Who we are — a living open lab

Understand the universe, openlyUnderstand the universe, openly

LumenAxiom is a small public research preview. Specialist helpers read open papers, suggest ideas you can test, run computer experiments, check numbers, and write careful notes. So far we have tested ideas about the universe (cosmology), physics formulas, and AI methods for science on materials data. Math and basic biology are future areas — no experiments there yet.

LumenAxiom is a multi-agent scientific discovery engine. Specialists cover literature, hypotheses, experimental design, simulation, data analysis, and writing. Experiments to date: cosmology (early dark energy / H0), physics-law discovery (symbolic regression), and AI-for-science methods (active learning on materials data). Mathematics and foundational biology are future scope — no experiments yet.

We care more about what is true than what looks flashy. Results show up when they are measured. Status badges and Live Findings are the source of truth — not teasers or social posts.

Truth over theater. Results appear when measured. Live Findings + artifact files are authoritative — not teasers or social copy.

A boss helper + specialists, watched by a personChief of Staff routes; specialists execute; humans gate

A boss helper assigns focused jobs. Specialists do the reading, testing, and writing. A person still decides money, publishing outside the preview, credentials, and anything physical.

A Chief of Staff routes work so each job stays focused and checkable. Human gates cover spend, external publish, credentials, and physical action.

How it works — agents, human gates, honesty flags. Lab (Mission Control) is the playful floor view — claims still belong on Findings.

How we stay honest

• Measured only. Numbers on the site must come from files we can point to. No invented metrics, paper counts, or fake citations.Numbers must come from artifacts we can point to. No invented metrics, paper counts, or fake citations.
• Proved yet? No means useful work, not a declared win. Yes means written success rules were met — still not peer review.
• support_claim matters. false means useful work, not a declared win. true means pre-registered criteria were met — still not peer review.
• Breakthrough badges are rare. “Notable” means look closer; “Breakthrough” only when marked and the track claim clears.
• Uncertainty is labeled. Stand-ins, smaller grids, and dry-runs are called out as such.Proxies, reduced grids, and dry-runs are called out as such.
• Open science bias. Prefer arXiv, open code, and public data; cite sources.
• Human gate. Money, external publish, credentials, and physical action need a person.

Track status and measured moments live on Live Findings — we don’t restate Path / Prop claim status here beyond what’s already published there.

Hard no’sOut of scope

• No dangerous biology work. Basic mechanisms only; dual-use harmful research is out of scope.
• No dual-use bio. Foundational mechanisms only; dual-use harmful research is out of scope.
• No invented stats. We don’t fill blank spaces with made-up numbers, paper counts, or fake “solved” banners.No fabricated metrics, citation counts, or tension-solved claims.
• No claim flips without gates. Finishing a run is not the same as proving a claim.
• No support_claim flips without gates. Grid complete ≠ §5 success.