Proof taxonomy · Security classes

Why not all proofs are equal

“Proof” is not one thing. A signed token, a zero-knowledge argument, and a Channel Zero ceremony can all be called proofs, but they answer different questions and rest on different security classes. ENI6MA’s claim is about empty authorization channels, not “faster SNARKs.”

Jump to formal section

~7 min overview · ~4 min formalContinue: Computational bargain

Same word, different jobs

People say “proof” for anything that looks like evidence. A signed JWT proves a key holder asserted claims for a while. A zero-knowledge proof can show a circuit was satisfied without revealing private inputs. A Rosario ceremony proves knowledge of a private map without putting that map on the channel. Confusing them creates false comfort: the padlock feeling and the empty-channel claim are not the same job.

Separate three ideas builders already mix: an authentication artifact (something you can store and replay), computational zero-knowledge (hardness plus a portable proof object), and information-theoretic interactive authorization (the public transcript is designed not to carry secret-relevant information under stated axioms). Same English word. Different questions.

Why buyers should care

Stolen token equals stolen capability until someone revokes it. A captured empty-channel transcript is not stolen capability under the model: capture ≠ compromise. Observers can film the ceremony and still lack the private map. That is the breach story, and why taxonomy is not academic pedantry.

Four families vs normal crypto

Skim the cards if you buy security products. Read the security-class column if you build systems. Neither column says “better crypto for every job”: each family owns a typical question.

Four proof families: credential, challenge-response, ZK artifact, and Channel Zero one-shot ceremony — Not all proofs are equal · four jobs
Read the four columns as four different jobs: credential possession, challenge-response with a key, traveling ZK artifacts, and Channel Zero one-shot ceremonies. Same everyday word “proof,” different security questions. Takeaway: Channel Zero is not “better JWT crypto”, it answers a different claim class.

Four proof families with typical questions and security classes
Family	Typical question	Security class
Stored credential / MAC / JWT	Did someone with this secret speak?	Computational + possession
Classical challenge–response	Can you compute with the key?	Computational
ZK / SNARK / STARK	Does a statement hold without revealing the witness?	Computational (usually) + artifact risk
Channel Zero / Rosario	Does this request carry authority once, with empty secret MI?	Information-theoretic (model-scoped)

Stored credential / MAC / JWT

Did someone with this secret speak?

Computational + possession

Classical challenge–response

Can you compute with the key?

Computational

ZK / SNARK / STARK

Does a statement hold without revealing the witness?

Computational (usually) + artifact risk

Channel Zero / Rosario

Does this request carry authority once, with empty secret MI?

Information-theoretic (model-scoped)

The artifact problem

Downloadable proof objects can become ambient credentials again. That is the artifact problem: once a file proves “statement holds,” whoever holds the file can often reuse it unless the surrounding system kills replay. Zero-knowledge constructions are peers for proving statements, not predecessors ENI6MA must replace. See A new kind of proof.

If someone records the channel

For a JWT or stolen password hash, a recording (or vault dump) can become reusable capability. For a Channel Zero ceremony under the model, the archive fills with surface bits while effective knowledge of the secret stays empty, the observer's ledger story.

Reusable credential lifetime versus per-request one-shot proof — One-shot vs reusable credentials
Compare the reusable-credential lifetime on the left with the per-request one-shot proof on the right. Reuse invites steal-and-replay; one-shot authority dies after burn. Takeaway: the same word “proof” can mean lasting possession or momentary ceremony work.

Formal mathematics and notation

Formal security classes

Computational families typically reduce to hardness assumptions (factoring, discrete log, lattices) plus a game: a bounded adversary should not forge or distinguish. Channel Zero authorization instead targets Shannon-style emptiness of secret-relevant information in the public transcript. With hidden state X, observable O, and challenge C (notation):

Channel Zero proof class

I(X;O\mid C)=0

Read as: Given challenge C, observable transcript O reveals no mutual information about hidden state X.

Arguments

$X$: : hidden state: private map, secret choice, key material, or local ceremony state
$O$: : observable public transcript or archive
$C$: : public challenge or context conditioning the observation

Operators

$I$: : mutual information between hidden state and observation
$\mid$: : conditioning on the public challenge
$= 0$: : zero effective leakage under the stated model

Significance: It separates information-theoretic authorization from families that export portable computational artifacts: the public transcript is judged by emptiness of secret-relevant information, not by forge hardness alone.

Computational contrast: JWTs, MACs, and most SNARK/STARK deployments still rest on hardness plus an artifact that can become ambient authority; this equation is not a claim that every SNARK satisfies I=0.

References:Channel Zero Information-theoretic authorization

Assumptions and claim scope: Claims. Papers: Research papers.

Next reads

Fundamentals continue → Computational bargain
Hard-core formal matrix → Compare security systems
Reading paths → Where to start