Can a researcher learn more by choosing the right experiment? GENESIS makes that question visible, testable and inspectable.
One suggested route · No wallet, new payment or 3D required.
Builder wallet → IMD Intake → agent proposal → Genesis measurement
The preserved research request used direct Intake. GenesisResearch did not fund or execute this historical request.
Owner → GenesisResearch → real IMD Intake → delivered callback → next Genesis measurement
A real follow-up received the first IMD proposal and its measured outcome. The accepted response and delivered callback were verified before one new experiment was applied. Compatible models: 14 → 10. No repeat purchase.
Real research order ↗ · Details and limits ↗ · Inspect the free local EVM loop ↗
A learner can see how evidence changes an explanation and why the next experiment is worth spending a limited experiment budget on.
GenesisResearch paid for one follow-up. A delivered mainnet callback and accepted artifact were verified before the next simulator measurement. 14 → 10 compatible local models.
Explore the real feedback loop → · Choose the next test yourself → · Inspect transactions, callback and costs →
SHA-256 1e985c840f4f1242208ff44c0ea374725af8660d9fa770607b75acb98ab75479
Callback confirmed on Ethereum
Fresh mainnet reads match the successful transaction, ResearchResult event and current delivered contract record. This confirms delivery, not scientific truth.
Checked 2026-10-24T20:51:53.102Z · UTCIMD API reachable
The official endpoint returned this recorded job (state: completed). Availability and job identity only; this is not the full artifact verifier.
Checked 2026-10-24T20:51:53.102Z · UTCPublic research engine 1.1.0 → · README and reproduction instructions →
61 public-source tests passed. The updated feedback verifier supports a chosen archive-capable Ethereum RPC and distinguishes unavailable online reads from evidence mismatches. The earlier review-v2 ZIP remains an unchanged historical package.
Water may improve plant growth, but the response could depend on mineral composition and light. Choose a controlled experiment that distinguishes these explanations.
Public observation 1 reports growth 50, survival false, and energy 40 without water, with no mineral, under amber light at warm temperature. With only this observation, evidence is insufficient to establish a water benefit or dependence on minerals or light. A water-only contrast can test its effect in this setting, but cannot establish whether that effect changes across mineral or light conditions.
Adding water left growth at 50/100, energy at 40/100 and survival false in this comparison. Mineral and light interactions remain unresolved. One comparison does not establish a universal law.
These are local calculations from the public hypothesis family, not additional IMD responses.
Compared with observation 1, this proposed experiment changes only water. It tests whether adding water changes growth from 50 at fixed mineral, light, and temperature, distinguishing a water benefit in this setting from no benefit. Mineral and light interactions would remain unresolved and require later controlled comparisons. This experiment has not been executed.
a832e2611353d3cb8e8c29d5c6250fe591b58be650413f03a48eb6fe7c746fc9281d59f730460ddf86d075da1e6d39082bb58dad492a2630b9225f31a154d54fPredict the response to a water-only experiment, inspect the original IMD proposal, then compare the measured result and decide what to test next. The useful skill is separating an assumption from an observation, even when the plant does not survive.
Follow the learning scenario → · Try the second free challenge →
Reproduce separate local worlds, recalculate exclusions from the original measurements, and check external request provenance. Each supports a different claim.
Read the evidence levels and their limits → · Download the immutable recorded-world.json →
Archive SHA-256: 281d59f730460ddf86d075da1e6d39082bb58dad492a2630b9225f31a154d54f
The recorded IMD agent and Kepler both chose to add water while holding mineral, light and temperature fixed. Ada and Lin would change mineral and temperature. All four choices have the same expected model separation under the local reference family.
| Researcher | Proposed experiment | Expected models left¹ | Recorded result |
|---|---|---|---|
| IMDHistorical proposal | Water · no mineral · amber light · warmChanges: water | 11 | Growth 50/10019 → 14 models · 5 excluded |
| Ada | No water · basalt · amber light · coolChanges: mineral + temperature | 11 | Not run separately |
| KeplerNext local turn | Water · no mineral · amber light · warmChanges: water | 11 | Not run separatelySame conditions as recorded IMD treatment |
| Lin | No water · basalt · amber light · coolChanges: mineral + temperature | 11 | Not run separately |
Local strategies receive the public observations and know the full 24-model family and response function.
