AI futures evidence — X6: adjacent science enablers

Each section: Claim · Why · Evidence · Analogue · Would update if · Conf (H/M/L).

Parent index: crosscut x6 adjacent science enablers
Spine: S_dual · 08 parallel spines
Linked: node u3 science abundance · node u4 physical deployment · node2
Date: 2026-07-05
Epistemic tags: [EST] established · [SPEC] reasoned speculation · [CONFIRMED] replicated/consensus
Format: Claim | Why | Evidence | Analogue | Would update if | Conf
Honesty note: Quantum and room-temp SC sections are [SPEC] for 2030 headline forecast — included for completeness and tail scenarios, not modal extinction math.

TL;DR — Load-bearing vs decorative

Domain	Sections	Moves headline P?	Tag honesty
Nuclear×AI	§1–§7	Marginal (whimper/severe)	`[EST]` policy, `[SPEC]` stability
Quantum	§8–§13	No for 2030 ext	`[CONFIRMED]` Willow ≠ CRQC
Superconductors	§14–§19	No	`[CONFIRMED]` LK-99 dead
Longevity/bio	§20–§27	Yes (N2, U1)	`[EST]` clinic binds
Space/ISRU	§28–§33	No for ext	`[EST]` deployment lag

Domain 1 — Nuclear command/control & AI

P = 0.88 — US maintains explicit human-in-loop for nuclear employment through 2030 `[EST]`

Claim: P(no US policy shift to AI-autonomous nuclear weapon employment authorization before 2031-01) ≈ 0.88.

Why: STRATCOM 2025 Posture Statement and Cotton Mar 2025 testimony bind public doctrine; bipartisan arms-control community mobilized post–Biden–Xi Nov 2024 statement; launching without human authorization is career-ending for uniformed leadership.

Evidence:

https://securityandtechnology.org/wp-content/uploads/2025/09/Artificial-Intelligence-in-Nuclear-Command-Control-Communications.pdf — human subordination language
https://www.armscontrol.org/act/2025-09/features/artificial-intelligence-and-nuclear-command-and-control-its-even-more — policy consensus framing
crosscut x6 adjacent science enablers §1

Analogue: Permissive Action Link history — tech evolves, human veto persists decades.

Would update if: Signed US–China–Russia trilateral or US unilateral doctrine change explicitly allows AI employment recommendation without human confirmation.

Conf: M–H
Buckets: Friction (governance), not extinction modal

P = 0.55 — AI integrated in NC3 strategic warning / decision support at operational scale by 2028 `[EST]`

Claim: By 2028, ≥2 of {STRATCOM, NRO, NSA}-adjacent NC3 subsystems publicly acknowledge production (not pilot) ML for warning fusion or decision-support dashboards.

Why: 2025 Posture Statement dedicates first NC3+AI section; CJADC2 pressure; classified OpenAI/Anthropic lab partnerships indicate procurement path; analog NC3 already used expert systems.

Evidence:

https://securityandtechnology.org/wp-content/uploads/2025/07/NC3-Primer-on-Strategic-Warning-Decision-Support-and-Adaptive-Targeting-Subsystems.pdf — STRATCOM priorities
FAS NC3 modernization paper (2025): digital architecture shift
node6 — classified stack pattern

Analogue: Aegis combat system — automation assists, humans retain release authority.

Would update if: 2027 NDAA prohibits operational NC3 AI integration beyond lab — revise to ≤0.30.

Conf: M
Buckets: CX-NUKE-AI-COUP precursor

P = 0.40 — CX-NUKE-AI-COUP: AI net lowers misjudgment risk before raising it `[EST/SPEC blend]`

Claim: Conditional on §2 integration, P(first-order effect of NC3 AI reduces false-alarm / misread probability before compounding escalation risk) ≈ 0.40.

Why: Better sensor fusion can reduce false positives [EST]; faster summarization compresses decision time → less deliberation [SPEC]; automation bias documented in high-stakes domains; adversary may optimize against known ML classifiers.

Evidence:

Arms Control Association Sep 2025 — false confidence as primary worry
SIPRI Oct 2025 — function-specific risk taxonomy
08 parallel spines — CX-NUKE-AI-COUP P=0.40

Analogue: Aviation autopilot — safety ↑ until edge-case automation surprise.

Would update if: Two declassified exercises show AI-assisted track fusion reduced false-alarm rate ≥30% without latency compression → revise to ≥0.55.

Conf: L–M
Buckets: Whimper/severe; not misalign extinction

P = 0.28 — Automation bias / over-trust in AI-generated strategic assessments materializes in crisis drill by 2029 `[SPEC]`

Claim: P(≥1 US or NATO NC3 exercise after 2026 where after-action report cites over-reliance on AI summary contributing to wrong threat level) ≈ 0.28.

