Introduce new kind of talk asset for audio

[fcamblor]

There might be some needs where we could attach some Audio assets to talks for :

• Podcasts
• pre-talk AI generated summary (based on other assets such as slide deck)
• post-talk AI generated summary (based on talk transcript)

Devoxx BE is currently providing such assets which might easily be integrated on talk details screen, see https://devoxx.be/talk/crafting-intelligent-applications-with-quarkus-langchain4j/

Important side note: providing some small description alongside the audio asset is important, particularly for the disclaimer part "AI-generated (Experimental): may contain inaccuracies, please verify facts."