← Back to all writing

AI futures evidence — X7: UBI / SWF variants

July 5, 2026

Futures index · 中文 · Main forecast

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


Parent: crosscut x7 ubi distribution variants
Extends: node u2 distribution
Labor gate: node1
Education: crosscut x4 education reskilling
Meaning: UBI post work meaning AI
Date: 2026-07-05
Format: Claim | Why | Evidence | Analogue | Would update if | Conf

Each section documents one probability or timing claim for S_soc X7. Probabilities are subjective elicitation unless noted as derived from U2.


TL;DR — top 5 P’s

PPoint est.Role
CX-UBI-SPEENHAMLAND0.55Cash transfer without hour reduction lowers real wages
EV-SWF (any G7 material by 2030)0.14Statute-level SWF/compute dividend — unlocks U3/U4
Federal UBI ≥$12k/yr by 20320.12Cash-only post-labor floor — low but non-zero
UBS expansion (≥1 G7 major)0.28Acemoglu-preferred services universalism
U4 path conjunction (transfer + direction + meaning partial)0.18Joint success metric for modest flourishing

Bucket leverage (if EV-SWF fires): U4 +4–6pp; friction −3–5pp; 28% → ~32–34% any-U rough order. If X7+X4 both fail: whimper +4–6pp; friction holds 53% modal.


P = 0.55 — CX-UBI-SPEENHAMLAND: UBI without hour reduction lowers real wages

Claim: ~55% that unconditional cash transfers enacted without coupled hour-reduction, wage boards, or strong bargaining power reduce real wages via employer capture — not merely fail to raise them.

Why: Speenhamland outdoor relief (1795–1834) allowed employers to pay less; modern EITC incidence literature shows partial capture; OpenResearch found −$1,500/yr labor income on $12k/yr transfer (−12.5% offset). Alaska dividend too small to test at post-labor scale.

Evidence:

Analogue: Speenhamland 1795–1834 — parish relief → wage suppression until Poor Law reform 1834.

Would update if: National UBI + 32-hour week law or labor share ↑ ≥2 pp within 3 yr of enactment.

Conf: M


P = 0.14 — EV-SWF: G7-scale SWF or compute dividend enacted by 2030

Claim: ~14% (range 0.08–0.22) any G7 country enacts material social wealth fund or compute-dividend statute — budget line + dividends ≥5% median income equivalent.

Why: Sanders S.4825 sets Overton window but 0 cosponsors; Altman 5% voluntary ≠ law; Norway template took decades; federal gridlock baseline (Node 6); U2 P(material G7 redistribution by 2029) 0.25–0.35 — X7 uses lower bar for any G7.

Evidence:

Analogue: Alaska Permanent Fund 1982 — 26 years from Prudhoe Bay to dividend culture.

Would update if: ≥10 Senate cosponsors + CBO score on S.4825 in 2027; or EU digital-dividend directive with enforceable penalties.

Conf: L–M


P = 0.10 — US federal AI SWF enacted at Sanders scale (≥30% equity) by 2032

Claim: ~10% S.4825-class 50% equity excise becomes US law — tail-adjacent, not zero.

Why: Constitutional takings politics; industry + antitrust opposition; no cosponsors; Altman counters at 5%; Trump admin prefers voluntary “public partner.”

Evidence:

Analogue: Warren wealth tax 2019 — agenda shift, no enactment.

Would update if: Second-majority Senate with Sanders-class coalition + AI macro crisis binding vote.

Conf: L


P = 0.22 — Voluntary / negotiated 5–10% public stake in ≥2 frontier labs by 2029

Claim: ~22% administration or successor deal: equity slice to sovereign fund without Sanders compulsion — B/E variant path.

Why: Altman-Trump discussions Jun–Jul 2026; “American public as partner” rhetoric; CHIPS Act precedent for conditional federal stake; OpenAI $852B valuation → 5% ≈ $42.6B.