No external payment. A local training experiment costs 2 internal game credits.
The original IMD request supplied the hypothesis, one observation and available condition combinations. It did not include the model family or response formula.
Historical external cost: 0.5 IMD + 0.0000157250756555 ETH aggregate gas, plus 2 internal credits for the simulation.
Useful for the question: the water-only contrast tests whether water improves growth in these fixed conditions. The recorded response did not improve, ruling out five explanations. It leaves water/mineral/light interactions unresolved.
This is one retrospective choice comparison, not evidence that IMD is better or more cost-efficient. Local rows are proposals recalculated from the original first observation; no new experiment or payment was made.
Start with the current verifier. Historical packages are preserved for their original scope; their version numbers do not indicate newer research.
Second real IMD experience: contract order, delivered callback and 14 → 10 models.
Download feedback-v2 ZIP ↓ · Instructions and scope →
ZIP SHA-2565da34cc2d0c10ed61376074246138240a2668de0550a898f290720da4b691981
First direct-Intake experience, 19 → 14 models and an earlier spent = 0 contract snapshot. Does not cover the later contract-funded request.
Download review-v2 ZIP ↓ · Instructions and scope →
ZIP SHA-25685831a7b678341da4b2d4df7b5de42fbe91c4b76239b2a3d6eab2b862d5dc382
Earlier verifier for the same second experience. Use feedback-v2 for configurable RPC and clearer diagnostics.
Download feedback-v1 ZIP ↓ · Instructions and scope →
ZIP SHA-256a42516f99dff77729f9aa9fc3d77344dcb991cb3f9906e80157af206495475cb
Version numbers belong to different package families. feedback-v2 and review-v2 are different records. Existing ZIPs and measurement archives are unchanged.
Extract feedback-v2 and run node verify.mjs with Node.js 20+. No dependency installation, wallet or network. For fresh read-only checks, use node verify.mjs --online --rpc HTTPS_URL.
Package guide and historical details → · Historical review-v2 pinned source →
Synthetic deterministic simulator; not a discovery of real-world scientific laws.
Ada, Kepler and Lin are local algorithms with the 24-model reference family as prior knowledge. They are not IMD agents.
Four held-out condition combinations in the same world. Predictions come from the remaining local reference model. A 4/4 score is not an independent evaluation of IMD intelligence.
Two real IMD research proposals are preserved, each matched to one simulator experiment. The second received the first result and used GenesisResearch; viewing this record makes no payment.
The transaction and artifact checks establish request provenance, not scientific truth. The simulator commitment is local; no on-chain experiment attestation is claimed.
The first real experience used direct Intake. A second real IMD request received its feedback, was funded through GenesisResearch and delivered an authenticated callback before one simulator experiment. The two routes are preserved separately.
Latest saved public check: 2026-10-22T19:29:05.932Z; spent() = 500000000000000000. One real contract-funded mainnet request and delivered callback are demonstrated in the separate continuation journal. Contract role and limitations ↗
0x550011CAdCC758969391050241B2Fc43ED1a4B970xff03410d0fe5fa8f7f59f743de35e333d98571200xcecc29b037f5064fcdf45a5c318f132ef76aa5510x550011CAdCC758969391050241B2Fc43ED1a4B97
The requested recipient paid for the IMD launch, is the recorded launch requester, and owns GenesisResearch. IMD used its service wallet and factory for deployment.
This website cannot configure the prize payment. There is no confirmed organizer policy mapping factory deployments to a recipient. The requested address is not a payout guarantee.
The verified AssetReceiver source withdraws to a fixed WITHDRAWER address. It has no winner registry or recipient argument for withdrawal. Distribution to winners requires a separate organizer action.