From the ZX81 to CBOR-Web: How a Dyslexic Brain Learned to Eliminate Noise
In 1983, a 16-year-old kid discovers computing on a ZX81 borrowed from his brother. He’s dyslexic. His brain doesn’t process information like others. Every page read three times. Every concept deconstructed, rebuilt, digested differently. Not a disability—a filter.
That kid is Eddie Plot. Delivery driver, chauffeur, maintenance technician, auto mechanic, military helicopter mechanic, market vendor, IT professional. Hands-on in the morning, intellectual in the afternoon, creative in the evening, inventor at night. 20 careers. No prestigious degree. But a discipline forged by necessity: eliminate the noise, keep the signal.
The problem: 96% noise
At 59, father of three—Mathieu who protects forests and wildlife, Eloïse who takes care of people’s health, Vincent who ensures hospitals never run out of electricity—Eddie founds Deltopide and creates CBOR-Web.
The principle? Exactly what his dyslexic brain has been doing for 40 years. A standard website sends 50 KB of noise to an AI—menus, ads, scripts. The useful information? 2 KB. That’s 96% waste. Billions of requests per day. Terawatt-hours burned for nothing.
The solution: from noise to signal
CBOR-Web eliminates this noise at the source. A structured, signed, verifiable file. AI accesses the essentials directly. Less noise, fewer tokens, less energy. And traceable sources that make disinformation harder.
The energy urgency
American data centers consume 176 TWh per year. By 2028: up to 580 TWh. Household electricity bills have risen 42% since 2019. What if a standard born from a brain that spent its life filtering noise could reduce this footprint by 96% on the web layer?
Open source, by conviction
The protocol is open source. Because a father watching his children work for the common good cannot lock the solution behind a paywall.
CBOR-Web. From noise to signal. As always.
Discover CBOR-Web
Learn how binary formats are replacing HTML for AI agents—and why it matters for the future of the web.
Read the article →