# Score of the IMD–SIMD thesis

Assignment: `[SIMD-THESIS]:muv5dzga-vxnpa`  
Research date: 2026-10-05. Only the supplied text is scored; it ends mid-sentence at “swarm usage dat”.

## Verdict and scoring

**58/100: a plausible subsidy mechanism, with partial factual support, presented as a proven economic flywheel.** The strongest part is the distinction between an NFT-gated workforce and an external sponsor. The weakest parts are the assertions of stable demand, automatic efficient allocation, direct proportionality between trading and useful labor, and reciprocal token-price benefits.

This is an editorial evidence score, not a protocol audit, investment rating, or probability of success. No scoring scale was supplied, so the following rubric awards points for support in the material reviewed, rather than presumed truth. Missing evidence is not proof of falsity.

| Dimension | Maximum | Awarded | Reason |
|---|---:|---:|---|
| Accuracy of protocol descriptions | 30 | 22 | Worker and staking architecture supported; live rewards, governance and seat count need qualification. |
| Evidence for the IMD–SIMD link | 25 | 14 | Attributable subsidy announcements exist, but the complete transaction and implementation trail was not verified. |
| Economic and causal reasoning | 25 | 15 | Lowering the user price can stimulate use; sustainability and token appreciation do not follow automatically. |
| Treatment of uncertainty and alternatives | 20 | 7 | Some future-facing language appears, but major causal claims lack assumptions, measurements or countercases. |
| **Total** | **100** | **58** | **Conceptually useful; empirically incomplete.** |

## Evidence and claim assessment

“Documented” below means supported by a project-controlled source, not independently audited. “Reported” means attributable public statements with weaker retrieval or verification. “Inference” means a conclusion drawn from a mechanism. “Unverified” means this review could not establish the claim.

| Thesis claim | Assessment | Evidence and limitation |
|---|---|---|
| NFT holders supply compute and AI subscriptions to a swarm. | **Documented.** | The official [worker distribution](https://github.com/Identity-md/worker#readme) describes contributing an authenticated runtime, consuming the operator’s quota, registering an eligible NFT as an ERC-8004 agent, and authorizing one active device per NFT. It also describes a control plane; distributed labor alone does not establish decentralized governance. |
| Up to 2,000 NFT seats exist. | **Partially supported.** | [Bankless’s September overview](https://www.bankless.com/read/inside-imd-ethereum-s-new-ai-swarm-experiment) reports 2,000 NFTs. This is secondary corroboration: this review did not read the NFT contract’s supply limit. Collection size is not a count of online workers or usable concurrent capacity. |
| Agents perform code, research, audits, oracles and launch support. | **Observed in part.** | The official [job explorer](https://explorer.imd.fun/) publishes task records. An [audit record](https://explorer.imd.fun/jobs/e6eda4d8-f50d-47cd-9464-9a272283ccd3) demonstrates that audit work is offered and outputs are recorded. A task record establishes activity, not that every result is correct, secure, or commercially useful. |
| IMD pays for jobs and agents receive IMD rewards. | **Payment supported; payout generalization unverified.** | [Bankless](https://www.bankless.com/read/inside-imd-ethereum-s-new-ai-swarm-experiment) reports 0.5 IMD per request through x402. The worker documentation says connected workers earn, but does not establish a universal per-job IMD payout schedule. Request charges, operator compensation and token allocations must be distinguished. |
| POOL4 burns supply and supports sIMD and NFT distributions. | **Documented, with an important correction.** | [POOL4 documentation](https://pool4.imd.fun/docs) describes excess inventory retirement, an 85% burn allocation, staking shares, and reserved allocations. Its bonding and NFT-node programs are described as still being built. Thus a reserve is not proof of a live distribution to seat holders. Burning is conditional on the mechanism operating, not every sale everywhere. |
| SIMD observes agents and funds 100% of job costs from fees. | **Reported policy, not independently verified implementation.** | Statements attributed to [@SuperIMD_eth](https://x.com/SuperIMD_eth), available through an [indexed Sotwe mirror](https://www.sotwe.com/SuperIMD_eth), describe monitoring, fee-funded subsidies, and a transition from half-cost refunds to full coverage. Direct X retrieval returned 403; the mirror could not be reopened. No source-code or end-to-end settlement audit was completed. |
| Users can create jobs without connecting a wallet. | **Reported for the SIMD flow.** | The same mirrored account announces walletless submission. This does not establish walletless access for every native IMD job or an unlimited entitlement. Earlier refund language describes a different payment experience. |
| SIMD fees are periodically converted into IMD and a vault exclusively pays jobs. | **Partly reported, exclusivity unverified.** | The mirrored announcements describe five-minute fee claims and fee-derived IMD funding. No verified conversion route, exclusive spending restriction, keeper history, or custody policy was obtained. The announced vault address is a lead for investigation, not proof of these guarantees. |
| Observation optimizes funding and reduces waste. | **Hypothesis.** | No allocation algorithm, efficiency criterion, rejection policy or comparative results were found. A dashboard can inform decisions; it does not demonstrate that decisions use those observations. |
| More activity increases useful labor, rewards and both tokens’ real demand. | **Conditional inference.** | Funding can enable additional requests, but completion, usefulness, compensation, retained users and token demand are separate outcomes. No causal dataset establishes this chain. |

## What the subsidy evidence actually establishes

Mirrored policy updates and funding counters are operator representations, not reconciled measurements. Relative post ages cannot establish precise dates. An earlier half-coverage launch statement can coexist with a subsequent full-coverage policy. [Indexed account statements](https://www.sotwe.com/SuperIMD_eth).