Evidence:

Analogue: TARP warrants 2008 — crisis-driven equity participation.

Would update if: Explicit admin rejection of any equity stake in writing; or DOJ antitrust blocks cross-lab fund structure.

Conf: L–M


P = 0.12 — Federal UBI (≥$12k/yr/adult unconditional) by 2032

Claim: ~12% unconditional federal floor at post-labor-relevant scale — below activist $24k, above Alaska.

Why: OpenResearch cooled enthusiasm; fiscal cost ~$3T+/yr at $12k; Acemoglu opposition; no AI-funded society-scale test; acceleration Congress anti-entitlement.

Evidence:

Analogue: Nixon Family Assistance Plan 1969 — near-miss, abandoned.

Would update if: CBO scores funded UBI with compute/SWF revenue and passes Senate with hour-reduction bundle.

Conf: L


P = 0.28 — UBS expansion (healthcare/education/housing guarantees) in ≥1 G7 by 2030

Claim: ~28% major expansion of service universalism — closer to Acemoglu menu than Yang UBI.

Why: Sanders SWF spend side includes “decent standard of living” bundle; Nordic template; US ACA incremental path; Dem coalition prefers services over cash (P 0.55).

Evidence:

Analogue: NHS 1948 — service universalism post-shock.

Would update if: US Medicare-for-All fails 2028 cycle with no UBS substitute in platform.

Conf: L–M


P = 0.18 — “Compute dividend” / universal basic compute at scale (@C6–C8)

Claim: ~18% meaningful public inference entitlement — monthly token/GPU credits ≥ professional tier, not rhetoric alone.

Why: Altman 2025 “universal basic compute” pivot from 2021 cash fund; EU digital sovereignty talk; administrability and abuse harder than cash; marginal cost nonzero.

Evidence:

Analogue: Lifeline phone program — universal floor, not abundance.

Would update if: Federal entitlement bill for inference credits passes committee with ≥$5B/yr appropriation.

Conf: L–M


P = 0.06 — Altman American Equity Fund (2.5% cap + 2.5% land) enacted as designed by 2032

Claim: ~6% the 2021 Moore’s Law for Everything equity+land fund becomes law — distinct from 2026 5% voluntary SWF seed.

Why: Never introduced as bill; Altman pivoted to compute-as-right; land tax politically toxic; superseded by voluntary negotiation track.

Evidence:

  • Moore’s Law for Everything — $13,500/yr/adult projection over 10y
  • UBI post work meaning AI §Altman table
  • Forbes Jul 2024 — Altman UBI study results shifted his public emphasis toward ownership/compute

Analogue: Buffett rule / wealth tax variants — decades from proposal to minimal enactment.

Would update if: White House submits American Equity Fund as formal legislative proposal.

Conf: L


P = 0.62 — Cash-only transfer (variant A) fails U2 normative bar @C8+ even if income met

Claim: ~62% unconditional cash without Jahoda structures (time structure, social contact, collective purpose, status, activity) → anomie — blocks U2 bucket.

Why: Putnam Bowling Alone; retirement depression literature; lottery winners; Bostrom status problem; U2 same estimate in parent node.

Evidence:

Analogue: Gulf rentier youth underemployment — income without purpose.

Would update if: National community-service corps >10% adult participation with measured wellbeing gains.

Conf: M


P = 0.70 — Cash-only (A) without X4 direction policy blocks U4 flourishing path

Claim: ~70% that U4 “machine usefulness + managed inequality” requires either services/JG (C/D), capital dividend (B/E), or successful reskilling (X4) — not standalone cash.

Why: Acemoglu so-so automation dominates (P 0.65); cash doesn’t change deployment direction; OpenResearch employment dip signals Speenhamland risk; U4 needs wage-share recovery not just transfers.

Evidence:

Analogue: Post-2008 UI expansion without industrial policy — demand patch, structural inequality persists.

Would update if: Randomized national cash UBI shows labor share ↑ alongside transfer.

