Syntopic

AI, knowledge graphs & the cyber humanities

A podcast from the two cofounders of Syntopic, Chad Nantais and Paul Salazar, on AI, knowledge graphs, provenance, and the digital and cyber humanities, and on what they mean for podcasters, researchers, media organizations, and academic institutions. The site is built the way Syntopic builds every archive: each claim on it is tied to the moment it was spoken.

1 episode · 1 hour · 65 claims · 29 entities

Three ways in

  • Listen in order. The episodes are each readable in full or playable from the page.
  • Follow an idea across episodes. Every concept, technology, field, and person the hosts discuss has its own page, gathering what was said about it: Concepts, Use Cases, Technologies, Fields, People.
  • Search it. Use the box at the top left. Search a name and you get the passages, not just the pages.

Concepts

  • Human in the Loop in 1 episode — The design principle that a person reviews and approves an automated system’s output.
  • Hallucination in 1 episode — An AI system’s confident output of things that are not true, here above all quotations nobody said.
  • Provenance in 1 episode — The documented history of where something came from, taken from the art and wine trades.
  • Augmented Intelligence in 1 episode — AI used to extend human judgment rather than replace it.
  • LLM Epoch in 1 episode — Chad’s name for the period after ChatGPT’s release in late 2022, when no text can be assumed to be wholly human-written.
  • Long-Form Conversation in 1 episode — Extended, unscripted dialogue of the kind long podcasts record.
  • Secondary Orality in 1 episode — Walter Ong’s term for the orality of electronic media (radio, television, and now podcasts), as distinct from the primary orality of cultures before writing.

Use cases

  • Podcasters in 1 episode — Independent and growing podcasters, the first audience the series addresses.
  • Cultural Heritage Institutions in 1 episode — Archives, libraries, and national institutions holding large collections of historical recordings.

Technologies

  • Gemini in 1 episode — Google’s family of large language models and the chatbot built on them.
  • Knowledge Graph in 1 episode — A structured map of entities and the relations between them, the form a Syntopic site takes: people, works, and concepts as pages, connected across episodes.
  • Large Language Model in 1 episode — The class of AI models, such as GPT, Gemini, and Claude, that generate text by predicting the next word.
  • Whisper in 1 episode — OpenAI’s speech-to-text model.

Fields

  • Cyber Humanities in 1 episode — A newer name for work at the meeting of computing and the humanities, close to the digital humanities, with an emphasis on protecting cultural memory.
  • Semantic Web in 1 episode — The project of publishing data on the web as linked, machine-readable meaning through ontologies, linked open data, and knowledge graphs.

People

  • Marshall McLuhan in 1 episode — The Canadian media theorist whose idea of media as extensions of the human runs under Syntopic’s case for spoken word.
  • Jacob Coxon in 1 episode — A former pretraining researcher at OpenAI and Anthropic who resigned from Anthropic in 2026 and warned publicly about the frontier labs’ race, including on The Daily Show with…
  • Jiddu Krishnamurti in 1 episode — The Indian philosopher whose talks survive as recordings and transcripts.
  • Jon Stewart in 1 episode — The American comedian and host of The Daily Show.
  • Mortimer Adler in 1 episode — The American philosopher and editor behind the Great Books of the Western World and its index, the Syntopicon.
  • Walter Ong in 1 episode — The American scholar of orality and literacy who coined secondary orality.

Most recent