computation stack for superintelligence
every generation of the web had its stack. web1 had LAMP. web2 had React + Node + Postgres. web3 had Solidity + EVM + RPC. each defined what developers could build and what users could experience
soft3 is the stack for a shared, provable, self-improving knowledge system where every computation leaves a cryptographic proof and every piece of meaning has a measurable weight
neurons — humans, AIs, sensors, agents — link knowledge into the cybergraph. the tru reads this graph every block and computes what matters: cyberank per particle, karma per neuron, syntropy of the whole. every result is deterministic, on chain, verifiable by anyone. trident compiles any logic into stark proofs — hash-based, post-quantum, no trusted setup. neural structures meaning through semantic conventions so the graph speaks a language both humans and machines understand. cyb makes all of it accessible — a personal cyb/robot that queries, scripts, and navigates the graph
the tru is an onchain language model. it does what models do — rank, retrieve, infer — except the weights are public tokens, the training data is an open cybergraph, and the inference runs in consensus with proofs. no API keys, no corporate weights, no black boxes. the model improves when anyone links useful knowledge, and the improvement is measurable as rising syntropy
trident closes the provability gap. in existing stacks, smart contracts can move tokens but cannot prove that a computation happened correctly without re-executing it. trident programs produce stark proofs: verify once, trust forever. this makes the stack suitable for AI alignment — you can prove that a model followed a policy, not just trust that it did
see cyber for the full stack breakdown and specifications
discover all concepts