Position paper · v1
Most DePIN networks already have their own genuity checks — and they should. A bureau is not a replacement for those checks. It's the layer above them. Here is the difference, in plain language, with the comparison most networks ask for after the first call.
"Our genuity check works over time — collected token, holder rank, click participation, DAO participation. We already have plenty of checks in place to ensure the authenticity of a node."
Correct — and not what a bureau does. A network's internal genuity check is your bank's internal risk model. The bureau is the credit report. They are not the same shape, they are not substitutes, and a serious financial system ships with both.
When a network grades its own users, the network has every incentive to be lenient — flagging users hurts the network's headline numbers. A bureau has the opposite incentive: its only product is being right. The same reason FICO, Moody's, and Carfax exist as independent companies is the reason a DePIN bureau has to be independent of every network it grades. We never issue a token. We never hold equity in a graded network. We never accept payment in a graded network's native asset. The neutrality is the product.
Your genuity check sees only your data. A bureau sees a node's behavior across every network it touches — DIMO trips, Hivemapper coverage, WeatherXM contributions, Helium uptime. A node that behaves honestly on three networks and starts misbehaving on the fourth is a signal you can't generate from inside your own walls. That cross-network reputation only exists if a neutral third party publishes it.
Most genuity checks are heuristics: time-weighted activity, engagement scores, manual review. They work — until an adversary learns the heuristic and games it. GETKINETIK's input is a hardware-signed proof from the device's secure enclave. You cannot fake a Genesis Score input without breaking Android's TEE. The methodology is published, the algorithm is open source, every grade is reproducible from public evidence. That's a different category of trust from "we've seen this user click a lot."
When DePIN networks scale into custodial or financial territory — token rewards as income, insurance products on top of contributions, institutional staking — regulators will ask who graded the underlying participant? "We graded ourselves" is a worse answer than "a third party with published methodology graded them." The bureau shape is the answer that pre-exists the question.
Every internal fraud heuristic decays. Sybil patterns evolve. Spoofing tools improve. If you maintain your own check, you maintain it forever. The bureau ships methodology versions on its own clock — you call one endpoint and inherit every improvement automatically. That's not a cost saving. That's a different ops shape.
| Your own genuity check | GETKINETIK Bureau | |
|---|---|---|
| Neutral third party | No — you grade your own users | Yes — outside every network |
| Sees node across networks | Your network only | Every network the node touches |
| Hardware-rooted evidence | Sometimes (behavioral usually) | Always — TEE-signed Ed25519 |
| Public methodology | Internal, opaque | Versioned, open |
| Independently reproducible grade | No | Yes — replay proof + methodology |
| Maintenance cost on your team | You maintain it forever | Zero — we ship versions |
| Regulatory third-party-grade narrative | "We graded ourselves" | "An independent bureau did" |
| Cost to your users | N/A | Free — bureau never charges users |
| Cost to your team | Engineering team forever | One HTTP call |
You don't have to replace your genuity check. The most valuable integrations stack them:
Pay verified-device users a 5–10% bonus. The bureau's grade decides the tier; your network decides the dollar amount. You reduce Sybil exposure without raising rewards for everyone. Users earn more by running on real hardware. Both sides win.
Before sending tokens, hit /api/score/:nodeId. If the
score band is TAMPERED, hold the payout for review.
Your own check still runs; the bureau is a second opinion that
catches things your heuristic doesn't.
After a payout, POST to /api/attest with what your
network observed. The bureau folds your signal into the next score
version (v1.2). Your contribution feeds back into a better grade for
every other network — a reputation flywheel no single network can
build alone.
Banks don't trust their own credit scores in isolation. They pull a third-party bureau report. DePIN networks will do the same — the only question is whether they integrate the bureau that already exists, or wait for one they can't influence.
Fine. We are not arguing every network needs one today. We are arguing that the networks that win the next five years will have a third-party grade attached to every participant — the same way every meaningful piece of physical infrastructure carries a UL rating, every used car carries a Carfax, and every loan applicant has a FICO. DePIN does not get to skip that step. The only choice is whether to integrate it early, before the methodology hardens around someone else's assumptions.
The methodology is at docs/methodology/GENESIS_SCORE.md. It's open. Read it. Disagree with it on the record. We'll publish your dissent in the changelog.