← @bgSponge

A Proposal for the Future of Scientific Communication

Summary

Paradigma proposes Flywheel, a research ledger similar to git, to replace static PDFs with transparent, machine-readable graphs that record the entire history of scientific progress and negative results.

  1. Flywheel replaces static papers.

    Unlike PDFs that package finished claims, Flywheel graphs record work in progress, including dead ends, to make the research process transparent and easier to audit.

    A graph is a record of work in progress, with the dead ends still attached

  2. Automation lowers review costs.

    Transparency allows models to spend compute attempting to break results, shifting review from manual reconstruction to verifying formalized lemmas and recorded steps.

    Making the process transparent changes the cost of checking

  3. Codex drives mathematical exploration.

    In a prototype, a Codex session using an internal math checkpoint and a Flywheel graph successfully pushed the frontier of the Riemann hypothesis research.

    This Codex session had two things at its disposal: the Flywheel graph as context, and our internal math checkpoint

  4. Small results lack completeness.

    The harness proved the 2×2 case for the Riemann hypothesis criterion, but Paradigma admits this says nothing about 3×3 or larger matrices.

    Showing the 2×2 case says nothing about 3×3 or larger

  5. Human distillation remains uncertain.

    Paradigma is unsure if future results require a human distillation step to explain meaning or if users will engage at higher levels of abstraction.

    Maybe every result of this kind should come with a distillation step

Read the original ↗Save to your Sponge

Summarized with Sponge. Quotes are the author’s own words.

More from @bg

All readings →