Encoding outperforms rote repetition. Producing it reliably, at scale, is the unsolved part. That's the Gorgo hypothesis.

Gorgo is an encoding pipeline. Language-agnostic. Domain-general by design.

Built as a system that bridges French to memory as successfully as it may anatomical terms, drug names, latin roots, Japanese, Dutch or Hindi.

  • It's built by encoding. The principle is old and well-established—Allan Paivio's dual coding, the isolation effect, the keyword method.

    What was never settled is whether you could produce encoding of high enough quality, reliably and at scale, to make the effect durable. That's the open problem.

  • The method improves retention when the cue is good enough to regenerate and gets practiced. The literature's own failure cases trace back to weak, non-regenerable cues. Gorgo is a bet on both: high quality, at scale.

  • It began with French, and a problem.

    After 10 years designing one of the world’s largest systems, the map by which 2B+ people navigate, I decided to begin a new adventure. One that wouldn’t be successful without learning spoken and written French. Familiar dread set in. Rote learning had never worked for my visually inclined, autistic brain. Streaks designed like children’s games were the antithesis of a ritual I’d do anything to avoid.

    I asked myself: if not this, then what? How do I actually remember things? Gorgo was born—a meeting of mnemonic practice and a challenging problem to solve.

Experiment Status

We are testing a single claim: That an automated pipeline generates keyword mnemonics of the same quality as those made by skilled human mnemonists, and higher quality than those made by typical humans, across structurally distinct domains.

Quality is operationalized through the dimensions the keyword-method literature has repeatedly tied to mnemonic effectiveness—acoustic link, interactiveness, and imageability/vividness (Atkinson & Raugh 1975; Campos et al. 2004; Beaton et al. 2005). This study targets one thing: whether the pipeline produces those dimensions at expert level. Durable retention is a separate, still-contested question—and it depends on cue quality and retrieval integration, which is precisely why this study targets cue quality first. It's the upstream variable the durability debate keeps landing on.