Conf: M


P = 0.18 — U4 path conjunction: {material transfer} ∩ {direction partial} ∩ {meaning partial} by 2032

Claim: ~18% joint success — aligns with U2 “broad sharing” P 0.22 and my_putopia U4 bucket; conjunction discount applied.

Why: Product of {EV-SWF or UBS 0.22–0.28} × {machine-usefulness tilt 0.25} × {U7 not collapsed 0.38} × {Speenhamland avoided 0.45} ≈ 0.09–0.20 central 0.18.

Evidence:

Analogue: Post-WWII bundle — GI Bill + unions + tax, not single instrument.

Would update if: CBO distributional score shows bottom quintile +20% from AI policies alone by 2030.

Conf: L–M


P = 0.68 — C4 policy window produces labor rules not distribution settlement (N1 mirror)

Claim: ~68% modal output = WARN tweaks, disclosure, EU oversight — not SWF/UBI — same as U2/N1 stitch.

Why: Vettting coalition wins over transfer coalition; Sanders moratorium P 0.03; SWF P 0.10; culture-war jobs frame (N1 P economic anxiety ≫ x-risk 0.85).

Evidence:

Analogue: Y2K compliance industry — massive activity, no structural reform.

Would update if: SWF or UBI equivalent passes during 2027 Q2 salience peak.

Conf: M–H


P = 0.58 — Path D modal friction through C8 (distribution variant failure)

Claim: ~58% branch mass: acceleration + inequality + institutional lag — no utopia bucket fully satisfied via X7 instruments.

Why: c8 default; concentration P 0.73; Ghost GDP sector-visible P 0.58; variant mix = B-lite talk only.

Evidence:

Analogue: 1995–2025 internet — consumer surplus ↑, inequality ↑, no post-labor settlement.

Would update if: EV-SWF + labor share inflection in same 24 mo window.

Conf: M


P = 0.58 — Ghost GDP sector-visible by C6 without transfer enactment

Claim: Corporate margins ↑, payroll flat in AI-heavy sectors legible before economy-wide crisis — anti-utopia early signal for X7 (transfers lag).

Why: So-so automation; Block −40% HC; Canaries −16%; machines spend $0 on discretionary goods.

Evidence:

Analogue: 1920s farm income collapse before macro crisis.

Would update if: AI-exposed sectors show payroll ↑ and margin ↑ simultaneously through 2028.

Conf: M


P = 0.55 — Acemoglu anti-UBI stance shapes Democratic distribution menu through 2030

Claim: ~55% major Dem packages emphasize services/jobs (C/D) over unconditional cash (A/F).

Why: Power and Progress Ch.11 explicit anti-UBI; Brookings/NBER pro-worker lane; OpenResearch gave ammunition to skeptics.

Evidence:

Analogue: Nordic services universalism vs US cash debates.

Would update if: 2028 Dem platform centers unconditional $1k+/mo federal without services bundle.

Conf: M


P = 0.35 — OpenResearch 2024 results materially cooled US cash UBI coalition

Claim: ~35% that the −4.1 pp employment / −$1.5k labor income findings shifted elite opinion against variant A — not fatal but measurable.

Why: Altman-funded study; NBER credibility; City Journal / skeptical uptake; Amodei “UBI dystopian” contrast.

Evidence:

Analogue: Oregon Medicaid lottery — one RCT moves priors for policy elites more than public.

Would update if: Follow-up OpenResearch papers show sustained wellbeing + neutral employment at year 5+.

Conf: M


P = 0.85 — Alaska-scale dividend alone insufficient for post-labor transition

Claim: ~85% that transfers ≤$2k/yr/adult cannot substitute for employment income at C8 displacement levels — rules out B analogue-only success.

Why: Alaska ~$1–2k vs ~$40k+ median wage; no work-participation effect because too small; post-work needs $20–30k+ UBI_post_work_meaning_AI.md.

Evidence:

  • Alaska Permanent Fund dividend history — no post-work society
  • node u2 distribution §Alaska too small

Analogue: Alaska 1982 — proof of concept, not scale.

Would update if: National dividend >$10k/yr funded by AI levy sustained 5+ years.

Conf: H


P = 0.15 — Norway-style compute/electricity → SWF template copied in G7 by 2032

Claim: ~15% G7 routes compute lease royalties or datacenter electricity surcharges into permanent fund (not general revenue).

Why: Oil → GPFG needed visible exhaustible rent; AI rent = monopoly + data, harder to classify; Gulf may lead US.

Evidence:

Analogue: Alaska 26-year lag discovery → meaningful dividend.

Would update if: UAE/Saudi announce citizen compute dividend >$5k/yr.

Conf: L


P = 0.08 — Federal Jobs Guarantee pilot (H.R.7566 / S.3864) enacted by 2028

Claim: ~8% Federal Jobs Guarantee Development Act of 2026 passes — competitive grants for local JG pilots, not national employer-of-last-resort.

Why: Introduced Feb 2026; pilot scale; fiscal cost; no AI-specific branding; progressive lane crowded by SWF talk.

Evidence:

Analogue: TANF block grants — pilot rhetoric, constrained scale.

Would update if: Markup with ≥$5B authorized in 119th/120th Congress.

Conf: L


P = 0.15 — Yang Freedom Dividend legacy retains state/local pilot influence through 2028

Claim: ~15% Yang 2020 frame still drives state/local guaranteed-income pilots ($500–1000/mo) — F as Overton, not federal A.

Why: Stockton SEED, Compton, dozens of Stanford Basic Income Lab tracked pilots; no federal VAT politics; 2020 primary normalized “freedom dividend” phrase.

Evidence:

Analogue: Cannabis — state labs precede federal by decade+.

Would update if: Zero new municipal/state GI pilots launched post-2027 despite C4 shock.

Conf: L–M


P = 0.40 — Newsom EO N-6-26 “universal basic capital” produces CA statute by 2029

Claim: ~40% California converts EO study into binding worker-ownership / capital-dividend / SWF-seed legislation — strongest E/B state track.

Why: Newsom “universal basic capital” rhetoric May 2026; 180-day recommendations ~Nov 2026; CA as policy lab (CCPA model); employee ownership (NCEO) alignment.

Evidence:

Analogue: CCPA 2018 → national pressure.

Would update if: Nov 2026 recommendations are study-only with zero 2027 bill introduction.

Conf: M


P = 0.38 — State-level AI-adjacent revenue / dividend measure (CA/NY/WA) by 2028

Claim: ~38% ≥1 large state enacts compute tax, data-center surcharge, equity-severance mandate, or citizen dividend fund with enforcement.

Why: CA SB 53 ecosystem; Newsom EO; states as labs when federal gridlocked; overlaps U2 estimate.

Evidence:

Analogue: CA cap-and-trade before federal carbon policy.

Would update if: GAAIA preemption without carve-outs blocks CA labor+distribution (Node 1 P 0.12).

Conf: M


P = 0.14 — US federal token/compute/electricity excise enacted by 2028

Claim: ~14% narrow levy on inference or datacenter electricity becomes law — funds B dividend indirectly.

Why: Administrability ↑ vs robot definition; UK electricity proposals; Mark Cuban token tax rhetoric; Summers anti-robot-tax but pro-electricity base.

Evidence:

Analogue: Carbon border adjustments — sector tax before income overhaul.

Would update if: EU energy package explicitly exempts AI inference through 2028.

Conf: L


P = 0.45 — X4 fails AND X7 SWF blocked → whimper +4–6pp (joint failure)

Claim: ~45% conditional on both CX-EDU-RESPONSE and EV-SWF absence — whimper mass rises +4–6pp vs baseline.

Why: Education is political substitute for transfers; Acemoglu path dead; cash UBI blocked by Speenhamland fear + GOP; meaning collapse channel opens.

