FIRST ORDER
See the frontier.
Identify what changed, what is genuinely accessible, and what the evidence establishes.
SECOND ORDER
See the consequences.
Construct newly valuable possibilities, their missing means, and viable productive arrangements.
THIRD ORDER
Improve the machinery.
Test better methods for discovering, constructing, and funding fresh opportunities.
Start with the paper. Inspect the evidence.
The new edition develops origin-neutral tool selection, a capability-creating decision architecture, the precise connection to MuZero, Frontier Replay, commercial participation and liquidity, and independent tests of improved invention.
To read: open the PDF above. To edit: use the LaTeX manuscript and bibliography. To reproduce: follow the README; the small analytical study needs only Python’s standard library.
Evidence, not a launch claim. The native accounts and economics are simulated. The paper does not establish real revenue, deployed open-world discovery, or general recursive self-improvement. Historical-source verification is not current-main production qualification.
What was checked
The supplied historical-source verification reports 62 native tests; the supplied Bayesian prototype passed 19; the composition/observer snapshot passed 25; and the new analytical companion passed 15. Its independent oracle agrees across 54 sensitivity cases. The supplied report records four fresh simulated accounts with accepted local audit and unchanged resume. Publication independently rechecks the 67 file hashes and the 19, 25 and 15 test suites; it does not rerun the native accounts or establish current-main compatibility.
Read research edition 2.0
Previous edition 1.1.0
The earlier 32-page edition and its public source remain available unchanged. Its Frontier Replay analysis and 68-test companion retain their original scope; they are not edition 2 verification results.