Post by @samsethi@podcastindex.social

Sam Sethi :pc2red: ⁂

@samsethi@podcastindex.social 6 days ago

On next weeks Podnews Weekly our special guest will be @ryanwi from Car Curios.

Ryan has proposed a new podcast annotation format which I liken to Apple’s proprietary ‘Timed Links.’

Should Podcasting 2.0 adopt this format?

@adam @dave @james @js @nathan @misener

https://www.podcastannotation.org

Podcast Annotation Format

Timestamped entity and topic references for podcast audio, anchored to the seconds where they are actually discussed.

Nathan Gathright

@nathan@xoxo.zone 6 days ago

@samsethi @ryanwi @adam @dave @james @js @misener I love structured data as much as the next nerd, and wish this idea the best of luck. But I’m convinced that using LLMs to derive data from transcripts will remain the purview of players not feeds.

https://xoxo.zone/@nathan/116613844156212080

John Spurlock

@js@podcastindex.social 6 days ago

@nathan @samsethi @ryanwi @adam @dave @james @misener

totally agree - from what I've seen in feeds, podcast publishers should focus on the hard part (audio + video) and leave metadata extraction to where it can be done best, the apps/crawlers

Adam Curry :pci: :pc2blue:

@adam@podcastindex.social 5 days ago

@js @nathan @samsethi @ryanwi @dave @james @misener

Whisper and ffmpeg are very powerful for users, even abstracted by a UI

Ryan Williams

@ryanwi@hachyderm.io 5 days ago

@adam @js @nathan @samsethi @dave @james @misener

Separate question. Do you think chapters are still mostly hand-curated today? I've had decent success generating them with an LLM on a limited set, but wasn't sure how most podcasters/publishers actually do it.

~8 more replies

James Cridland

@james@bne.social 5 days ago

@ryanwi @adam @js @nathan @samsethi @dave @misener Some podcast hosts will make them using an LLM. So does Apple Podcasts for playback. I make mine by hand because it’s one keystroke while editing.

(Going back to the original question - I believe that super-chapters are the way here. That chapters json file could include a lot more than it does. I tried to connect you to @theDanielJLewis earlier.)

Adam Curry :pci: :pc2blue:

@adam@podcastindex.social 5 days ago

@james @ryanwi @js @nathan @samsethi @dave @misener @theDanielJLewis

I want my podcast app to make other decisions about chapters or other relevant content and contextual data presentation for me. Based mainly on a transcript.

Once an app can use on-device "AI" the possibilities for a customized listening app are limitess

Adam Curry :pci: :pc2blue:

@adam@podcastindex.social 6 days ago

@nathan @samsethi @ryanwi @dave @james @js @misener

I tend to agree. Podcast apps will be able to take advantage of models om-device to build custom data environments based on user preference

John Spurlock

@js@podcastindex.social 6 days ago

@samsethi @ryanwi @adam @dave @james @nathan @misener

site looks ai-generated

it would be great if podcast publishers delineated mentioned entities using stable identifiers (useful across shows/episodes) with something agreed upon like wikidata ids: https://www.wikidata.org/wiki/Wikidata:Identifiers

Wikidata:Identifiers - Wikidata