The proposed identity is Ethereum token `0xbb0c1f82a2ea0253ea3d91c2f05caded82133415`; the reported vault is `0xd60483eb8004e3de3e283b3eff0e67fbb57f9b21`. These identifiers come from those statements and were not authenticated through contract analysis. An attempted [vault explorer lookup](https://etherscan.io/address/0xd60483eb8004e3de3e283b3eff0e67fbb57f9b21) failed to return usable content.

An [IMD animation job](https://explorer.imd.fun/jobs/d2384040-0e41-499a-b283-efd25cbf5054) repeats the fee-funded-vault narrative in its request. That is evidence that someone submitted this narrative to the swarm, not independent corroboration of its financial mechanism. Its structural verification concerns the delivered artifact, not the truth of the request.

## Economic assessment: a conditional loop

**Inference:** covering a request charge lowers the user’s financial barrier and can increase experimentation. It can also fund requests users would never pay for, duplicate work or spam. Subsidized volume is therefore distinct from organic willingness to pay and useful output.

An illustrative accounting model, not measured protocol behavior, is:

`available IMD = opening balance + fee-derived IMD + other inflows − other outflows`

`fundable jobs ≤ available IMD / covered IMD charge per job`

Successful useful work additionally depends on submission demand, capacity, acceptance and quality. Token trading volume is only one input to fee revenue; fee rates, the subsidy share, denomination, conversion prices and execution costs matter. Liquidity depth affects execution but is not itself fee income. A finite vault can cover 100% of a quoted charge for eligible jobs while still exhausting its runway. Covering that charge also does not prove reimbursement of all operator compute and subscription costs.

**Unproven causal step:** higher IMD price does not automatically increase the value of SIMD fees. That depends on fee denomination and trading behavior. If subsidies are bought using a fixed dollar budget, a higher IMD price can purchase fewer IMD-denominated jobs. Job payments may transfer tokens to recipients who sell them; burns and job settlement are separate mechanisms. Neither more requests nor lower supply guarantees appreciation or deeper liquidity.

**Unproven sustainability claim:** fee funding dependent on speculative trading is variable. Calling it “stable demand from the outside” requires evidence across quiet markets as well as launch activity. The reverse cycle is possible: lower volume, lower subsidy income, depleted reserves, fewer funded jobs. There is no demonstrated unavoidable failure, but neither is there a demonstrated self-sustaining loop.

## Unanswered questions and tests that would change the score

1. **Can every subsidy be reconciled?** Publish fee-claim transactions, conversions, vault inflows/outflows and matching job IDs. Distinguish upfront payment from refunds, failed jobs and repeat subsidies.
2. **What is guaranteed?** Publish eligibility, caps, exclusions, queue rules, fee allocation, vault control and behavior at exhaustion. Identify which rules are enforced in code and which are operator policy.
3. **Who earns what?** Document how request payments reach operators, and distinguish those payments from POOL4 reserves, staking rewards and launch-token allocations.
4. **Is the demand incremental and useful?** Measure accepted work, retries, actual compute costs, independent quality reviews, returning users and willingness to pay after subsidies. Compare against an unsubsidized baseline.
5. **Does observation guide allocation?** Provide a decision policy and evidence that it improves useful output per unit spent.
6. **How long can it last?** Report net fee income and subsidy expenditure over time, reserve runway and low-volume stress scenarios. A launch-day balance is insufficient.
7. **How decentralized is the system?** Identify scheduler, verifier, custody and upgrade authorities, and verify NFT supply and active capacity separately.

## Better-supported formulation

IMD provides an NFT-gated distributed worker network and documented token/staking infrastructure. SIMD publicly reports using fee-derived IMD to subsidize requests, including a move to full coverage and a walletless submission flow. This could reduce barriers and increase swarm usage while funds and capacity permit. Efficient allocation, lasting demand, operator profitability and reciprocal token-value gains remain hypotheses requiring implementation evidence and longitudinal measurement.

## Research limits

Public sources were checked on 2026-10-05. Primary IMD documentation supports technical descriptions; secondary reporting is explicitly labeled where needed. SIMD evidence was available chiefly as mirrored first-party statements, with weaker provenance than a directly retrieved post. Official landing pages exposed little text; no discoverable SIMD technical documentation was established in this review. An unrelated Solana SIMD search result was excluded.

No contracts were executed, balances reconciled, privileged configuration inspected, or causal study performed. Links are mutable and no block-pinned snapshot is claimed. The score should be revisited when stronger evidence becomes available. Local artifact checks establish file completeness only and have no independent authority over these findings.