[voiceinteraction] minutes October 23, 2024

https://www.w3.org/2024/10/23-voiceinteraction-minutes.html
And below as text


   [1]W3C

      [1] https://www.w3.org/

                             - DRAFT -
                           Voice Interaction

23 October 2024

   [2]Agenda. [3]IRC log.

      [2] https://lists.w3.org/Archives/Public/public-voiceinteraction/2024Oct/0006.html
      [3] https://www.w3.org/2024/10/23-voiceinteraction-irc

Attendees

   Present
          debbie, dirk, gerard

   Regrets
          -

   Chair
          debbie

   Scribe
          ddahl

Contents

    1. [4]follow-up from Dirk's presentation last time
    2. [5]assistant terminology
    3. [6]Summary of action items

Meeting minutes

   review changes to Architecture document

   [7]https://lists.w3.org/Archives/Public/
   public-voiceinteraction/2024Oct/0003.html

      [7] https://lists.w3.org/Archives/Public/public-voiceinteraction/2024Oct/0003.html

   for issue 52

   debbie: could just say a little more about the difference
   between traditional and GenAI systems in 3.1

   dirk: should we use another word besides "traditional"?

   we can't think of anything better

   dirk: redraws Figure 3

   dirk: Section 5.1 is just about traditional systems

   debbie: why is LLM in Figure 3?

   dirk: could gray out LLM to show that it's not involved. the
   same for Figure 4 (output)

   debbie: should say something about why the LLM boxes are grayed
   out

   dirk: agrees

   dirk: would like to add the word "agentic" because that's an
   established term

   debbie: I agree

   dirk: will add that

   debbie: should we say something about GenAI at the beginning
   (Section 1) about how these specifications apply to GenAI

   ACTION: Debbie to add that in Section 1

   dirk: should we add a use case for GenAI, like an avatar

   dirk: chat application with a GenAI system while you're driving

   ACTION: debbie to write up a sample use case

   ACTION: dirk to write up a sample use case

  follow-up from Dirk's presentation last time

   [8]https://docs.google.com/presentation/d/
   13ONyFA96wsWqnEV3jCTvHl6ze7rCspjZuhZ9bU66U50/edit?usp=sharing

      [8] https://docs.google.com/presentation/d/13ONyFA96wsWqnEV3jCTvHl6ze7rCspjZuhZ9bU66U50/edit?usp=sharing

   dirk: it seems like the taxonomy introduced last time was ok
   but data types were missing
   . should we mention those categories in the document?

   debbie: I don't think so, interesting, but going off on a
   tangent

   dirk: could try to contact author and see if he has anything to
   add

  assistant terminology

   debbie: list discussion had a lot of opinions

   dirk: this term, "personal" has two meanings, knows about you
   or nothing is shared

   debbie: I never thought it meant nothing is shared

   dirk: "changing personal assistant" would require a lot of
   changes

   dirk: can add something to say that personal doesn't mean
   nothing is shared

   dirk: will make a suggestion for text about why we call these
   personal assistants

   debbie: it doesn't affect anything in the architecture or specs

   dirk: we're not making any assumptions about where the
   component is running, and we may suggest not to share data
   unnecessarily

   debbie: that's kind of a best practice

   debbie: we can postpone reviewing GitHub issues until next time

Summary of action items

    1. [9]Debbie to add that in Section 1
    2. [10]debbie to write up a sample use case
    3. [11]dirk to write up a sample use case

Received on Wednesday, 23 October 2024 15:55:07 UTC