SUPERSTRATEGY – probabilistically optimised, frontier-forecasted decisions.

Question: Will the performance of leading GPU clusters be at least 10× that of early-2026 leading clusters by 1 Jan 29?

Resolves: TRUE if Epoch's GPU-clusters data shows leading-cluster performance ≥10× the early-2026 leader by the date; otherwise FALSE.

Question: Will an open-weight model be within 5% of the leading closed model on a major composite capability index as of 1 Jan 31?

Resolves: Will resolve as true if, by 01 Jan 31, Artificial Analysis, Stanford HELM or a comparable independent composite evaluation places an open-weight model within 5% of the leading closed model using the same published scoring method. It will otherwise resolve as false.

Why it resolved: Artificial Analysis’ August 6, 2026 v4.1.1 update says Claude Opus 5 remains #1 on the Artificial Analysis Intelligence Index with a score of 63 (https://artificialanalysis.ai/articles/artificial-analysis-intelligence-index-v4-1-1). Artificial Analysis’ Kimi K3 comparison page states Kimi K3 scores 60 on the same Intelligence Index v4.1.1, while GPT-5.6 Sol scores 61 (https://artificialanalysis.ai/models/comparisons/kimi-k3-vs-gpt-5-6-sol). Since 60/63 = 95.24%, Kimi K3 is within 5% of the leading closed model. Moonshot’s Hugging Face repository states that full Kimi K3 model weights are released under the Kimi K3 License (https://huggingface.co/moonshotai/Kimi-K3/commit/c5d1dd4c428bd1ce8b88c5044f3b6ccde9e3b721), satisfying the open-weight condition.

Question: Will AI-driven diagnostics, treatment or personalised care be in mainstream, system-wide use across the NHS before 1 Jan 31?

Resolves: TRUE if NHS/DHSC reporting documents system-wide deployment of AI diagnostics/treatment/personalised care before the date; otherwise FALSE.

Why it resolved: NHS England’s official “Quality strategy for NHS-funded care in England,” published 14 July 2026, reports that an AI tool developed by Skin Analytics “is being rolled out across the NHS” to assess potentially cancerous skin lesions and recommend discharge or referral. It also reports an AI tool predicting risks such as falls and viruses is being rolled out across more than two-thirds of ICBs. These are NHS-reported AI diagnostic/care deployments beyond pilots. Source: https://www.england.nhs.uk/long-read/quality-strategy-for-nhs-funded-care-in-england/

Question: Will at least 80% of UK businesses report using AI as of Q4 2030, with at least half of adopters delegating multi-step work without case-by-case approval?

Resolves: Will resolve as true if, by 01 Jul 31, official UK business data covering Q4 2030 show that at least 80% of businesses use AI, and an official or robust representative survey shows that at least half of adopters delegate multi-step work without case-by-case human approval. It will otherwise resolve as false.