Self-Improving Intelligence
Theoretical models and conditions for recursive improvement.
Completion: Reading edition complete · Lean: No Lean claim · Confidence: Recorded written review
Before you start
Suggested background: Begin with the introduction and glossary. Then use the volume's own definitions to identify the background needed for its subject.
Recorded results and limits: Theoretical publication with 30 statement/proof pairs and finite reference constructions. No deployed autonomous agent or indefinite acceleration result.
Remaining work: Obtain focused review of the theorem chain and design a separate empirical evaluation of the integrated architecture.
The background advice is an editorial starting point. The work itself states its prerequisites and hypotheses.
Reading routes · Glossary · Claim links · Download this work
Read
Written proofs
Read the written arguments in the PDF. The manuscript source remains available below.
Find the statement number or section in the reading files, then read its assumptions and argument. A claim map may cover only part of the written result.
Find this work among other written proofs
Lean
The links below are guides and claim maps. No Lean source is linked here. See the evidence record for the recorded coverage.
Code
These links include code, examples, checks, or build tools. Read the edition guide for setup instructions and what each command does. All software is beta and work in progress.
Find this work among other code projects
Evidence
- Completion, Lean coverage, confidence, and cited evidence
- Source-file receipt (identity and availability)
- All files in this source edition
The evidence records describe the reviews and runs completed for this edition.