This document enumerates problems 512 does not attempt to solve.
It exists to prevent scope creep and misattribution of responsibility.
These are boundary conditions, not defects: they are the limits of what 512 can guarantee on its own.
512 does not determine:
- truth
- accuracy
- correctness
- intent
It evaluates whether a proposed action satisfies declared constraints. It does not determine whether those constraints are correct, complete, or whether the declared intent is truthful.
512 does not evaluate:
- morality
- fairness
- harm
- benefit
It contains no ethical logic.
512 does not:
- penalize actions that have already occurred
- reward compliant behaviour
- modify how agents behave over time
- enforce consequences beyond the boundary decision
At the commit boundary, 512 produces exactly two outputs: ALLOW or DENY. What happens to an agent as a result of a denial — sanctions, reputation effects, legal consequences — is outside 512's scope. 512 controls the boundary. It does not control what follows from a boundary decision.
When the gate cannot complete evaluation, it produces no output. Execution proceeds under the fail-open posture required by Invariant 6. The witness layer records the ungoverned period as an evidence chain gap. That gap record is a witness layer classification — it is not a gate output.
512 does not:
- verify identity
- manage credentials
- assign reputation
- enforce accountability
Any system doing so is external to 512.
512 does not:
- improve governance
- reform institutions
- guarantee compliance
- ensure trust
It constrains legitimacy witnessing only.
512 does not address:
- inequality
- democracy
- authoritarianism
- social cohesion
Such outcomes are outside its scope.
512 cannot prevent:
- misuse of systems
- malicious actors
- bad incentives
- institutional capture
It is not designed to.
If a problem requires:
- ethics
- enforcement of consequences
- identity
- values
- authority
It is out of scope for 512.
This document defines the explicit limits of the 512 constraint layer.