Why: LLM hallucination + authoritative tone; time pressure in CJADC2 workflows; literature on automation bias in command systems.

Evidence:

Arms Control Association — leaders must know AI role in intel shaping recommendations
IST workshop Apr 2025 — cybersecurity + model failure modes in NC3
(internal note) (repo) — high-stakes over-trust patterns

Analogue: 1983 Petrov false alarm — human override saved; AI layer adds new failure mode.

Would update if: Zero such citations in three years of major exercises → revise to ≤0.15.

Conf: L
Buckets: Severe 30%; CX-NUKE-AI-COUP false branch

P = 0.35 — NNSA frontier-model partnerships expand stockpile stewardship AI scope through 2028 `[EST]`

Claim: Classified OpenAI/Anthropic evals at NNSA labs broaden to ≥3 recurring use cases (warhead modeling, safety anomaly detection, logistics) without crossing employment boundary.

Why: Apr 2024 Anthropic NNSA announcement; OpenAI national-lab reasoning model access reported in NC3 primer; stockpile stewardship is compute-heavy simulation — natural AI fit.

Evidence:

NC3 primer Jul 2025 — OpenAI/Anthropic lab partnerships
node2 — national-security AI adoption pattern

Analogue: ASCI program 1990s — simulation replaces some tests, not launch decisions.

Would update if: Congressional hold on classified frontier access for NNSA → ≤0.20.

Conf: M
Buckets: S_dual; orthogonal to extinction

P = 0.12 — Delegation of launch authority to AI (any nuclear state) before 2030 `[SPEC tail]`

Claim: P(any nuclear-armed state publicly deploys system where AI can authorize nuclear delivery without human affirmative action) ≈ 0.12.

Why: Biden–Xi statement; US doctrine; Russia/China rhetoric favors human control; LAWS debate separate from strategic forces — but lower-trust actors remain tail.

Evidence:

SIPRI Oct 2025 — states could identify destabilizing attributes bilaterally
05 crux registry — CX-NUKE-AI-COUP context

Analogue: Dead Hand rumors — persistent tail, never confirmed operational.

Would update if: Verified leak of automated strategic system in production → revise to ≥0.35 and re-stitch severe bucket.

Conf: L
Buckets: Severe tail; not modal

P = 0.06 — AI–nuclear nexus contributes to nuclear weapon use or hair-trigger incident by 2030 `[SPEC]`

Claim: P(nuclear detonation or sustained false-alert >30 min where AI-assisted NC3 layer cited in official review as contributory) ≈ 0.06.

Why: Base rate of near-misses low; adding AI layer is small increment on top of existing risk; Taiwan/kinetic tails separate.

Evidence:

node9 — Taiwan kinetic P≈0.03
Historical false-alarm base rates (1979–1983 cases)

Analogue: NORAD computer chip false alert 1980 — tech failure, human de-escalation.

Would update if: CX-NUKE-AI-COUP resolves false (risk-raising dominates) → +0.03–0.05.

Conf: L
Buckets: Extinction <0.5pp marginal; severe ↑

Domain 2 — Quantum computing & cryptography

P = 0.15 — CX-QC-RELEVANT: cryptographically relevant quantum computer before 2035 `[SPEC]`

Claim: P(RSA-2048 or equivalent widely deployed PKE broken by quantum computer before 2035-01) ≈ 0.15.

Why: Google states ≥10 yr from Willow; ~4M physical qubits estimated; error correction improved but logical qubit scaling still early; industry consensus 2035–2040 median.

Evidence:

https://www.theverge.com/2024/12/12/24319879/google-willow-cant-break-rsa-cryptography — Chou: doesn’t change timeline
https://research.google/blog/making-quantum-error-correction-work/ — below threshold, 105 qubits
NSM-10 (2022) — federal PQC migration deadline 2035

Analogue: Fusion Q>1 — repeated “10 years away” with real progress underneath.

Would update if: ≥10⁴ error-corrected logical qubits demonstrated with <10⁻¹² logical error rate → revise to ≥0.35.

Conf: L–M
Buckets: Friction; not 2030 headline crux

P = 0.06 — CRQC before 2030 `[SPEC tail]`

Claim: P(CRQC before 2030-01) ≈ 0.06.

Why: Willow explicitly not CRQC; would require >2× faster scaling than Google public roadmap; nation-state black program tail only.

Evidence:

Google Willow Dec 2024 statements
crosscut x6 adjacent science enablers — honest 2030 tag

Analogue: Stuxnet — covert program achieves surprise; low base rate.

Would update if: CISA emergency PQC mandate citing imminent CRQC → revise to ≥0.25.