Evidence:

Analogue: Post-2008 — TAA underfunded + no structural wage policy → lasting scar.

Would update if: EV-SWF fires without X4 success but labor share still recovers (capital dividend sufficient).

Conf: M


P = 0.30 — X4 success + UBS (C) without SWF still opens partial U4 (+2–3pp)

Claim: ~30% conditional on EV-EDU-FED or strong corporate placement — Acemoglu bundle achieves partial U4 without cash/SWF.

Why: Machine usefulness + skilled deployment; wage pass-through in augmented roles; insufficient for U3 golden age without capital sharing.

Evidence:

Analogue: German dual VET — deployment absorption without post-scarcity dividend.

Would update if: X4 success but Ghost GDP accelerates with zero transfer talk (pure so-so wins).

Conf: M


P = 0.35 — Shared-gains narrative wins in media/policy @C4–C6 (not enactment)

Claim: ~35% “Workers must share AI gains” becomes bipartisan talk-track — necessary not sufficient for X7 success.

Why: Newsom framing; Sanders 79% poll; Trump “public partner”; Altman equity rhetoric.

Evidence:

Analogue: Healthcare public option 2009–10 — narrative win, law loss.

Would update if: No major 2028 candidate mentions AI dividend.

Conf: M


P = 0.40 — Staged transition (tax fix → levy → pilot dividend → expand) starts by 2028

Claim: ~40% at least two stages of happy_path §C pipeline underway legislatively — not completed.

Why: Newsom EO + SB 951 + Sanders + Altman = stage 0–1; enactment weak but activity nonzero.

Evidence:

Analogue: ACA waiver experiments 2011–13.

Would update if: Zero legislative activity 2027 after Block-scale events.

Conf: M


P = 0.73 — Top decile captures ≥60% AI-attributed surplus through C8 (concentration anchor)

Claim: Primary competitor to all X7 variants succeeding — rents stay with compute owners unless EV-SWF class event.

Why: Ghost GDP; SWF low enactment; 1980s computer revolution base rate; circular AI capex beneficiaries concentrated.

Evidence:

Analogue: Oil majors 1970s before Norway GPFG matured.

Would update if: Labor share + SWF dividends + wage growth jointly offset ≥40% of AI productivity to bottom 50%.

Conf: M


P = 0.25 — Finland / Kenya RCT evidence generalizes to AI-funded fiscal UBI

Claim: ~25% — low — that existing RCTs predict outcomes when transfers are AI-tax-funded at society scale.

Why: All pilots externally funded; no inflation/tax incidence test; OpenResearch US scale still <$15k/yr effective; Kenya entrepreneurship context differs.

Evidence:

Analogue: Drug trial in healthy volunteers vs comorbid population.

Would update if: Nation-scale AI levy funds UBI with published macro general-equilibrium results.

Conf: L


Correlation notes (X7-specific)

  • Substitute variants: P(A) and P(B Sanders) negatively correlated — same political window, different coalitions (ρ ≈ −0.30)
  • Complement: P(C UBS) and P(X4 EV-EDU-FED) (ρ ≈ +0.40)
  • Crisis override: Ghost GDP macro-binding ↑ P(voluntary 5% stake) (ρ ≈ −0.25 with concentration)
  • Do not multiply variant enactment P’s — at most one primary federal instrument by 2032

Bucket mapping summary

X7 outcome @2032U3/U4Friction 53%Whimper 24%Extinction 19%
EV-SWF fires+4–6pp−3–5pp−2–4pp~0
Modal patchworkcappedholds+1–2pp~0
Speenhamland partial cashblocked+2–3pp+3–4pp~0
X7+X4 joint failblockedholds+4–6pp+0.5pp race tail

Update log

DateChange
2026-07-05Initial X7 evidence pass — 25 sections; web 2024–26 (S.4825, OpenAI 5%, Newsom EO, H.R.7566, OpenResearch)

Section count: 25 evidence sections.