Fragmentation of our world view is the underlying problem leading to our problems in society, ability to identify with one ingroup and go into rivaly with another outgroup, optimizing particular metric while externaliting harm to another metrics, the ability to benefit the near term while damaging the long term, all of those are special cases of percieving parts rather than wholes, which ended up actually damaging the wholes and parts themselves through the interdependence of parts. We must take on account all technological, environmental, economical, geopolitical variables. Clusters of people compete for exponentially growing population resource growth through wars or cooperation in terms of power. Thanks to industrial revolution we have nash equilibrium with nukes and globalized economy, cold war. Those are more monitorable through satelites. More than exponentially growing democratized bio/cyberweapons with linear nature resources might destroy this. Two attractors that the society is heading to: Decentraliztaion - less regulatory control over groups with exponential tech leading to higher chances of existencial risks, but more freedom to do stuff, ancap dystopia? Centralization - can catch and deal with the landscape of existencial risks but too much regulations, potential for authoritian regime, centralized power, uneven conditions, it's not clear that what will be maximized is power of the top class instead of survival of the species anyway Evolutionary competition happens either way, between clusters of centralized groups with exponential tech maximizing their short term safety at the cost of long term global safety because of finite resources and for living stable conditions of the environment. Solving multipolar traps: What do these two attractors of catasrophies and dystopia have in common? Distrust, opaqity, secretness between agents... Democracy aka parcipitatiry governance doesn't work with secret goverment information about the state not shared to citizents because of fear of information disadvantage with rivals. Potential solution: Full trust and transparency between agents' doings/insides? Sweden military seems to partially do this, with less money being spent on making information secret and it seems to work. It can lead to information disadvantage, less asymmetries generating power over the other. Forced global transparency sidesteps this as everyone gets the same disadvantage resulting in more global safety with less overall tension. Absolute symmetry? Results in global coordination on global survival? If only people implicitly synchronized. Increasingly smaller actors can use destructive technology in increasingly destructive ways. How to create global survillance to manage the interconnected landscape of existencial risks that's not dystopic Orwell style? China is banning polarization technologies (social media) for example, regulating memeplexes. Honestly how synchronized they are is sometimes fascinating. But from our point of view and many of them its very dystopic. Global existencial risks are still a thing from negative sum games between superpowers. Shifting society to language of unified meaning and values instead of polarizing identity boxes. We dont want short term loosing to other agents or long term selftermination. We dont want cascading catasrophies (climate change, weaponized technology,...) and clear dystopia (asymmetries of power, almost everyone has almost no freedom) exponential tech -> decentralized power -> coordination failures, benefiting agent's safety short term at the cost of everyone, multipolar traps -> increasing catasrophies -> Amazon versus Microsoft's military ancap wars as there's no monopoly on violence (police) exponential tech -> centralized power -> AI monitoring sensors everywhere -> potential for Orwell autocratic feudalistic dystopia Sweden: Assuming rivals will probably find out secret goverment information anyway -> not investing any resource in hiding opaqueness -> population are what's going on in the country -> population has more trust in government -> less money spent on champaigning and convincing people of things, more emergent coordination, various goverment deparments and military know more what eachother is doing (less information transmission blockages, more coherence) and can more easily coordinate, corruption is also more seen thanks to the transparency and can be dealt with -> some information asymmetry disadvantage related to other rival countries but a lot of other disadvantages, different local minima valley in the adaptive landscape Too much information though leads to not everyone being able to process it so it doesn't matter of its transparent or not, information singularity, can ve processed via computational compression: AI summarization for example. It seems that something similar Is happening in Czechia, the new Czech president literally has a YouTube podcast. The trust in president got higher here but the trust in government apparently lowered (it seems that there are more left leaning people in the population who's values aren't that represented in the elected government because of bad weird rules and bad luck in last elections thanks to formings of coalitions to succesfully take down a populist corrupt oligarch) I love how internal symmetrification of information works for both the mind and the nation to achieve better health Better communications between government deparments is same as better communication between brain centers For USA: More strategic military capacity could emerge our of transparency -> leading to Russia and China trying to do the same thing to not lose the multipolar trap of racing to the top of transparency rather than race to the bottom of opaqueness -> more potential for international agreement and fundamentally adressing the multipolar trap aggressively better Multipolar traps are scenarios where the things that work well for individuals locally are directly against the global well-being. Forced transparencies strategies making them more capable of winning: International satelites, Opensource This culture more oriented to peacefulness could beat the counterculture? Becoming more warring without becoming more warring in all the most problematic ways Could this symmetry voulnerability create much more easier conditions for negative actors tho On the other hand open source software could be prefixed by collective of positive actors before negative actors finding out Decreasing pathological competition within the system Maintaining adequate strategic capacity More positive sum games Something that wont make one agent's start of a cascade make him not lose to other agents and influences the rest of the world Enlightened country is nice but it will still die to global issue of climate change thanks to USA/Russia/China Nature of what it makes it win isnt externaliting harm or driving arms races Obseleting destructive game theory And winning at some fundamental game theory at the same time More transparency -> more coordination and less collective action failures, less incentive evolutionary pressure for psychopaths, sociopaths, narrcisists, worse conditions for hiding the stuff that one is doing Transparency is hard to scale because of those negative actors Collective intelligence technologies such as interconnected global internet communication to the rescue? Noticing perverse incentives, accounting them and externalities, and creating progressively better incensives and more capable detterence to align the motivation of individual agents with the whole better Some of the exponential computational technologies that portend the fastest and worst existencial risks in many ways also portend possibility for better coordination systems Not AGI singleton disintermediating humans and running the world But AI and other capacities being to fascillitate collective intelligence. AI not making decisions by itself but processes of collective intelligences of humans making it. People voting yes no on binary propositions where both sides of the propositions suck. If it goes through it benefits something and harms something else. That kind of proposition is designed badly and doesnt factor how interconnected everything is so they yes no cant not polarize the population, that would be a stupid system. Before making the proposition, doing sense making what are all the interconnected things, what are all the values you take as design constrains to go through a better proposition crafting process of what is the best synergestic satisfier with the least theory of tradeoffs possible and what are better voting systems than binary that inherently polarize the population. More education in topics, incentifying more ability to influence those topics How to solve information singularity, more information than anyone can process? More journal articles on a topic than any expert can read so nobody can be an expert on anything. How to solve it? Make AIs run the world? Merge with AIs? Value systems have to bind guide and direct the technology and influence social systems that create the capacity to bind guide and direct technology and markets in the way that they are not currently doing it. Technology is a new capacity and that capacity if rightly directed can make new things possible. Steer global society away from selfdestruction towards consciousness utopia! Choices that take our attention in phones or capitalism hijack our basic needs instead of aligning with our highest goals. Use wisdom to develop wisdom in people through designing technology. Can we make the types of transparency the types of incentive alignment the types of capacity for deterrence that would exist at small tribal type scale possible at much larger scales, where the intcrease in effective coordination starts to be more effective, reducing huge amount of waste and duplication and failures of those kinds. What makes this thing adaptive and selective in game theoretic situation is also what makes it healthy in terms of wellbeing of people inside and relative to eachother.<ref>[https://civilizationemerging.com/ Daniel Schmachtenberger]</ref> Solving meta and meaning crisis by increasing collective transparency, trust, compassion, happiness as reaction for others' happiness, reducing addiction, where addiction is defined as selfdestructive/compulsive behavior that goes against one's goals and decreases cognitive flexibility. In order to keep wellbeing even in the chaos, just vibe with the environment and maybe embrace omnipotent misaligned AGI in climate change in collapsing civilization while helping remaining people survive with loving joyous compassion while in pure blissful peaceful equanimous stressfree vast spacious accepting state of mind. Andres mapped out the space of clusters of human emotions and transition functions between them with data science. Rebuilding meaning group mapped out the space of possible meanings corresponding to our values encoded in a language model. Put this space of our values into existing AGI language models. Encode it in maths used to model biological systems as fundamental priors corresponding to neural correlates and synchronize it with biology inspired selforganizing AGI - Selforganizing optical spiking transformers? Can we take the mathematical models on how people synchronize in neuroscience (vibing by same resonant wave modes) and use them to build artificial agents such that they synchronize with us. That could constrain their statespace of solutions to more values aligned attractors. It seems to me that mathematics of Active Inference combined with Connectome Harmonics is a very promising path towards aligned AGI. Because we can literally copy the brain (consonance, synchrony of bayesian values) as the same maths is used for brain and AI modelling. Every cell is a reinforcement learning agent on its own, not just neuron cells. But neurons communicate in long distances with axons like telegraph. One can do learning global functions that tell neurons how to respond to patterns in their neighborhood. <ref>[[Generalist AI beyond Deep Learning - YouTube](https://youtu.be/p-OYPRhqRCg?t=2825) Generalist AI beyond Deep Learning]</ref> These neurons as continuous Markov blankets? The whole thing as continuous neural celluar automata responding to input sending actions, to allow brain's resonant modes wave propagation computing? '''ab. Philosophy and other things '''abc. Mental constructions''' '''abcd. Truth''' What is truth? In the feelings of something being deeply true: The more overall consonant/symmetrical the dynamics (that the mental state corresponds to) is, the truer it feels, the more the insight wants to stick (how the definition interlocks with the whole mental framework is important) mathematicians: its derived from these axioms under these rewriting rules engineers: boom! it fookin works, i got Doom to run on this alien computer using selfconstructed biology inspired bioelectromagnetic selforganizing system with adaptive hiearchical crossscale harmonic bayesian beliefs with goals and actions 🥳 What seems true to us can be seen as sensory patterns that satisfy our learned neural classifiers aka energy sinks the most, that generate the highest positive valence (consonant dynamics) in the mental framework (background resonant wave modes) that we are inhabiting, that we annealed to, what software we or the environment installed to us.