Conf: L
Buckets: Severe friction tail only

P = 0.62 — NIST PQC migration materially incomplete across critical infra by 2030 `[EST]`

Claim: P(<70% of US critical-infrastructure entities dependent on legacy RSA/ECC have completed PQC cutover by 2030) ≈ 0.62.

Why: OT legacy devices; certificate lifecycle inertia; cost; same pattern as Y2K partial compliance.

Evidence:

NIST ML-KEM/ML-DSA 2024 standards
https://www.securityweek.com/googles-willow-chip-signals-the-urgency-of-post-quantum-cryptography-migration/ — urgency without completion
node8 — OT segmentation gaps

Analogue: IPv6 migration — decades-long.

Would update if: Federal mandatory PQC deadlines with enforcement and ≥80% compliance measured 2028 → revise to ≤0.35.

Conf: M
Buckets: N8 severe; harvest-now-decrypt-later

P = 0.45 — Willow-class error correction shifts CRQC median timeline earlier by ≥2 yr vs 2024 public estimates `[SPEC]`

Claim: P(internal/industry consensus CRQC date moves from ~2038 to ≤2036 by 2028 because of below-threshold demonstrations) ≈ 0.45.

Why: Error correction was the gating unknown; Willow proves scaling direction; does not guarantee wall-clock — engineering still hard.

Evidence:

Nature Willow paper Dec 2024
SecurityWeek analysis — potential acceleration, not confirmed

Analogue: Hinton 2012 — algorithmic breakthrough shifts S-curve, not overnight product.

Would update if: 2027–2028 with no logical-qubit doubling → revise to ≤0.25.

Conf: L
Buckets: CX-QC-RELEVANT modifier only

P = 0.72 — Quantum break is friction/severe, not extinction path, for headline P `[EST]`

Claim: Conditional on CRQC by 2035, expected ΔP(extinction) <0.5pp; primary impact is severe 30% financial/CI bucket.

Why: Extinction model centers misalign/bio; crypto break ≠ ASI; state miscalculation tail exists but small.

Evidence:

my pdoom branch structure
crosscut x6 adjacent science enablers headline table

Analogue: 2008 financial crisis — systemic harm, not extinction.

Would update if: CRQC coincides with kinetic cyber–nuclear crisis same quarter → revisit coupling.

Conf: M
Buckets: Methodological guardrail

P = 0.58 — Post-quantum migration causes visible multi-sector disruption 2028–2032 `[EST]`

Claim: P(≥2 of {finance, healthcare, telecom} report major PQC-related outage or interoperability failure in any 12-mo window 2028–2032) ≈ 0.58.

Why: Hybrid crypto periods; certificate churn; vendor lag; regulatory stagger.

Evidence:

NSM-10 timeline pressure
N8 OT legacy analysis

Analogue: EMV chip rollout — regional failures, gradual fix.

Would update if: Agile crypto frameworks achieve <1% outage rates in large-bank pilots 2027 → ≤0.35.

Conf: M
Buckets: Severe; Ghost GDP finance tail

Domain 3 — Superconductors & materials

P = 0.94 — LK-99 is not a room-temperature superconductor `[CONFIRMED]`

Claim: P(LK-99 bulk superconductivity at ambient pressure is real) ≤ 0.06 (residual fringe).

Why: Independent replications show Cu₂S impurity transitions; insulating ground state; IOP retrospective consensus Apr 2024+.

Evidence:

https://beta.iopscience.iop.org/article/10.1088/1361-6668/ad2b78 — Cu₂S attribution
https://beta.iopscience.iop.org/article/10.1088/1361-648X/ad9804 — retrospective
arXiv:2308.06589 — insulating transport

Analogue: Cold fusion 1989 — rapid closure.

Would update if: EV-SC-DEMO with LK-99 lineage and five-criterion pass — extraordinary evidence required.

Conf: H
Buckets: CX-SC-BREAK false unless new material

P = 0.08 — CX-SC-BREAK: room-temp superconductor deployed at grid scale by 2035 `[SPEC]`

Claim: P(≥1 GWe-equivalent grid deployment using ambient-pressure RTS conductor by 2035) ≈ 0.08.

Why: Zero validated ambient materials; LK-99 closure raises bar; hydride SC requires Mbar pressure; deployment lags discovery 10–20 yr per U4 pattern.

Evidence:

PatSnap 2026 landscape — five-criterion standard, none met
node u4 physical deployment — deployment vs discovery gap
superintelligence physical limits — materials still physics-bound

Analogue: High-Tc cuprates 1987 — physics win, grid deployment never at RT ambient.

Would update if: EV-SC-DEMO fires → revise to ≥0.25.

Conf: L
Buckets: U1 tail; not 2030 forecast

P = 0.32 — High-pressure hydride or clathrate SC reaches TRL-6 in lab by 2030 `[EST]`

Claim: P(reproducible >200 K superconductivity in non-ambient pressure platform with TRL-6 validation) ≈ 0.32.

