Browse · People
Person · Latent Space wiki

Toran Bruce Richards

AutoGPT’s founder described using programming techniques from games to chain language-model predictions, and advocated open source as a way to broaden access to AI.

Toran Bruce Richards is the founder of AutoGPT. His contribution centers on chaining language-model outputs through traditional programming techniques—an approach he described as drawing especially on techniques used in games. He uses Torantulino on GitHub, where he also links the handle @SigGravitas.[2][3]

AutoGPT and open access

In a 2024 OpenUK interview, Richards explained that he created AutoGPT on the belief that a language model’s next-word prediction could serve as a simulacrum of a single “thought”. Programming could then chain those outputs into what he called a “mind”. This was his conceptual framing for the system.[3]

Richards presented open sourcing as a way to give individuals access to capabilities otherwise concentrated in large companies. In the same interview, he argued that an individual could have “just as much ability” to use the technology as a corporate giant, opening opportunities previously dependent on substantial financial resources.[3]

Richards also presented Auto-GPT’s roadmap at Agent Weekend in 2023.[1]

Earlier foundations: games development

Before creating AutoGPT, Richards studied games development, according to his retrospective OpenUK account. That background supplies concrete context for his later explanation of AutoGPT: he explicitly identified programming techniques used in games as a means of chaining language-model predictions.[3]

Timeline

  1. milestone

    Presented Auto-GPT’s roadmap at Agent Weekend
    Richards also presented Auto-GPT’s roadmap at Agent Weekend in 2023.[1]

  2. milestone

    Explained AutoGPT’s conceptual framing in an OpenUK interview
    In a 2024 OpenUK interview, Richards explained that he created AutoGPT on the belief that a language model’s next-word prediction could serve as a simulacrum of a single “thought”. Programming could then chain those outputs into what he called a “mind”. This was his conceptual framing for the system.[3]

From the Latent Space archive

Original posts and separately reviewed explanations.

Sources

  1. 1
  2. 2
  3. 3

Model source and teaching reviews passed. Human editorial review pending. This profile describes the sources as of 2026-09-20.

Latent Space

Join Latent Space.

Sign in with Google, or get the newsletter by email.

Loading Google sign-in…

or subscribe by email

Free. Unsubscribe any time.