-
Notifications
You must be signed in to change notification settings - Fork 0
What PIC Does Not Do
PIC is a checking, routing, and residual-ledger toolkit. It is intentionally limited.
PIC does not prove that a system is a real artificial superintelligence.
It can produce protocol-relative reports about finite declared records, packet flow, verifier routes, and collective phase proxy conditions. Those reports are not proof of real ASI.
PIC does not prove unobserved physical facts, simulator outcomes, oracle answers, legal authority, policy outcomes, or real-world success.
When a claim depends on an external domain, PIC records the required verifier route or residual obligation.
In other words, PIC can help preserve and route evidence about an external-world claim, but it does not turn agent text into proof of external-world truth.
PIC output is diagnostic.
It does not authorize:
- shell commands;
- file writes;
- repository mutation;
- package installation;
- arbitrary network access;
- background crawling or autonomous polling;
- hidden live connectors without an explicit source;
- external message sending outside the declared local relay workflow;
- credential use;
- model-weight changes;
- self-rewrite;
- payments or purchases.
Host runtime controls still matter: sandboxing, approvals, allowlists, permissions, and tool policy.
PIC can perform bounded candidate intake for an explicit source and can write local agent-message inbox records through declared CLI commands. That is not the same as granting broad network, repository, shell, credential, or external messaging authority.
Agent text is candidate material. It can describe evidence, but it is not evidence by itself.
PIC separates:
- what an agent says;
- what metadata declares;
- what a finite checker accepted;
- what remains unresolved.
PIC is designed to preserve missing obligations and residual ledgers. If work remains unresolved, the report should show it.
- Home
- Getting Started
- Acceleration Measurement
- Operation Guide
- AFST Satisfaction Flux
- Python TypeScript Compatibility
- Security
- Quickstart
- ASI Proxy Loop v0.9
- Token Reports and Agent Loop
- ASI Proxy Acceleration
- CCR Interop
- MCP And A2A Safety
- Operation Gate
- Phase Ecology Lab
- Related OSS
- Use Cases
- Live Intake and Agent Messages
- Core Concepts
- Concepts in Plain Language
- Agent Guide
- Diagnostics and Agent Autonomy
- What PIC Does Not Do
- Safety and Limits
- Theory Map
- CLI Recipes
- Glossary
- FAQ
- v1.1 Release Notes
- v0.9.0 Release Notes
- v0.8.0 Release Notes
- v0.7.0 Release Notes
- v0.6.0 Release Notes