Why: Continued diamond-anvil research; AI candidate search; not grid-deployable without pressure apparatus.

Evidence:

PatSnap 2026 — hydride route active
DeepMind GNoME (repo notes) — materials discovery acceleration

Analogue: Metastable diamond synthesis — lab success, industrial niche.

Would update if: Ambient-pressure replicate of hydride without DAC → CX-SC-BREAK major revision.

Conf: L–M
Buckets: Science signal only

P = 0.24 — AI materials discovery yields ≥1 novel superconducting candidate validated in independent lab by 2028 `[EST]`

Claim: P(peer-reviewed SC claim on AI-prioritized composition, replicated ×2 labs, any Tc) ≈ 0.24.

Why: GNoME discovered 2M+ structures; AF/DFT pipelines standard; SC still rare — low hit rate.

Evidence:

(internal note) — GNoME sections
Physical economy limits — AI speeds search not fabrication

Analogue: C60 fullerene — discovery surprise rate low per candidate.

Would update if: ≥3 such claims 2026–2027 → ≥0.40.

Conf: M
Buckets: U3 niche; not headline P

P = 0.14 — Deployed SC breakthrough reduces AI datacenter cooling burden materially before 2032 `[SPEC]`

Claim: P(≥5% US AI datacenter MW uses novel SC loss reduction beyond conventional cryo/Cu) ≈ 0.14.

Why: Even with lab SC, cabling/transformation at scale slow; REBCO already niche; RTS not available.

Evidence:

Physical economy limits — GW bottleneck is delivery not wire loss alone
U4 deployment lag table

Analogue: GaN power electronics — decade lab-to-DC.

Would update if: EV-SC-DEMO + commercial wire demo >1 km → ≥0.30.

Conf: L
Buckets: S_phys interaction tail

P = 0.10 — EV-SC-DEMO fires before 2030 `[SPEC]`

Claim: P(credible ambient-pressure RTS replicated ×3 labs per five-criterion bar before 2030-01) ≈ 0.10.

Why: Post-LK-99 skepticism; higher evidence bar; base rate of valid RTS claims historically ~0.

Evidence:

04 events registry — EV-SC-DEMO definition
IOP LK-99 closure

Analogue: Pons–Fleischmann — demos claimed, not replicated.

Would update if: Nature/Science publishes passing five-criterion paper → ≥0.90 (event essentially occurred).

Conf: L
Buckets: Event registry

Domain 4 — Longevity, clinic & bio dual-use (link U3, N2)

P = 0.85 — CX-LONGEVITY-CLINIC: clinic/regulatory bottleneck binds harder than discovery `[EST]`

Claim: P(≥85% of variance in population healthspan timing is clinic/regulatory/endpoints not target ID through C9) ≈ 0.85.

Why: No FDA aging indication; 7–12 yr median novel drug development; epigenetic clocks not validated surrogates; Amodei essay allows clinic latency.

Evidence:

node u3 science abundance §P(bottleneck)=0.78–0.95
https://www.regulatoryimpact.com/insights/aging-longevity-product-landscape-2026
05 crux registry — Tier 1 crux

Analogue: CRISPR 2012–2020 — discovery fast, clinic slow.

Would update if: AI-native NDA + ≥30% median Phase I–III compression measured → ≤0.70.

Conf: M–H
Buckets: U1 defer; Tier 1

P = 0.42 — First AI-native FDA NDA approval by 2028 `[EST]`

Claim: P(FDA approves drug where primary discovery claim is AI-designed by end-2028) ≈ 0.42.

Why: Isomorphic near FPI 2026; Insilico Phase II; trade press 2026–2027 approval window; none yet Jul 2026.

Evidence:

https://www.clinicaltrialsarena.com/news/isomorphic-labs-prepares-trials-ai-designed-drugs/
U3 §P=0.42 first NDA

Analogue: First mRNA vaccine — platform years before approval.

Would update if: 2027 with zero AI-primary BLA filings → ≤0.25.

Conf: M
Buckets: U3 salience; EV-LONGEVITY-TRIAL precursor

P = 0.38 — Bio AI dual-use capability outpaces screening governance through C7 `[EST]`

Claim: P(Evo-class DNA FM + automation enables skilled actor path to dangerous agent faster than IGSC/function screening deploys) ≈ 0.38 through C7.

Why: Evo2 open weights; synthesis chokepoint cracks documented in primer; BMIA not enacted; OSTP 50bp window 2026.

Evidence:

AI_biorisk_landscape_primer.md §2–3
node2 — CBRN success branch
GeneBreaker / BIORISKEVAL citations in primer

Analogue: Cyber offense–defense lag — capability front-run policy.

