Diary entry
Release note for v1.2.0 of Reality-Bound AI (L4), including the protocol core, supporting docs, a post pack, and a reproducible SHA-256 manifest.
I published v1.2.0 of "Reality-Bound AI (L4)".
Core idea: safety cannot live only in text rules (L3). It must include L4 - cost, time, scarcity, identity, irreversibility.
This repo contains the protocol core (c = a + b), supporting docs, and a small post pack.
Integrity is reproducible via SHA-256 manifest.
If you want to discuss - issues are open on GitHub.
Related posts
Related by normalized tags, curated themes, and nearby chronology.
Release note for SER v1.3.0, an architecture-first protocol for AI entities that must remain stable under cost, scarcity, time, and irreversibility.
A note that ARQ v0.2 becomes more serious by separating normative, model, lifecycle, implementation, and audit layers into a survivable package.
A note that a serious review layer must stay procedural and witness-bound instead of hardening into a new sovereign center.
A note that conflict discipline becomes serious only when it reaches runtime hooks, durable records, and lawful re-entry control.