multi abstraction level physical simulation using blockchain forking model to reconcile abstraction error leakage w/ longest worldline winning, lower levels retroactively overriding higher when deltas accumulate to some threshold level, agent actions reevaluated against new chain
parallel evaluation of spatial regions w/ non-intersecting light comes
higher levels speculative evaluation w higher actually levels being the confirmation cycle
allowing low res predictions to be concurrently evaluated w convergence to whichever one is verified
this enables heuristic monetization as miners can compete on prediction algo, profiting off efficient system modeling; alphafold reaping rewards in protein folding simulation, fluid dynamics simulators in turbulent regions, etc
generalization of prediction markets to physics
tradeoff btwn wasted resources on truncated worldforks and preferential response opportunity; compare to chess game conditional actions
users can broadcast intent to subscribe to certain future state and miners who can construct edit path from current state earn feesm
requires a DAHG structure; directed acyclic hypergraph, as nodes necessarily inherit from multiple parents, a hyperedge capturing causality better
cellular automata has comparable structure, just without a regular grid. wolfram's hypergraph physics is similar but lower levelt
typed hypergraph allows for arbitrary abstraction while guaranteeing future reconcilibility
types also allow mapping of external systems onto a universal domain; all systems are an abstraction from physical reality and hence can be represented in terms of abstracted physics
platonic realm can be integrated as well through orthogonal abstractions; ones not mapped to any specific region of space
properly designed package manager can interact w this system by representing pure functions as mappings btwn specific types in platonic space
path to profit
open source software has no real sustainable monetization model for many reasons but in large part to heterogenous deployment targets; impossible to attribute usage and hence invoice accurately
decentralized typed hypergraph abstraction system enables this, with limits ofca
another potential usecase is in minimizing perceived realtime multilayer gaming jank; recall competing predictor model subscription earlier, ML model trained on a gamers playstyle can minimize divergence of their avatars actions in case of a network partition; think super smash
current implementation uses very simple model of players to simulate possible future during short term lag events; works but is suboptimal. possible to improve upon
comparable to deploying ML model of advertising target demo to user devices to avoid exfiltrating sensitive data
world participants can choose to subscribe to certain predictors, paying for priority access to future-verified timeline; comparable to trading alpha signals
being able to predict world allows to respond to it faster than those w/ lower risk tolerance