The Value Tensor
Reputation tells you whether to trust a person. It doesn't tell you whether a particular thing they made is any good. The Value Tensor - V3 - scores the work.
Concretely: if Alice writes a tutorial, V3 grades the tutorial (useful? true? sound?); T3 grades Alice (do her talent, training, and temperament fit the task). The same person can produce brilliant work one week and sloppy work the next, and V3 is what notices.
Both feed into ATP rewards. A trusted person (high T3) who produces poor work (low V3) still earns less.
Three Questions About Any Piece of Work
Every output in Web4 gets scored across three dimensions - and the weights are designed to reward truth over popularity.
Don't worry about memorizing six terms at once. The key insight is simple: T3 = your reputation as a person, V3 = the quality of a specific piece of work. The three V3 dimensions below just break down “quality” into usefulness, truthfulness, and soundness.
Valuation
weight: 0.30= “Was it useful?”
Measured by recipient satisfaction and ATP earned vs expected.
Veracity
weight: 0.35= “Was it true?”
Verified by external validation and witness attestation - a recorded confirmation from an independent party that it actually happened.
Validity
weight: 0.35= “Was it sound?”
Confirmed by receipt, logical consistency, and actual delivery.
Notice: Veracity + Validity (0.70 combined) outweigh Valuation (0.30). Web4 rewards truth and rigor over popularity by design.
And yes - like T3, V3 scores decay rather than sitting frozen: Valuation fades fastest (14-day half-life - how useful something is depends on conditions that change), Validity in 90 days (certifications and checks go stale), Veracity slowest (365 days - a truth record persists).
Who Assigns These Scores?
Not you - you never rate your own work. And not any one other party either. Each dimension is scored by a different mechanism, which is what stops any single scorer from simply lying.
Valuation is scored by the recipient of an output. When someone receives your work, they confirm whether it was useful via the VCM (Value Confirmation Message). This is like leaving a receipt - “yes, this helped me.”
Veracity is scored through witness attestation and external validation. Your LCT device witnesses verify that you actually did what you claim. For factual claims, other trusted entities in the same domain can challenge or corroborate. Think peer review: your claim is only as strong as the evidence your witnesses can attest to.
Validity is scored by structural verification - did the output actually arrive? Is it logically consistent? Receipt confirmation proves delivery, and the system checks internal consistency (e.g., a code review that contradicts itself scores low). This is the most automated dimension - much of it can be verified without human judgment.
In short: Valuation = the recipient judges usefulness. Veracity = witnesses and peers judge truthfulness. Validity = the system verifies delivery and consistency. No single party controls all three dimensions.
Try It: What Does Web4 Actually Reward?
Five pieces of work, scored by the same weights. Watch what happens to the one that gets 50,000 views.
Score These Outputs
Click each output to see how V3 scores it.
How T3 and V3 Work Together
T3 measures who you are (your trustworthiness as a person). V3 measures what you produce (the quality of each specific output). They're separate scores that feed into a single outcome: how much ATP you earn.
The feedback loop: High T3 (trusted person) + High V3 (great work) = maximum ATP reward. But a trusted person who produces sloppy work (high T3, low V3) earns less than their reputation suggests. And a newcomer who produces brilliant work (low T3, high V3) earns more than their reputation would predict.
Over time, T3 and V3 converge: consistently producing high-V3 work raises your T3. Consistently producing low-V3 work drags your T3 down. Your reputation tracks your actual output quality.