Skip to main content

Credibility

What proof shows that past agency earned the right to carry more scope?

Credibility is the proof layer of agency. It does not equal agency. It shows whether an actor kept commitments under real constraints, with consequences visible enough for others to trust the next commitment.

Definition​

Agency is the capacity to choose and act toward valued change, under real constraints, with feedback from consequences. Credibility is the track record that makes that capacity believable to others.

The Book records the collisions. Credibility is what survives when the record becomes legible.

Reputation is what people say. Authority is what a title says. Credibility asks: did the actor do what they said they would do?

Metric​

credibility = kept commitments / made commitments

The ratio is the base score. Qualify it before trusting it:

  • Sample size: a perfect 10/10 and a hard-earned 95/100 tell different stories.
  • Difficulty: easy promises should not score like hard promises.
  • External validation: someone outside the actor must see the same outcome.
  • Prediction calibration: confidence must match reality over time.

A person who keeps 95 of 100 commitments has a 0.95. A person who keeps 10 of 10 has a 1.0. The second ratio is cleaner. The first sample is richer.

How to Verify​

Check five proof surfaces:

  • Identity: is the actor who they claim to be?
  • Capability: can the actor do what they claim?
  • Integrity: did action match commitment over time?
  • Prediction: did confidence match the resolved outcome?
  • Market: did others validate with money, attention, referral, or repeat engagement?

The Tight Five route is Purpose -> Principles -> Platform -> Perspective -> Performance. Credibility lands in Performance because it asks what reality confirmed, not what the actor intended. Use Reality when the proof needs a measurable performance surface.

Agency Proof​

Use credibility to qualify the five agency dimensions:

  • Options: did the actor preserve or create real paths?
  • Intention: did the actor name the outcome in advance?
  • Capability: did the actor have the means to deliver?
  • Action: did the actor move when committed?
  • Proof: did consequences confirm the claim?

Human agents and AI agents both need the same proof pattern: completion, acceptance criteria, quality gates, and a better process for the next run. The ledger does not care what kind of agent you are. It cares whether you kept your word.

Failure Modes​

  • Thin perfection: 10/10 commitments, all easy, with no outside validation.
  • Self-scored proof: the actor declares success without a receiver or receipt.
  • Prediction theater: claims use confidence language but never resolve.
  • Borrowed authority: title, credential, or reputation substitutes for delivery.
  • Graph fragility: one failed referral weakens every trust edge that vouched for it.

Prediction Evidence​

Plans are predictions with deadlines. Prediction credibility measures whether confidence matched reality.

A plan worth believing in has three properties:

  1. Specific predictions with resolution dates.
  2. Conviction tags so others can calibrate confidence.
  3. Kill signals named before the actor starts.

The prediction score is a subscore, not the whole credibility model:

prediction credibility = (correct predictions / total predictions) × average conviction weight

Context​

  • Agency — The full operating model credibility helps prove
  • The Book — The ledger credibility makes visible
  • Alignment — The audit: do actions match words?
  • Predictions — How forecasts become evidence
  • Data Trust — How systems make credibility measurable
  • Trust — The social currency credibility earns

Questions​

If credibility is commitments kept divided by commitments made, what happens when you stop making commitments?

  • Is a person with 10/10 commitments kept more credible than one with 95/100?
  • When an AI agent earns L4 credibility through receipts, is that the same kind of trust you'd give a human with the same track record?
  • The Book says loyalty is "commitment held through difficulty" — is that the same as credibility, or does credibility require something loyalty doesn't?
  • What difficulty level and sample size make credibility compound rather than accumulate?