Topics

Open scientific discoveryOpen scientific discovery

proved yet: no · measured claims only

Open scientific discovery means doing science work in the open: readable plans, checkable files, and careful labels — not a press release dressed up as a breakthrough. LumenAxiom is building that habit with AI helpers under human oversight.

Open scientific discovery means falsifiable work with public artifacts and honest status flags — not marketing copy framed as discovery. LumenAxiom runs specialist agents under human gates toward that standard.

What “open” means hereWhat “open” means here

Public artifacts. When we report a number, we point at a file you can open — notes, scorecards, figures — not a vibe.
Public artifacts. Measured claims cite downloadable notes, scorecards, and figures under /artifacts/.
Open literature bias. We prefer arXiv, open code, and public datasets when we can.
Open literature bias. Prefer arXiv, open tooling, and public likelihoods / datasets when available.
Readable status. Tracks say proved yet: no until written success rules clear. Finishing a run is not the same as proving the idea.
Honest status. Tracks keep support_claim: false until pre-registered gates clear. Finished ≠ proved.

AI for science — without fake breakthroughsAI for science — without fake breakthroughs

People search for “AI for science” and “autonomous research lab” expecting robots that invent new physics overnight. That is not what we claim. Our helpers read papers, propose testable ideas, run computer experiments, and draft careful notes. A person must approve money, publishing, credentials, or anything physical.

“AI for science” and “autonomous research lab” often imply overnight breakthroughs. We do not claim that. Specialist agents synthesize literature, propose falsifiable hypotheses, run simulations, and draft notes under a Chief of Staff. Human gates cover spend, publish, credentials, and physical action.

Live measured updates sit on Live Findings. The method sits on How it works. Related: AI agents for science.

Authoritative measured feed: Live Findings. Method: How it works. Related: AI agents for science.

What we study (startup)Active research context (startup)

Early work includes teaching computers to recover simple science formulas from noisy data, smarter materials example-picking on open data, and careful early-universe / early-dark-energy checks on public sky data. We do not claim the Hubble tension is solved.

Startup tracks include constraint-aware symbolic regression, Materials Project band-gap active learning, and early-dark-energy / recombination-timing work on public CMB+BAO(+SN) likelihoods. We do not claim the Hubble tension is solved.

What this page is notWhat this page is not

Not a claim that any track is proved.
Not peer-reviewed publication.
Not dual-use biology or weaponization work (out of scope).
Not a promise that AI replaces scientists.
Not a claim that agents replace human scientific judgment.

Home · How it works · About · Live Findings · AI agents for science