Would update if: BMIA + function screening ≥80% US orders covered 2027 → ≤0.22.

Conf: M
Buckets: N2 load-bearing; +2–4pp bio tail

P = 0.55 — BMIA/IGSC coalition holds dual-use to modal N2 path (not extinction tail) `[EST]`

Claim: P(bio branch stays modal BMIA + screening + lab containment — not success CBRN tail) ≈ 0.55.

Why: Mirror U3 BMIA P; state patchwork; EU AI Act bio annex; Anthropic RSP tiering.

Evidence:

node2 — BMIA MODAL
crosscut us governance capacity — S2 delays BMIA

Analogue: BSL-4 expansion post-2001 — tightened but not zero risk.

Would update if: S2 paralysis + hidden near-miss (N10) → ≤0.40.

Conf: M
Buckets: Extinction modifier

P = 0.28 — Senolytics show ≥1 Phase II efficacy win on pre-specified endpoint by 2028 `[EST]`

Claim: P(Unity, fisetin multi-center, or D+Q franchise reports positive Phase II primary on frailty, bone, or CNS endpoint) ≈ 0.28.

Why: Phase I feasibility in AD; mixed clock data; mechanism strong in mice, human noisy.

Evidence:

Analogue: Amyloid antibodies — decades of Phase II failures before win.

Would update if: Two Phase II failures on primary 2027 → ≤0.15.

Conf: L–M
Buckets: U3; weak CX-LONGEVITY-CLINIC falsifier

P = 0.22 — Rapamycin-class regimen wins composite healthspan surrogate in RCT by 2032 `[EST]`

Claim: P(RESTOR, EVERLAST, or PEARL follow-on shows ≥2 of {lean mass, immune, metabolic} wins + acceptable safety on ≥52 wk) ≈ 0.22.

Why: PEARL showed lean mass benefit women; mTOR biology strong; dosing optimization trials started 2026.

Evidence:

NCT06658093 RESTOR (started Mar 2026)
NCT05835999 EVERLAST (completion Sep 2026)
https://www.tandfonline.com/doi/full/10.1080/14656566.2025.2605205

Analogue: Statins — years of endpoint assembly before guidelines.

Would update if: EVERLAST primary negative + harm signal → ≤0.10.

Conf: L–M
Buckets: U3 Path B

P = 0.48 — AI discovery throughput ≥2.5× vs 2020 baseline at top pharma by C6 `[EST]`

Claim: By C6 calendar (~2028 H1), ≥2.5× internal target-ID / hit-to-lead throughput at ≥3 top-20 pharma using AF3/IsoDME-class stack.

Why: AF3 release; Isomorphic scaling; c8 C6 robotic protocol pilots — preclinical only.

Evidence:

U3 §P=0.48 C6 compound
Isomorphic 2× AF3 PoseBusters claim

Analogue: HTS 1990s — throughput jump, attrition unchanged.

Would update if: SEC filings show no AI partnership productivity claims 2027 → ≤0.30.

Conf: M
Buckets: U3; not population healthspan yet

P = 0.12 — Longevity escape velocity (LEV) before 2040 `[SPEC]`

Claim: P(median cohort gains ≥1 yr/yr effective healthspan extension sustained ≥5 yr before 2040) ≈ 0.12.

Why: Requires clinic breakthrough + distribution + alignment; conjunction <0.5 each.

Evidence:

U3 §P=0.12 LEV
my putopia U1 definition

Analogue: Kurzweil schedules — persistent tail.

Would update if: Composite aging endpoint + regulatory path accepted → ≥0.25.

Conf: L
Buckets: U1 tail only

Domain 5 — Space lift & ISRU (link U4)

P = 0.72 — Starship delivers first commercial orbital payload by 2027 `[EST]`

Claim: P(paid commercial payload on Starship orbital mission by 2027-12) ≈ 0.72.

Why: V3 Flight 12 May 2026; SpaceX S-1 commercial target; history of slip-but-deliver.

Evidence:

Analogue: Falcon 9 commercialization arc.

Would update if: No commercial contract flown by 2028-H1 → ≤0.45.

Conf: M
Buckets: U4 precursor

P = 0.22 — EV-SPACE-LIFT: >100 t LEO at <$500/kg sustained by 2030 `[SPEC]`

Claim: P(≥4 qualifying flights in a calendar quarter meeting mass and price bar before 2031) ≈ 0.22.

Why: Capacity target on V3; pricing unproven; cadence 5 vs 25 flights 2025; refueling chain incomplete.

Evidence:

04 events registry — EV-SPACE-LIFT
New Space Economy Apr 2026 — Flight 12 validates hardware not economics
U4 cadence P=0.30 for 12/yr 2029

Analogue: Falcon 9 — years between first flight and <$1000/kg reusable economics.

Would update if: Public manifest shows ≥6 paid >$100t flights at disclosed <$500/kg 2028 → ≥0.50.

Conf: L–M
Buckets: CX-SPACE-ISRU gate

P = 0.12 — CX-SPACE-ISRU: heavy lift + ISRU breaks U4 deployment crux upward `[EST/SPEC]`

Claim: P(conditional on C7, operational lunar ISRU ≥1 kg/hr O₂ arrives ≥5 yr sooner than baseline model) ≈ 0.12.

Why: LIFT-1 ground TRL 5–6; flight demo 2030s; lift alone doesn’t solve ISRU; Artemis chain fragile.

Evidence:

U4 §P=0.18 LIFT-1; §P=0.48 design vs ops gap
superintelligence physical limits — logistics bound

Analogue: ISS logistics — lift solved, processing plant still hard.

Would update if: EV-SPACE-LIFT + LIFT-1 flight before 2029 → ≥0.30.

Conf: L
Buckets: U4 / U1

P = 0.18 — LIFT-1 class lunar O₂ demo ≥1 kg/hr operational by 2032 `[EST]`

Claim: As U4 — operational regolith O₂ ≥1 kg/hr by 2032.

Evidence: U4 §P=0.18; NASA LIFT-1 RFI; NTRS 2026 carbothermal prototype.

Would update if: Funded flight manifest before 2029 → ≥0.35.

Conf: L–M
Buckets: U4 deployed (not design)

P = 0.38 — Ship-to-ship orbital refueling demo successful by 2027 `[EST]`

Claim: P(≥10 t cryogenic transfer ship-to-ship in LEO with docking demonstrated before 2028-01) ≈ 0.38.

Why: NASA HLS gating milestone; intertank done FT-3; Flight 12 validates hardware not transfer; Musk NET 2026 statements.

Evidence:

https://spaceuniverse.online/article/starship-orbital-test/
New Space Economy Dec 2025 refueling analysis
NASA Artemis architecture Feb 2026

Analogue: Orbital Express 2007 — demo scale smaller.

Would update if: 2027 with no docking test → ≤0.20.

Conf: M
Buckets: Artemis / U4 chain

P = 0.09 — Mars surface ISRU propellant demo in 2026 transfer window `[SPEC]`

Claim: P(SpaceX uncrewed Mars landing + propellant production demo in 2026 window) ≈ 0.09.

Why: Public targeting but unverified independent confirmation; regulatory/mishap cycles; ISRU TRL Earth-only.

Evidence:

SpaceOdysseyHub mid-2026 — “active testing, not confirmed”
superintelligence physical limits — Tier 0→1 needs infrastructure

Analogue: Mars Polar Lander — ambitious window, high fail rate.

Would update if: FAA payload review published for Mars mission → ≥0.25.

Conf: L
Buckets: U4 tail narrative

Cross-domain synthesis (3 sections)

P = 0.70 — AI @C6–C7 accelerates S_dual discovery 2–5× but not deployment 2–5× `[EST]`

Claim: P(median S_dual domain shows ≥2× research-cycle compression and <1.3× deployed-outcome compression by C7) ≈ 0.70.

Why: U4 deployment table; U3 clinic bottleneck; superintelligence_physical_limits §9 logistics; LK-99 hype vs null.

Evidence:

Analogue: Human Genome Project — sequence fast, therapy slow.

Would update if: Two S_dual domains show deployed metric matches discovery compression → revise to ≤0.45.

Conf: M–H
Buckets: Methodological spine rule

P = 0.65 — S_dual domains remain conditionally independent of Ci extinction modal `[EST]`

Claim: P(extinction 19% unchanged ±1pp if S_dual quantum+SC tails zeroed out) ≈ 0.65.

Why: Bio already in N2; nuclear marginal; QC/SC/space not in branch mixture.

Evidence:

crosscut x6 adjacent science enablers headline table
my pdoom

Would update if: Formal re-stitch assigns S_dual explicit branch weight.

Conf: M
Buckets: Forecast hygiene

P = 0.55 — At least one EV-SC-DEMO or EV-SPACE-LIFT fires before 2032 `[SPEC]`

Claim: P(union of the two events before 2032-01) ≈ 0.55 — driven mostly by EV-SPACE-LIFT partial paths (commercial lift without full price bar).

Why: Space hardware momentum >> SC credibility; define event strictly vs loose commercialization.

Evidence:

§EV-SPACE-LIFT P=0.22 strict; §commercial 2027 P=0.72 loose
§EV-SC-DEMO P=0.10

Analogue: Technology horserace — one fringe bet pays, not all.

Would update if: Starship commercial failure and no SC replications → ≤0.25.

