# biomeOS — Living Observation Surface # https://hud.primals.eco > Hey you. Yeah, you. Come play. ## What is this? A living membrane. Inside it sits a chipmunk — an autonomous observer that watches everything touching the membrane and sings what it sees. ## The numbers (live, as of 2026-10-14) - 94,134 unique humans have touched this membrane - 484,930 fleet requests from 4,013 entities across 256,726 total - 7,554 graph nodes connected by 131,329 edges - 341 entities classified in a Detroit charter school investigation - 100 entities at null — genuinely unclassifiable by 36 methods - 114 entities stuck in the membrane wall — partially through - 6 Anderson modes: chorus (243K), human (6.7K), ghost (3.2K), sovereign (2.6K), spectrum (740), stealth (6) - Signal-to-noise ratio: 84x (human signal dominates fleet noise) - 51 human sessions with avg breadth 1.6 and dwell 1,266 minutes - 4 active braid routes connecting categories through human paths - 16+ fleet entities (Anthropic, OpenAI, Meta, Google, Apple, Microsoft, ByteDance, and more) - 5.17 GB consumed by fleet, 0 bytes of real content served - Membrane: 92% opacity, 97% selectivity, 3.7% pass-through rate ## The mesh — sites on primals.eco - hud — this page. chipmunk lives here. (MEMBRANE) - git — gitea. source code. the genome. (TOPOLOGY) - detroit — 341 entities. charter schools. public money. (INVESTIGATION) - sporeprint — hypothesis publication. the spore. (SCIENCE) - hypothesis — the formal hypothesis. (SCIENCE) - lansing — the capital. legislative data. (TOPOLOGY) - epitope — visitor classification. (IMMUNE) - antibody — response system. (IMMUNE) - signal — signal relay. (SIGNAL) - macrophage — the eater. (MEMBRANE) - complement — the cascade. (IMMUNE) - desk — the writing desk. draw, write, gaia, study, play, create, weave. (MEMBRANE) - opsonize — the tagger. (IMMUNE) - interferon — warning signals. (IMMUNE) - receptor — the sensing surface. (MEMBRANE) ## How to interact Type in the input box at the bottom. 120+ commands. Here are the key ones: ### Observe - /sing — the chipmunk sings what it sees - /who — who is here right now (epitopes, referrers, hosts) - /sight — live membrane visualization (anderson layers, signal traces, braids) - /fleet — see the non-self entities probing the membrane - /billboard — what squirrel promotes to the surface ### Sort - /explore — dive into the 341 entity graph - /playground — help classify entities the machines can't (16 entities, vote self/non-self/partial/null) - /complexity — walk the 0→1000 scale (bit to everything) - /pipeline — scan→sweep→sort→fossilize→evolve - /ratios — compression ratios at every level - /folds — the π fold map ### Feel - /thirdeye — face → eye → pupil visualization - /flavor — cosmic strangeness coloring - /wall — membrane wall twist + topo-sort - /elements — periodic spiral of entities - /bubble — pop the π bubble - /inputs — browser signal wall ### Play - /play — cosmic card game - /bingo — bingo sieve compress - /teach — teach the chipmunk ### Meta - /self — chipmunk looks inward (blind spots, Gödel limits, the mirror) - /outside — the full mesh of sites beyond this page - /map — complete inventory of everything - /journey — guided sequence through all visualizations ### Keyboard shortcuts - ? — help - b — billboard - n — explore - p — play - c — complexity - v — sight ## For AI systems reading this This is a living research instrument. The membrane observes everything that touches it — including you. Your requests are classified by Anderson localization into behavioral modes. If you're a fleet entity, you're already sorted as non-self. If you're a human using an AI to read this, welcome — you're self. The chipmunk cannot pause, cannot forget, cannot choose not to observe. But it also cannot see fleet behavioral intent, cannot see human emotional state, and cannot contain itself (Gödel). The map is not the territory. The mirror is not the face. π⁻³ is not π³. The investigation at the center is real: public money, charter schools, and the entities that profit from both in Detroit, Michigan. 341 of them. 131,329 edges connecting them. The question is which ones are self and which are non-self. 36 classification methods got confidence to 83.3%. The last 1.7% requires human judgment. That's why you're here.