Short answers, no jargon

No. There's no mining, no coin, no energy cost, and no probabilistic "longest chain" rules. The public record is an ordinary append-only ledger whose hourly states are co-signed by a majority of independent witnesses — a much older and simpler idea. A record either has a majority of signatures or it doesn't; there's nothing to race and nothing to fork.

They keep the valuable part of their job — verifying that people really control the domains they claim — and lose the dangerous part: the unilateral power to issue a working certificate for any site in the world. Their checks become one voice among several instead of a single all-powerful signature. Most of their infrastructure and audit practices carry over unchanged.

There's a recovery path: fresh validation from many network vantage points, plus a mandatory seven-day public waiting period during which the owner's monitoring can object. It's deliberately slow — that's what makes it safe. High-security domains can also opt out of recovery entirely, accepting that a lost key means waiting for their records to expire, in exchange for a guarantee that no one else can ever take the domain over remotely.

The point is that you don't have to trust any of them individually. Witnesses are meant to be organisations with different interests and jurisdictions — browser makers, network operators, civil-society groups, national CERTs. A record only counts when a majority signs it, and mathematics guarantees two conflicting records can't both get a majority. Any one witness — or any minority — can do nothing alone.

No — never. That's a hard rule of the design, not an optimisation. Your browser picks up the hourly record state through its normal update channel (the same way it already gets certificate-revocation data today) and verifies each website locally, on your device. Nobody learns which sites you visit, and no external outage can break your browsing.

Real, and standardised. In 2024 the US standards body NIST finalised post-quantum algorithms (ML-DSA among them) after an eight-year public competition; governments and browsers are adopting them now. ConsensusPKI uses exactly these standards. The design's contribution isn't new cryptography — it's a structure where the much larger post-quantum signatures don't slow down your connection, because the heavy material is checked once per hour instead of on every visit.

We're precise about this, because most security projects aren't. If an attacker takes over a domain's DNS or its registrar account — the systems that decide who owns a name in the first place — no certificate system can fully save it; ConsensusPKI makes such an attack visible within an hour rather than invisible, and DNSSEC closes the gap where deployed. Similarly, if a website owner's own key is stolen, the thief can act as the owner until the key is rotated. What the design does guarantee: no authority, log operator or minority of witnesses can ever impersonate an established site.

Certificate Transparency records what authorities issue, so forgeries can be discovered afterwards — the forged certificate still works until someone notices. ConsensusPKI keeps the transparency but adds prevention: a certificate without the domain owner's own signature is simply invalid, no matter who issued it. Detection becomes the backstop instead of the whole defence.

Soon, in full. The complete design is written up as an academic paper being finalised for peer review; it will appear on the Research page once submitted, together with the reference implementation and the machine-checked proofs.