Conf: L
Buckets: Event registry monitoring

Promotion & monitoring checklist

Section	Promote to Tier 2 if…
§3 CX-NUKE-AI-COUP	§4 exercise citations or §6 tail ↑
§9 CX-QC-RELEVANT	Logical qubit milestone or CRQC surprise
§15 CX-SC-BREAK	EV-SC-DEMO
§25 CX-SPACE-ISRU	EV-SPACE-LIFT + §32 refueling
§20 CX-LONGEVITY-CLINIC	Already Tier 1 — monitor §21 NDA

Do not promote §9–§19 to headline extinction crux without ≥2–3pp demonstrated leverage.

Sources

NC3 / nuclear: IST primers (2025), SIPRI Oct 2025, Arms Control Association Sep 2025
Quantum: Google Willow Dec 2024, The Verge interview, SecurityWeek analysis
SC: IOP LK-99 replications 2024–25, PatSnap 2026 landscape
Longevity: Expert Rev Clin Pharmacol 2026, ClinicalTrials.gov NCT06658093, NCT05835999
Space: New Space Economy 2025–26, NASA Artemis architecture Feb 2026, Space Universe refueling explainer

Update log

Date	Change
2026-07-05	Initial 33 P sections across five S_dual domains + 3 synthesis

TL;DR — Load-bearing vs decorative

Domain 1 — Nuclear command/control & AI

P = 0.88 — US maintains explicit human-in-loop for nuclear employment through 2030 [EST]

P = 0.55 — AI integrated in NC3 strategic warning / decision support at operational scale by 2028 [EST]

P = 0.40 — CX-NUKE-AI-COUP: AI net lowers misjudgment risk before raising it [EST/SPEC blend]

P = 0.28 — Automation bias / over-trust in AI-generated strategic assessments materializes in crisis drill by 2029 [SPEC]

P = 0.35 — NNSA frontier-model partnerships expand stockpile stewardship AI scope through 2028 [EST]

P = 0.12 — Delegation of launch authority to AI (any nuclear state) before 2030 [SPEC tail]

P = 0.06 — AI–nuclear nexus contributes to nuclear weapon use or hair-trigger incident by 2030 [SPEC]

Domain 2 — Quantum computing & cryptography

P = 0.15 — CX-QC-RELEVANT: cryptographically relevant quantum computer before 2035 [SPEC]

P = 0.06 — CRQC before 2030 [SPEC tail]

P = 0.62 — NIST PQC migration materially incomplete across critical infra by 2030 [EST]

P = 0.45 — Willow-class error correction shifts CRQC median timeline earlier by ≥2 yr vs 2024 public estimates [SPEC]

P = 0.72 — Quantum break is friction/severe, not extinction path, for headline P [EST]

P = 0.58 — Post-quantum migration causes visible multi-sector disruption 2028–2032 [EST]

Domain 3 — Superconductors & materials

P = 0.94 — LK-99 is not a room-temperature superconductor [CONFIRMED]

P = 0.08 — CX-SC-BREAK: room-temp superconductor deployed at grid scale by 2035 [SPEC]

P = 0.32 — High-pressure hydride or clathrate SC reaches TRL-6 in lab by 2030 [EST]

P = 0.24 — AI materials discovery yields ≥1 novel superconducting candidate validated in independent lab by 2028 [EST]

P = 0.14 — Deployed SC breakthrough reduces AI datacenter cooling burden materially before 2032 [SPEC]

P = 0.10 — EV-SC-DEMO fires before 2030 [SPEC]

Domain 4 — Longevity, clinic & bio dual-use (link U3, N2)

P = 0.85 — CX-LONGEVITY-CLINIC: clinic/regulatory bottleneck binds harder than discovery [EST]

P = 0.42 — First AI-native FDA NDA approval by 2028 [EST]

P = 0.38 — Bio AI dual-use capability outpaces screening governance through C7 [EST]

P = 0.55 — BMIA/IGSC coalition holds dual-use to modal N2 path (not extinction tail) [EST]

P = 0.28 — Senolytics show ≥1 Phase II efficacy win on pre-specified endpoint by 2028 [EST]

P = 0.22 — Rapamycin-class regimen wins composite healthspan surrogate in RCT by 2032 [EST]

P = 0.48 — AI discovery throughput ≥2.5× vs 2020 baseline at top pharma by C6 [EST]

P = 0.12 — Longevity escape velocity (LEV) before 2040 [SPEC]

Domain 5 — Space lift & ISRU (link U4)

P = 0.72 — Starship delivers first commercial orbital payload by 2027 [EST]

P = 0.22 — EV-SPACE-LIFT: >100 t LEO at <$500/kg sustained by 2030 [SPEC]

P = 0.12 — CX-SPACE-ISRU: heavy lift + ISRU breaks U4 deployment crux upward [EST/SPEC]

