The Manifesto
What is Vetted?
Expert Judgment as Infrastructure.
Vetted is credibility infrastructure for hiring. Domain experts organize into Guilds, review candidates against peer-defined standards, and stake tokens on their assessments. Accurate judgments earn rewards. Poor ones get slashed. The result is a signal that does not exist anywhere today: expert evaluation with real economic consequences attached to it.
How Raw Judgment Becomes a Trustworthy Signal.
Most systems aggregate opinion. Vetted does something structurally different: it transforms professional intuition and the judgment that comes from having done the work into an economically accountable signal, verified against real-world outcomes.
When an expert reviews a candidate, they stake tokens before seeing the application. Their score is submitted privately. They cannot see what other reviewers are scoring. Once all scores are in, the protocol computes the median, strips the extremes, and averages only the scores that fall within the consensus band. Reviewers whose scores align earn rewards. Those outside it face slashing.
This does two things simultaneously. It makes honest evaluation the only rational strategy, because gaming the system by scoring too high or too low puts your stake at risk. And it produces a candidate score that reflects genuine collective judgment, not the loudest voice or the most generous reviewer.
The signal turns opinions into consensus with economic consequences for inaccuracy.
But consensus alone is not enough. Consensus can be wrong.
Protecting the Outlier.
The deepest flaw in any consensus system is majority bias: genuinely exceptional candidates with unconventional backgrounds get filtered out because their profile is unfamiliar to reviewers trained on conventional paths.
Vetted addresses this structurally. Any expert on the platform, or the candidate themselves, can trigger an appeal. The original scores, the reasoning behind them, and the challenger's argument are reviewed side by side. If the appeal succeeds, the candidate is admitted and the appealing expert earns reputation plus a portion of the first-year salary whenever that candidate gets hired. If it fails, the challenger is slashed.
This creates a counter-incentive to majority bias. Experts who consistently identify talent that consensus missed build outsized reputation and earn real upside from successful appeals. The system tolerates and pays for accurate dissent.
Personal Conviction
The endorsement layer adds another dimension. When Guild-approved candidates apply for specific roles, experts bid in a sealed auction to endorse them. Top three stakes win. Sealed bids mean no one can see what others are staking. This surfaces genuine conviction. An expert who endorses a candidate is betting capital on a specific prediction: that this person will succeed in this specific role, with this team, in this context. That is a materially harder prediction than "this person is generally qualified," and it is treated as such. Endorsers who get it right earn a share of the placement fee. Endorsers whose candidates fail due to skill mismatch face proportional slashing.
The Feedback Loop That Makes It Deterministic
Here is what separates Vetted from every other trust signal in hiring: the outcome feeds back in.
After each hire, the platform tracks retention and performance. That data flows back to the experts who endorsed the candidate and to the review panel that admitted them. Experts who consistently identify talent that succeeds build compounding reputation and earn higher shares of future reward pools. Experts whose judgments prove poor face slashing and reputation decay. Guild standards are refined based on what the outcome data actually shows predicts success.
Every review generates data. Every data point updates the model. The reward structure ensures that the experts with the best track records gain the most influence on future evaluations. The signal improves with every completed cycle.
Credibility shifts from assumption to a measurable, compounding output.
Why Blockchain
If you arrive at solving trust from any direction, you eventually reach the same conclusion: accountable systems need infrastructure that no single party can manipulate.
All reviews, stakes, slashing events, and outcomes are recorded on-chain. Auditable by anyone. Manipulable by no one. Expert reputations are portable across protocols and contexts. They cannot be erased by a platform closing or changing its rules. Staking and slashing create consequences for poor judgment that are architecturally impossible to enforce in any centralized system.
Centralized platforms can approximate parts of this. They cannot deliver all of it simultaneously: transparent process, private identities, economic stakes, portable reputation, and outcome-linked feedback in a single system where no single party controls the outcome.
That combination requires trustless infrastructure. We chose blockchain because the alternative was asking you to trust us. And the whole point of Vetted is that you shouldn't have to trust anyone.
The Underlying Primitive
Vetted starts with hiring because it is concrete, measurable, and universally flawed. But the underlying primitive, structured expert consensus with economic accountability, majority-bias protection, and outcome verification, extends to any domain where expertise is needed and trust is absent.
The core belief is simple: the people best qualified to evaluate talent are the experts in that domain. Vetted is the first system that makes it economically rational for them to do so, economically costly for them to be wrong, and structurally impossible for the majority to quietly bury the ones they got wrong.
Three parties. One infrastructure.
Experts monetize their judgment, set the standards in their field, and earn from the value created by the candidates they endorse. Companies get signal they can stake decisions on, strengthening their judgment under uncertainty. Candidates get evaluated by people who actually understand their work, regardless of background.
Each one's need is the other's solution.
Everything can be generated. Except trust.