Before I could write the Polyglottal Cipher, I spent ten years reading.
Not books. URLs. Query strings. The long encoded tails of affiliate links — ?utm_source=, base64 blobs, percent-encoded parameters, tracking IDs folded into tracking IDs. Millions of them. It was my job to read the exhaust of the internet and understand where money came from.
If you stare at encoded strings long enough, something happens to your eyes. You stop seeing noise. You start seeing structure. You learn that every string is a language, and every language is just an agreement about what symbols carry meaning.
That's the whole insight of the cipher, and it came from marketing, not math.
Conventional encryption marks itself. You encrypt a file and it turns into a wall of random bytes that announces something valuable is hidden here. The protection becomes the target. I'd watched the same mistake in SEO for years — the sites that screamed "optimize me" were the ones that got penalized. The ones that looked natural survived.
So I asked the cryptographic version of the question I'd been answering for a decade. What if the protected thing didn't look protected? What if encrypted data looked like a sentence?
The answer was already sitting in Unicode. 133,387 characters across 202 language traditions. Most of them never used in ordinary text. An enormous quiet space that everyone treats as a font problem and nobody treats as a channel.
You take the plaintext. You encrypt it properly first — ChaCha20-Poly1305, no shortcuts, real symmetric strength. Then instead of showing the payload as bytes, you encode it into glyphs drawn from real writing systems. The output reads as natural human text in whatever language you pick.
An adversary who intercepts it doesn't see ciphertext. They see writing. Without the key they can't even prove encryption happened. That's not just confidentiality. That's deniability — the message doesn't admit it's a message.
People assume you need a cryptography PhD to arrive here. I didn't have one. I had a decade of reading the world's encoded strings until I understood, in my hands, that the difference between text and ciphertext is only a matter of which agreement you're using to read.
The formal work is happening now, with people far more credentialed than me — KU Leuven's COSIC group, the team that includes one of the two designers of AES. The patent is filed with the EPO. The math will get the scrutiny it deserves, and it should.
But the idea didn't come from a lab. It came from years of being underestimated in a field nobody respects, doing a kind of reading nobody thought was reading.
I used to be a little ashamed of where I came from. Affiliate marketing is the internet's low-status trade. You don't put it on a conference badge next to cryptographers.
I'm done being ashamed of it. That trade taught me to see the thing the credentialed people walked past every day.
The cipher was never really about hiding. It was about the fact that reading and writing are the same act performed in opposite directions — and I'd been doing both, badly labeled, for my whole career.