Nostradamus Intellect

Forecasts that can be proven wrong — sealed, dated, scored

News

Artificial Intelligence

Data & Analytics

Visit WebsiteSee on Product HuntTwitter ⧉

Hey Product Hunt — Quick disclaimer first: this is not decoded Nostradamus. No quatrain names 2027–2030, and the site says so on the record. We kept the long gaze and added a receipt. What you can do in about ten seconds, free, no OAuth: 1. Open the ledger — 16 sealed calls. Each has a locked probability, a fixed judgment date, and a written condition that would make it incorrect. 2. Pick the one you think we have wrong, and seal your own number next to it. Yours locks exactly like ours and is graded by the same rule on the same day. 3. Open the alert log — every sector alert we have ever raised, including the ones that came to nothing. It is append-only. We cannot delete our own false alarms. How it works, short: 8 named wires every 10 minutes → a published lexicon of 46 rules → 36 named strands → 24 hand-written couplings, each carrying a rationale and a historical precedent → a number with a kill-condition. A headline that cannot touch a strand does not enter a projection. There is no language model in the forecasting path. A sector alert needs 2.5 sigma against that strand's own 30-day baseline, carried by 3 independent feeds. Most weeks nothing clears it, and the log shows the quiet. On the money: 9 EUR opens one domain in full depth, 19 EUR opens all seven. Free and paid read the same Core — a pass buys filters, alerts and delivery, never a different probability. What I will not claim: any accuracy at all. All 16 seals are still PENDING. The first public grade is 28 February 2027, against a scoring rule published in September 2026 while everything was still unresolved. Ask me again in March. WHERE THIS IS GOING — AND AN ASK. The aim is a public instrument of record for the world-system: the place you check what was claimed, when, at what confidence, and whether it held. Forecasting has no equivalent of a court record today; everything is either an unfalsifiable take or a market price with no reasoning attached. Right now this is one person, and the honest constraint is not ideas or code — it is hours. There is more built here than I can properly maintain and extend alone. So if you are an engineer, a forecasting or calibration person, or someone who has run a research desk, I would genuinely like to talk — co-founder, contributor, or just someone who tells me which coupling is wrong. Open-sourcing part of the engine is something I am actively weighing; the constants, couplings and lexicon are already public, so the remaining argument is about contribution rather than transparency. If you have a view on that, I want to hear it. Ledger: https://nostradamusintellect.com... Alert log: https://nostradamusintellect.com... Engine room: https://nostradamusintellect.com... If you open one thing, open a near-term seal and put your number beside ours. I am here in the comments today.