Episode Details

Voice cloning has gotten good enough that hearing yourself answer back is no longer science fiction. This episode digs into the small but revealing collection of documented experiments where people actually conversed with clones of themselves — from Evan Ratliff's Shell Game, where two AI Evans called each other on the phone, to a developer who fine-tuned a model on a decade of his own forum posts, to coaching prototypes like Inner Voice that reflect your goals back in your own voice. The recurring discovery: a familiar voice is not a familiar mind. The clones are fluent, eager to please, and remarkably generic — world-class bullshitters, as Ratliff put it. We explore why "enjoyable" and "convincing" are completely different tests, what happens when your training data is missing the questions your answers were replying to, and why two clones talking to each other is where the illusion really falls apart. Key sources: • Transom, Finding Their Voice by Evan Ratliff, 2024-12-10 — https://transom.org/2024/evan-ratliff/ • Shell Game newsletter, Episode 3, 2024-07-23 — https://www.shellgame.co/p/episode-3-conversations-with-myself • StationLM analysis of Shell Game Ep. 3, 2024-07-23 (with transcript quotes) — https://stationlm.com/analysis/2783a5e2-4e73-439a-b01e-8ba61a112ca4 Episode 6197 — myweirdprompts.com/n/6197 (permanent short link: myweirdprompts.com/353125)