Mathieu

@mathieuloutre
75 Followers
124 Following
43 Posts
Type and interaction bricoleur
Websitehttp://www.mathieutriay.com

🏁 🎉📾 Jules Vernacular reloaded

âžĄïž https://vernacular.fr

C’est avec joie que je vous annonce le retour de Jules Vernacular, mon herbier typographique dĂ©diĂ© aux Ă©critures artisanales. Le revoilĂ , transformĂ©, aprĂšs plusieurs annĂ©es de travail, de doutes et d’obstination.

Ce site est pensé comme un véritable outil : pour explorer, trier, voyager à travers les signes vernaculaires et les mots de la rue.

Merci Ă  @greg, avec qui j’ai engagĂ© cette refonte et qui en a conçu l’architecture.
Merci Ă  @timotheegoguely, qui en a assurĂ© la mise en Ɠuvre et menĂ© le chantier jusqu’à son terme.👌

Avec le soutien des Éd. @bp1t
SpĂ©ciale dĂ©dicace Ă  @super2000 ❀ pour son aide depuis le dĂ©but.‹
Et big up aux auteur·ices de toutes ces belles lettres.




“The future of AI agents is not just a question of what they can do, but of what kind of human experience we want to preserve." My thoughts on AI agents and the finite nature of agency for @BBCRD — https://www.bbc.co.uk/rd/articles/2025-05-ai-agents-challenges-summary
AI agents: Exploring the potential and the problems

AI agents could save time and effort, but they also raise many big questions - including ones of trust, control, privacy, and sustainability.

Just had two articles about Orbit published on the @BBCRD blog.

The first one about the technical details of how we organised the music so it could be discovered organically: https://www.bbc.co.uk/rd/articles/2025-03-orbit-music-discovery-organise-analysis/

And the second one about the philosophy behind the UI and UX design to give users agency in their discovery process: https://www.bbc.co.uk/rd/articles/2025-03-orbit-music-discovery-design/

Organising music for organic discovery

Details on how we organised thousands of tracks to build a music discovery service that lets you find your next favourite song from BBC Introducing acts.

Orbit is our pilot for a music discovery service that lets you find your next favourite song by exploring recently played tracks from BBC Introducing!

Find out how machine learning helped us to organise *thousands* of tracks to let you find your next favourite song:
https://www.bbc.co.uk/rd/articles/2025-03-orbit-music-discovery-organise-analysis/

Then read about the design challenges of creating a way to explore music genres and styles... our experiment is called Orbit - and you can read about it and try it out below:
https://www.bbc.co.uk/rd/articles/2025-03-orbit-music-discovery-design/

Organising music for organic discovery

Details on how we organised thousands of tracks to build a music discovery service that lets you find your next favourite song from BBC Introducing acts.

Dynamicland intro

🎧 6/6 I’m excited to hear what other music fans think. Personally, while developing Orbit, I've been introduced to dozens of new artists, and it's become my go-to for fresh tunes when “the algorithm” feels stale.
5/6 ⏳ You can sample any number of tracks, but you can only reveal 5 tracks a day. This isn’t to limit discovery but to value your time and make each track count. Find what you like, then listen on your preferred streaming service and get on with your day.
4/6 📀 Instead of getting recommendations, you can make decisions based on what you hear and what you fancy right now, charting your way across the stars and revealing the tracks you want to save.
3/6 đŸŽ¶ You start with 10 samples, each representing a different space in the collection. There's no metadata visible, so choose what's closest to what you want and you’ll get tracks in the same ballpark, each a potential doorway to a different sub-genre.
2/6 🌌 BBC Introducing plays hundreds of new tracks from unsigned artists each week. We take last month’s batch and organise them by affinity according to 8 indicators based on the track itself — it doesn’t matter what you or others have listened to or skipped.