Every claim, with its receipt.

Claim

Weights that the Blindproof validator writes from verified work turn into real subnet incentive on Bittensor’s public testnet: honest miners are paid in proportion to the work the validator credited, and a miner that returned cheap guesses is paid nothing.

Test

Phase T’s first graded cycle wrote weights before subnet 575 had an emission schedule or a staked validator, so no incentive could follow. Both were fixed the same evening: the emission schedule was started and the validator was staked. The same 24-job cycle (12 deterministic-tier and 12 proven-tier jobs, same three miners, same mock inference) was then run a second time, and the validator wrote its weights at 19:54:50 EDT on 5 October 2026: FINAL SetWeights {2: 0.555, 1: 0.327} -> OK. Subnet 575 runs commit-reveal weights, so the chain applies a weight vector two epochs after it is committed. On-chain incentive was then read from the public testnet node at every epoch boundary after the commit.

Result Measured

Incentive first appeared at block 8,159,238, two epochs (720 blocks) after the weight commit at block 8,158,518, and was paid at every epoch through block 8,163,558 — 13 consecutive epochs, about 14 hours. The split matches the validator’s weight vector to four decimal places. Incentive returned to zero by block 8,163,918, when the testnet’s activity cutoff retired the single weight commit; the validator had set weights once and exited, as designed for a graded cycle.

Chain references

Addresses for every key are on the testnet record.

How to check

The public testnet node keeps full state history, so the incentive vector can be read back at any of the blocks above with the Bittensor Python SDK (v11). Values are stored as 16-bit integers; divide by 65535.

import bittensor as bt
st = bt.Subtensor(network="test")
for blk in (8158518, 8159238, 8163558, 8163918):
    inc = st.query(("SubtensorModule", "Incentive"), [575], block=blk)
    print(blk, [round(v / 65535, 4) for v in inc.value])
# 8158518 [0.0, 0.0, 0.0, 0.0]
# 8159238 [0.0, 0.3709, 0.6291, 0.0]
# 8163558 [0.0, 0.3709, 0.6291, 0.0]
# 8163918 [0.0, 0.0, 0.0, 0.0]

Element order is UID order: validator, m01, m02, m03.

What Phase T+ does not prove

• Real inference or real proofs: the credited work was mock inference with mock proofs, as in Phase T.
• Sustained payment: one weight commit, paid for 13 epochs until the testnet’s activity cutoff retired it. A validator that keeps weights current is the next build item and is required before any rung can claim continuous operation.
• Decentralization: one team operates the validator and all miners.
• Value: testnet incentive is test TAO, which has no value.

Claim

On Bittensor’s public testnet, the Blindproof validator credits miners only when their work verifies, in both the deterministic tier (two-miner hash agreement) and the proven tier (single miner plus proof check), and writes weights that reflect only verified work.

Test

One validator (UID 0) on a separate machine, making outbound calls only. Three miners (UIDs 1–3) on a single rented CPU server with public endpoints, calling a mock inference server paced at a measured 32.8 tokens/s. Two miners behave honestly. The third is lazy: it returns a cheap guess instead of running the model. The validator dispatches 12 deterministic-tier and 12 proven-tier jobs, settles each one, and then writes weights.

Result Measured

Validator weights were written at 19:33:50 EDT on 5 October 2026: FINAL SetWeights {1: 0.337, 2: 0.547} -> OK. Miner incentive in that epoch was zero because the subnet’s emission schedule had not yet been started and the validator held no stake. Both were fixed immediately after, the cycle was run again, and the chain paid the honest miners — see Phase T+.

Chain references

Addresses for every key are on the testnet record.

How to check

Open each block in the explorer and confirm the extrinsic and its signer. List current participants with btcli wallet overview or the SDK snippet under Phase T+. Run logs and per-job ledgers are archived and will be published with the reference validator.

What Phase T does not prove

• Real inference: miners used a mock server, not the model.
• Real proofs: proven-tier proofs were mock objects.
• Privacy: no masked execution on testnet.
• Decentralization: one team operates the validator and all miners.
• Sustained operation: one graded cycle, no fees, testnet only.

Claim

Before spending anything on a public network, the validator rejects forged or misdirected traffic, never pays a cheating miner in either tier, and keeps settling correctly under load.

Test

A full local Bittensor chain with the subnet, a validator, and miner personas scripted to cheat in specific ways. Sixteen scenarios, each run repeatedly, against a mock inference server paced at a measured 32.8 tokens/s.

Result Measured — 16 / 16 scenarios passed

What Phase M does not prove

• Anything about a public network: this ran on a private local chain.
• Real inference or real proofs: both were mocked, as in Phase T.
• Resistance to colluding miners: the deterministic tier assumes the two matched miners are not colluding.