<ref>[[Subjective signal strength distinguishes reality from imagination | Nature Communications](https://www.nature.com/articles/s41467-023-37322-1) Subjective signal strength distinguishes reality from imagination]</ref> If the world model feels stronger, we assign world as ontologicaly primal. If the self model feels stronger, we assign consciousness as ontologicaly primal. If its equal or one signal, it merges or transcends. You have a thesis and an antithesis and then Hegel's idea is you have a synthesis which is a kind of higher order truth that understands how those relate in a way that neither of them do and its at a higher order of complexity. The first part would be: Can I steal man each of these, can i argue each one well enough that the proponents of it are like: "Totally you got that!" and not just argue it rhetorically, but can I inhabit it where i can try to see and feel the world the way someone seeing and feeling the world that way would, because once i do then, I don't want to screw those people, because there's truth in it, and i'm not going to go back to war with them, I'm going to go to finding solutions, it could actually work at a higher order. If i don't go to a higher order, then there's war, the higher order approach can generate cooperation, coordination, peace, love, friendship, connection, mutual understanding, empathy, through complexity. [[File:69.png|500px|Alt text]] “It seems plain and self-evident, yet it needs to be said: the isolated knowledge obtained by a group of specialists in a narrow field has in itself no value whatsoever, but only in its synthesis with all the rest of knowledge and only inasmuch as it really contributes in this synthesis toward answering the demand, ‘Who are we?'” – Erwin Schrödinger in Science and Humanism (1951) It follows from the Free Energy Principle that much of the functional and anatomical hierarchical structure in the brain reflects hierarchical structures in nature and culture of which we evolved to keep track. https://twitter.com/JakeOrthwein/status/1650589310939451393? '''abc. Final contemplation about physicalist and mental theories of everything and oneness''' The deeper you go into deflatory more predictive models about empirical data, the less intuitive (comprehensible for our usual by evolution preihstalled mental data structures) one becomes when trying to communicate it to others ((ontology agnostic) quantum mechanics, Amplituhedron) In this book i mostly reason from ontology agnostic classical thermodynamics of bayesian statistics, but adding some more Quantum theory (Chris Fields) or Amplituhedron (going beyond spacetime, which is not fundamental but emergent there) would make it extremely hard grasp for the classical neural classifiers in our mind. Its all just in different contexts useful labels for models for predicting data, and their validity is being supported by empirical data. In a certain lens when constructing models we are approximating the brain dynamics trying to predict our sensory data and those brain dynamics have an extreme complexity and the more you try to describe the nature of experience, the more iffeable it seems. We may never fully know the experiental or scientific truth as we may never construct a model that predicts all empirical data. Easiest theory of everything for free is taking the finite set of all measurments that we got recorded so far and create input output lookup table from it and that's it. Then slowly deflate and compress it by various symmetries. Machine learning is also helping us find patterns in dynamical systems. There may be different theories of everything, with many isomoprhisms, and their count depends on the amount of abstraction of the different theories of everything. It's all many lenses useful for different domains going from different angles. All of this is for engineering useful set of interacting data structures in our experience. Ontology agnostic bayesianism aka metarationality doesn't care what is fundamental, if everything is matter or consciousness or both. Freedom from fundamentality of fundamental symbolic concepts "is fundamental" including freedom from this freedom! Philosophical problems as linguistic confusion? Everything is just a nebulous dream-like rainbow-like mental construction. No source, destination, concepts. Groundless ground. Emptiness is empty too. If you think you understand emptiness then you don't understand emptiness. In another lens this mental construction is the most useful one for generating constant sense of liberation when it comes to concepts. For futher liberation do some meditation. "Truth" is end of mind, end of knowledge, and that's freedom. Metaalgorithm that when applied to all mental constructions creates the most satisfying total mental framework as it creates symmetries between all dualities, language, concepts. It dissolves all abstract thought in general. =Unfiltered unsorted chaos= [Episode 45: Leonard Susskind on Quantum Information, Quantum Gravity, and Holography - YouTube]([Episode 45: Leonard Susskind on Quantum Information, Quantum Gravity, and Holography - YouTube]([Episode 45: Leonard Susskind on Quantum Information, Quantum Gravity, and Holography - YouTube](https://www.youtube.com/watch?v=XTSdPSOcdjI))) Sean Carroll [[2112.10510] Transformers Can Do Bayesian Inference]([[2112.10510] Transformers Can Do Bayesian Inference](https://arxiv.org/abs/2112.10510)) Transformers Can Do Bayesian Inference What is convolution [But what is a convolution? - YouTube]([But what is a convolution? - YouTube]([But what is a convolution? - YouTube](https://www.youtube.com/watch?v=KuXjwB4LzSA)&t=693s)) “Over the next 10 years, we are aiming to build a fully autonomous AI Scientist, and to use that AI Scientist to scale up biology research.” [Future House]([Future House]([Future House]([Future House](https://www.futurehouse.org/articles/announcing-future-house))))