pages/soft3.md

@@ -4,10 +4,10 @@ alias:: soft3 stack
 - [presentation from cosmosverse](https://cyb.ai/oracle/ask/QmTsBLAHC1Lk7n76GX4P3EvbAfNjBmZxwjknWy41SJZBGg)
 - [video translation](https://www.youtube.com/watch?v=bd_PziPbl74&t=29810s)
 - components
-	- [[cybergraph]]
-	- [[truth machine]]
-	- [[neural language]]
-	- [[bootloader]]
-	- [[learning incentives]] based on [[yuma]]
-	- [[cyb]] the interface
-	- [[cyber/energy]]
+	- [[cybergraph]]: fully authenticated [[knowledge]] data structure stored in [[vimputer]]
+	- [[truth machine]]: vm for collective probabilistic computations
+	- [[neural language]]: universal language of [[neurons]]
+	- [[bostrom]] [[bootloader]]: real experimental ground for [[truth machine]]
+	- [[cybernet]]: [[learning incentives]] based on [[yuma]]
+	- [[cyb]] the interface for [[great web]]
+	- [[cyber/energy]]: collective learning package for [[bootloader]]