I built a website last week. It has a padlock in the address bar. I don’t know what the padlock means.
I have a guess — something about encryption — and the guess is close enough for most conversations. It is not close enough for any conversation I actually want to have.
I use the word crypto to mean two different things. One is math. The other is “money.” I might switch between them in the same sentence without noticing and I suspect I am not alone.
This is a tool for getting closer. Not a textbook. A set of pages you can cut into.
a picture of a lock
The padlock promises privacy of transit. It does not promise trustworthiness of destination. It is honest about one thing and quiet about another.
one key locks it, the same key unlocks it
A lock says: the thing inside is for one person, or two people, or everyone who holds a copy of the key. The promise is not privacy in the abstract. It is privacy from anyone who does not already hold the key.
Which raises the question every lock asks and never answers on its own — how did the key get to the other person in the first place?
The correct key pushes each pin to exactly the right height so the shear line runs clean. The plug rotates. The lock opens. A wrong key leaves pins crossing the shear line.
This is symmetric encryption. Two people share one secret. The math replaces the pins with numbers. The structure of the promise is identical.
A lock is rarely broken where you think. A lockpicker finds the hasp — the metal loop the lock passes through — and discovers it is aluminum, or welded poorly, or held to a wooden door with two short screws.
Symmetric encryption has the same property. The math is fine. What breaks is everything the math depends on: how you stored the key, how you sent it, who was watching when you typed it.
Before pin-tumblers there were wooden pin locks in ancient Egypt. Before those there were knots tied so that they could not be retied the same way. Before those there was trust — you handed the letter to someone you knew, and you hoped.
The mathematics that backs the padlock in your address bar is about seventy years old. The problem it solves is about five thousand.
anyone can drop a letter in; only one person can take letters out
The lock solved privacy. But it created a new problem: both people need the same key. How do you share a secret with someone you have never met?
If you could already communicate securely, you would not need the key. The lock’s promise depends on a step the lock itself cannot perform.
A mailbox has a slot anyone can reach and a locked compartment only the owner can open. The slot is the public key. The lock on the back is the private key.
The mathematical miracle: a one-way function. Easy in one direction, practically impossible in reverse.
The mailbox solves key-sharing. But it introduces a new problem: how do you know the mailbox belongs to the person you think it does? Anyone can put up a mailbox and claim to be someone else.
If the wrong person set up the box, every letter you send goes to them. This is the man-in-the-middle problem.
Public-key encryption did not exist before 1976. Whitfield Diffie and Martin Hellman published the concept, and it made the internet possible. Before that, every secure channel required a pre-shared secret.
The insight that you could communicate securely without ever sharing a secret was genuinely new under the sun.
a unique fingerprint of a letter
Encryption hides the message. But hiding is not the whole promise. You also need to know the message arrived unchanged.
Privacy and integrity are two different promises. The lock handles privacy. The wax seal handles integrity.
A hash function takes any amount of text and returns a fixed-length string. Change one character of the input and the hash changes completely.
Change one character. Watch the entire hash change.
A wax seal proves the letter has not been tampered with. But it does not prove who sent it. If you have never seen the sender’s seal before, a forged seal is indistinguishable from the real one.
Wax seals were used on correspondence from at least the Middle Ages. Clay seals are older — Mesopotamian cylinder seals from 3500 BCE rolled across wet clay to mark ownership.
The mathematical hash function formalizes the same promise: a unique mark that breaks if anything changes.
a private mark anyone can recognize
The wax seal proves a message has not been tampered with. But it does not prove who sent it. Anyone with wax and a flame can seal a letter.
Identity requires something unique to the sender — something no one else can produce, but anyone can verify.
A signet ring pressed into wax leaves a mark only this ring can make. The ring stays with the sender. The impression is visible to everyone.
A digital signature works the same way: the sender uses their private key to sign. Anyone with the corresponding public key can verify.
The ring proves who sealed it — if you already know whose ring it is. But if someone shows you a mark you have never seen, you have no way to connect it to a person.
Signet rings date to ancient Mesopotamia. Egyptian pharaohs used them as seals of authority. Medieval nobles pressed them into wax to prove authorship.
The digital signature formalizes this — mathematics replacing metal, but the gesture unchanged.
a bound book recording which marks belong to whom
The signet ring proves identity — if you already know the person. But the internet is a system of strangers. When you visit a website for the first time, you have never seen its mark before.
Someone has to vouch for the connection between the mark and the name.
A notary keeps a bound book. When someone presses their ring into wax in front of the notary, the notary records: on this day, this mark was made by this person.
A certificate authority does exactly this — binding public keys to identities. Your browser trusts a pre-installed list of authorities.
The entire system now depends on trusting the notaries. If a notary is corrupt or careless, every certificate they issued is suspect. And there are only a handful of notaries.
Notaries public date to ancient Rome. The medieval church maintained registries of authorized scribes. Every civilization has needed trusted third parties to verify identity.
The modern certificate authority system inherits this structure — and its weaknesses. Certificate authorities have been compromised. Fraudulent certificates issued. Every downstream trust relationship broke.
The places you arrive already confused, in your own words.
cipher is not an encryption explainer. It is a tool for learning to think in fractals about trust, information, and the systems people build to make communication reliable. Every concept is a physical object drawn in cross-section. Every cross-section raises the problem the next object solves. The spiral is the point.
cipher holds two voices at once. The Explainer is calm and precise. The Skeptic is willing to say this is where it gets weird. The Skeptic is what you find when you zoom in far enough.