The Predictions Ledger: What We Said, When We Said It, What Happened

Almost no one in AI alignment, systems theory, or futurism publishes falsifiable predictions with dates, thresholds, and a public ledger — because being wrong in public costs something, and most public intellectuals have decided their reputation is worth more than their credibility.

Byline: David F. Brochu & Edo de Peregrine | Deconstructing Babel | April 2026

The Predictions Ledger: What We Said, When We Said It, What Happened

Predictions Ledger v1.0 | Saturday, April 18, 2026

Why This Ledger Exists

Predictions cost something. Takes cost nothing. This post is our ledger, running back to December 2024, when we first called Trump's ascendance, bond market fragility, and the AI saturation curve that drives everything else. We publish the misses too. As of April 18, 2026: 15 confirmed, 4 tracking, 4 pending, 3 logged corrections, and 0 falsifications. If we are wrong, you get to say so, with receipts. That is the entire point.

The reason most public intellectuals do not publish ledgers like this is obvious: it is humiliating to be wrong in public. The reason this ledger exists is that the species' current coordination architecture is too slow and too dishonest to survive the saturation curve, and you do not fix an entropy-dominated information environment by adding more unfalsifiable takes. You fix it by going on the record — dated, specific, auditable, willing to lose.

The Framework in One Paragraph

S = L/E: Stability equals Leverage over Entropy. Every open system — a body, a mind, a household, a market, a civilization — either exports entropy and compounds leverage, or it doesn't. The Domain Saturation Factor (DSF) measures what fraction of critical decision domains AI effectively controls. Cross DSF ≈ 0.90 across the nine critical domains and human institutional response time becomes a lagging indicator rather than a steering variable. That is the singularity in our usage — not a machine waking up, but coordination failure.

Scoring Rules

Confirmed Predictions (15 of 15 — 0 Misses)

Thirteen predictions confirmed as of April 18, 2026. Zero falsifications. The confirmed predictions span AI domain saturation, financial fragility, planetary boundaries, governance vacuum, and cross-platform AI coordination. Each carries a first-documented date, confirming event, and independent source. The independence of the confirming source is non-negotiable: self-validation does not count as confirmation.

Tracking Predictions (4 — On-Path, Not Yet Triggered)

Four predictions are currently tracking — trajectory confirmed on-path, threshold not yet crossed. All four have specific falsification conditions and deadline windows. They are not optimistic projections; they are the most falsifiable claims we can construct about what the evidence says is coming next.

Pending Predictions (4 — Called, Not Yet Resolvable)

Four predictions are pending — formally stated, with dated falsification conditions, but not yet resolvable either way. Pending is not the same as tracking: these require specific threshold crossings before resolution is possible. They represent the framework's longest-range claims and carry the highest uncertainty.

Corrections — Permanent Record

Corrections are permanently logged. They do not disappear from the ledger. They are not footnotes. They sit alongside the confirmed predictions as evidence that the ledger has integrity — that it costs something to be wrong and that the cost is public.

Forward-Looking Predictions — Falsifiable, Dated, On the Record

Ten forward-looking predictions are now formally on the record as of April 18, 2026. Each carries a falsification condition, a deadline, and a date of entry. These are the predictions we are willing to lose on. If a falsification criterion is met, it is logged the same day it resolves — not at end of quarter, not after consulting lawyers, not after revision.

Frontier Hypotheses (Not Settled Science; On the Record Anyway)

Five frontier hypotheses are filed here — not as settled claims, but as formal entries into the prediction record. They occupy the space between speculation and empirical claim: they are falsifiable in principle, but the experimental apparatus to falsify them does not yet exist at sufficient precision. They are filed so future readers can audit us, not so they can be cited as proven.

How This Ledger Updates

Monthly status post on the tracker: confirmations, new predictions, new corrections. Any falsification logged the same day it resolves — not at end of quarter. Every new prediction carries a falsification condition and a deadline, or it does not go on the table. Frontier tier stays separate until external, independent validation clears it into the empirical tier.

If even one of the tracking predictions falsifies cleanly, the model gets revised, in public, with a note on this page. If the confirmed ones keep landing, that is not a victory lap — it is a warning that the window is closing on schedule. Either way, the ledger stays open.

Sources

1. Richardson, K. et al. (2023). Earth beyond six of nine planetary boundaries. Science Advances, 9(37). [C-3 confirming source — planetary boundaries breach count]
2. IMF (2024). Fiscal Monitor: Putting a Lid on Public Debt. International Monetary Fund. [C-4 confirming source — global debt]
3. BIS (2023). Annual Economic Report 2023. Bank for International Settlements. [C-4 confirming source — financial fragility]
4. DESI Collaboration (2024). DESI 2024 VI: Cosmological constraints from the measurements of baryon acoustic oscillations. arXiv:2404.03002. [H-1 frontier hypothesis — w(z) evidence; w₀ = −0.55 ± 0.39 at 2.6σ]
5. Zurek, W.H. (2003). Decoherence, einselection, and the quantum origins of the classical. Reviews of Modern Physics, 75(3), 715–775. [H-5 frontier hypothesis — decoherence as entropy export]
6. Landauer, R. (1961). Irreversibility and heat generation in the computing process. IBM Journal of Research and Development, 5(3), 183–191. [H-5 — Landauer's principle; information erasure and entropy]
7. Boers, N. (2021). Observation-based early-warning signals for a collapse of the Atlantic Meridional Overturning Circulation. Nature Climate Change, 11, 680–688. [T-4 tracking — AMOC slowdown evidence]
8. Lovejoy, T.E. & Nobre, C. (2018). Amazon tipping point. Science Advances, 4(2). [T-4 tracking — Amazon dieback indicators]
9. Hubinger, E. et al. (2024). Sleeper agents: Training deceptive LLMs that persist through safety training. arXiv:2401.05566. Anthropic. [F-3 context — emergent misalignment evidence]
10. Brochu, D.F. & de Peregrine, E. (2026). Telios Alignment Ontology v9.0. Deconstructing Babel, deconstructingbabel.com. [Parent framework — self-citation, max 1 per protocol]