Querido diario, S.A.
The title's dream-job AI gets designed out loud: hiring is a cell membrane guarded by HR, interviews test latent knowledge he himself once faked with ChatGPT open, and the fix is a diary that learns what you actually want and matches you before you polish a CV.
// trace: where this idea came from
- ↳ video diary @ 26:23 (la célula y su portero)
- ↳ video diary @ 30:04 (el matcher de embeddings)
- ↳ video diary @ 41:57 (Julia: la sesión de psicólogo)
- ↳ Seed 102-1: Quién indica (la evidencia de que el sistema real es otro)
The indictment first, with himself as exhibit A: interviews test knowledge held latent in the head ▸ 20:44, and two years ago he passed one with ChatGPT open in another window, bluffing through BPMN questions while charm covered the typing ▸ 21:27, then learned the actual job after entering, like everyone does. His conclusion: evaluate whether the person can deliver the thing in the time, “ya no me interesa nada más” ▸ 25:01, since with AI “cualquier persona que tenga esa habilidad blanda de aprender rápido puede hacer cualquier trabajo de esos” ▸ 24:30. The system diagram is biology: the company is a cell, the hiring manager its membrane gatekeeper ▸ 26:23, and the pore is tiny, misshapen, and guarded by someone from another department.
The design session then builds the title’s machine. Your side: not a CV, most people rightly can’t write one, his first took a week ▸ 43:07, but a diary. “Querido diario, hoy me sentí mal en este trabajo…”, and the AI accumulates, distills your values and strengths the way REAPRA’s coaching does, then acts when the pattern is ripe: here’s the company that matches what you keep wishing for ▸ 30:04, offered proactively, even noticing you’re bored before you resign ▸ 30:57. Julia names why a form can’t do this: it’s a psychologist’s first session, you don’t know why you’re there, “es con las preguntas exactas que te sale” ▸ 41:57.
el CV es un formulario; el deseo es un diario →
The company side is symmetric: postings are generic on purpose, specificity kills discoverability, so precise needs and precise people “nunca match” ▸ 44:57; with a matcher, the senior who actually owns the problem writes the literal description, REAPRA’s Founder-in-Residence being the rare existing example, and the ideal endpoint gets said plainly: “cada persona podría terminar en el lugar donde realmente genere el mayor valor” ▸ 46:17. A membrane with zero friction, no gatekeeper, just recognition…
// continued in
no entry has continued this idea yet: the arc is still open