https://meta.aiyosg.com/MetaAware Progress2026-09-13T00:00:00+08:00https://meta.aiyosg.com/updates/how-metaaware-worksHow MetaAware works—from evidence to accountable action2026-09-13T00:00:00+08:00The Approach page now explains the seven-stage governance loop, evidence inputs, finding contract, epistemic states, natural-language boundary, and governed capability improvement.https://meta.aiyosg.com/updates/security-governance-horizonSecurity governance accepted as a MetaAware product horizon2026-09-13T00:00:00+08:00Secure delivery, runtime and network posture, and identity and payment integrity will be explored in stages after the Git evidence-engine MVP.https://meta.aiyosg.com/updates/expanded-git-evidenceExpanded research note: Public Git history is evidence, not ground truth2026-09-13T00:00:00+08:00The note now develops the observable limits, selection effects, and responsible curation of public development histories.https://meta.aiyosg.com/updates/expanded-self-certificationExpanded research note: AI should not certify its own work2026-09-13T00:00:00+08:00The note now develops why model agreement is not independent assurance and where an assurance chain should terminate.https://meta.aiyosg.com/updates/expanded-before-expertiseExpanded research note: Why AI-assisted development needs governance before expertise2026-09-13T00:00:00+08:00The note now develops how governance can reduce immediate risk while helping expertise grow.https://meta.aiyosg.com/updates/open-source-founder-habitsHow can open-source governance rules avoid encoding the founder's habits?2026-09-13T00:00:00+08:00Open source does not automatically create plural governance; rules still need provenance, counterexamples, and transparent decisions.https://meta.aiyosg.com/updates/quantify-governance-capabilityCan governance capability be quantified?2026-09-13T00:00:00+08:00Governance capability should be an evidence-bound multidimensional profile, not a single authority score.https://meta.aiyosg.com/updates/agentic-approval-accountabilityApproval and accountability boundaries for agentic AI2026-09-13T00:00:00+08:00Approval should bind specific effects, scope, time, and evidence; an agent must not expand its own authority.https://meta.aiyosg.com/updates/unknown-legitimate-resultWhy unknown is a legitimate result2026-09-13T00:00:00+08:00When evidence is insufficient, a structured unknown is more governable than fabricated certainty.https://meta.aiyosg.com/updates/measure-false-positives-negativesHow should a governance tool measure false positives and false negatives?2026-09-13T00:00:00+08:00Governance evaluation needs segmented error profiles, consequence measures, and reviewable reference judgments.https://meta.aiyosg.com/updates/natural-language-deterministic-engineNatural-language governance interfaces and deterministic engines can coexist2026-09-13T00:00:00+08:00Natural language can interpret the question while repeatable judgments come from an inspectable deterministic core.https://meta.aiyosg.com/updates/can-ai-governCan AI govern AI-assisted development?2026-09-13T00:00:00+08:00AI can participate in governance, but its authority must be bounded, its claims evidence-linked, and final accountability human.https://meta.aiyosg.com/updates/understanding-not-complianceGovernance should increase understanding, not only compliance2026-09-13T00:00:00+08:00Governance tools should be designed and evaluated as learning systems as well as control systems.https://meta.aiyosg.com/updates/public-previewMetaAware public research preview published2026-09-13T00:00:00+08:00The bilingual static website is now public. The governance CLI remains unimplemented.https://meta.aiyosg.com/updates/website-designPublic research website design accepted2026-09-13T00:00:00+08:00The site's information architecture, evidence boundaries, and publication gates were accepted.