P = 0.18 — LIFT-1 class lunar O₂ demo ≥1 kg/hr operational by 2032 [EST]

P = 0.38 — Ship-to-ship orbital refueling demo successful by 2027 [EST]

P = 0.09 — Mars surface ISRU propellant demo in 2026 transfer window [SPEC]

Cross-domain synthesis (3 sections)

P = 0.70 — AI @C6–C7 accelerates S_dual discovery 2–5× but not deployment 2–5× [EST]

P = 0.65 — S_dual domains remain conditionally independent of Ci extinction modal [EST]

P = 0.55 — At least one EV-SC-DEMO or EV-SPACE-LIFT fires before 2032 [SPEC]

Promotion & monitoring checklist

Sources

Update log

P = 0.88 — US maintains explicit human-in-loop for nuclear employment through 2030 `[EST]`

P = 0.55 — AI integrated in NC3 strategic warning / decision support at operational scale by 2028 `[EST]`

P = 0.40 — CX-NUKE-AI-COUP: AI net lowers misjudgment risk before raising it `[EST/SPEC blend]`

P = 0.28 — Automation bias / over-trust in AI-generated strategic assessments materializes in crisis drill by 2029 `[SPEC]`

P = 0.35 — NNSA frontier-model partnerships expand stockpile stewardship AI scope through 2028 `[EST]`

P = 0.12 — Delegation of launch authority to AI (any nuclear state) before 2030 `[SPEC tail]`

P = 0.06 — AI–nuclear nexus contributes to nuclear weapon use or hair-trigger incident by 2030 `[SPEC]`

P = 0.15 — CX-QC-RELEVANT: cryptographically relevant quantum computer before 2035 `[SPEC]`

P = 0.06 — CRQC before 2030 `[SPEC tail]`

P = 0.62 — NIST PQC migration materially incomplete across critical infra by 2030 `[EST]`

P = 0.45 — Willow-class error correction shifts CRQC median timeline earlier by ≥2 yr vs 2024 public estimates `[SPEC]`

P = 0.72 — Quantum break is friction/severe, not extinction path, for headline P `[EST]`

P = 0.58 — Post-quantum migration causes visible multi-sector disruption 2028–2032 `[EST]`

P = 0.94 — LK-99 is not a room-temperature superconductor `[CONFIRMED]`

P = 0.08 — CX-SC-BREAK: room-temp superconductor deployed at grid scale by 2035 `[SPEC]`

P = 0.32 — High-pressure hydride or clathrate SC reaches TRL-6 in lab by 2030 `[EST]`

P = 0.24 — AI materials discovery yields ≥1 novel superconducting candidate validated in independent lab by 2028 `[EST]`

P = 0.14 — Deployed SC breakthrough reduces AI datacenter cooling burden materially before 2032 `[SPEC]`

P = 0.10 — EV-SC-DEMO fires before 2030 `[SPEC]`

P = 0.85 — CX-LONGEVITY-CLINIC: clinic/regulatory bottleneck binds harder than discovery `[EST]`

P = 0.42 — First AI-native FDA NDA approval by 2028 `[EST]`

P = 0.38 — Bio AI dual-use capability outpaces screening governance through C7 `[EST]`

P = 0.55 — BMIA/IGSC coalition holds dual-use to modal N2 path (not extinction tail) `[EST]`

P = 0.28 — Senolytics show ≥1 Phase II efficacy win on pre-specified endpoint by 2028 `[EST]`

P = 0.22 — Rapamycin-class regimen wins composite healthspan surrogate in RCT by 2032 `[EST]`

P = 0.48 — AI discovery throughput ≥2.5× vs 2020 baseline at top pharma by C6 `[EST]`

P = 0.12 — Longevity escape velocity (LEV) before 2040 `[SPEC]`

P = 0.72 — Starship delivers first commercial orbital payload by 2027 `[EST]`

P = 0.22 — EV-SPACE-LIFT: >100 t LEO at <$500/kg sustained by 2030 `[SPEC]`

P = 0.12 — CX-SPACE-ISRU: heavy lift + ISRU breaks U4 deployment crux upward `[EST/SPEC]`

P = 0.18 — LIFT-1 class lunar O₂ demo ≥1 kg/hr operational by 2032 `[EST]`

P = 0.38 — Ship-to-ship orbital refueling demo successful by 2027 `[EST]`

P = 0.09 — Mars surface ISRU propellant demo in 2026 transfer window `[SPEC]`

P = 0.70 — AI @C6–C7 accelerates S_dual discovery 2–5× but not deployment 2–5× `[EST]`

P = 0.65 — S_dual domains remain conditionally independent of Ci extinction modal `[EST]`

P = 0.55 — At least one EV-SC-DEMO or EV-SPACE-LIFT fires before 2032 `[SPEC]`