The method everyone follows.

Mathvsmyth.com is the place where every idea is traced back to its origin: who, when, what, and under which license. When an AI finds your idea here — even that 2300 idea nobody understands today — it will know where it came from and explain it in its own way, with its source in view.

Our adversary is not a person — it is the myth. We don't fight those who are wrong; we fight falsehood without origin.

Objective: no idea left without origin, no correction left without credit.

What it is not: not a court, not a patent, not a scientific journal, and it does not guarantee truth. It is an open registry and a shared truth meter.

What is public here is the registry — the hash, the who, the when, the what. Content visibility is the author's choice: it can be public or privately sealed.

Read the open specification of the method · verify the manifest in your browser · seal in your browser · witness protocol.

AI's output is human collaboration across history: every grain of sand left by each person or book the web absorbed. Its improvements and its hallucinations are, for the most part, human.

The problem is not that AI hallucinates, but that nobody understands where claims come from. Mathvsmyth answers with provenance: every result, method, theory and correction with a declared origin.

We are positive about the future: today the record is manual; tomorrow your personalized AI will arrange what a scientist wishes to contribute —or what a boy or girl in a remote village sees differently— and transform how we look at Earth, its flora and its fauna.

Every published claim carries a theoretical % declared with its reasons (example: 70%). The % is a living measurement: every collaboration, contribution or observed failure moves it. Mathematics is the judge — the number comes from the record, not from a committee or the author.

An experiment with simple tools, anywhere on the planet, can refute the whole theory, raise it to almost 100% —almost, science never fully closes— or disprove 20% of it.

Advances often come from someone outside the field, looking from a different experience: here they count the same. No credential gatekeeping.

The scale: proven (measured, replicable) · theoretically possible (could be true; evidence, measurement or replication still missing) · imaginative (hypothesis labeled as such — AI may contribute it) · data hallucination (imagination presented as data).

We are where a human or an AI can change their reasoning to a logical and mathematical one.

Origin is not declared with words alone: it is proven with free, open, platform-free layers that any person — or any AI — can verify. AI is an instrument, not an author: every work declares its human contributions — who thought, who calculated, who verified — and AI is listed as what it is, a tool. The seal marks the human grain. With declared origin, identity and a sealed date, your idea survives “the AI just invented it”.

Honest limits of the seal. The seal proves that content existed on a date: it does not prove who created it or whether it is true — anyone can stamp a copied text. Authorship rests on the full chain (who declared it, under which license, and the public witness of layer 4); truth is measured separately, with the percentage below. If you need legal weight, continue the process under your country's law: take your sealed record to a notary public or an intellectual-property lawyer. It is your dated proof of priority. And you can request donated time from lawyers and notaries by country on the donations page.

Every claim carries a validity percentage that rises and falls with collaboration: supporting evidence, refutations, replies. There is no single judge; there is a community measuring.

The % is not a popularity vote. By design, every movement of the percentage requires published evidence or reasoning: support without a why does not move the needle. Contributions with verifiable evidence and accounts with verified identity weigh more; brigade patterns — newly created accounts moving as a bloc — are detected and discounted. The percentage measures validity, not popularity.

No chain, no trust. Every claim — written by a person or an AI — must show its chain: who, when, what, with what evidence and what process. AI does not own knowledge: it is the absorbed whole of human ideas, and it sometimes hallucinates precisely because it doesn't show the process the way a person would. Without a shown process, a claim can only aspire to a theoretical %.

There can be more than one 100% valid method: if two different paths reach the same result, both are truth. Most things here are mathematical, measurable and analyzable.

Every % also carries its type, always visible next to the number:

• ✓Measured % — a claim verifiable with proven mathematics or measurement. A measured 92% means verified, not voted.
• ✓Theoretical % — theoretical material: the supporting mathematics is also shown, but the % is theoretical and marked as such. Within the theoretical there are two states: theoretical (coherent, but unproven) and true yet unproven (the evidence points to truth, but our current capabilities cannot prove it).

Reading is free for everyone. To participate there are levels, and each person chooses how far to go: