Introduction
How could capabilities for server-side and client-side AI processing, e.g., to obtain embedding vectors from discussion forum threads, posts, and selections of content, enhance the Fediverse?
With respect to client-side AI and Web browser APIs, one can observe:
Obtaining Contextually Similar Discussion Threads
If servers could obtain embedding vectors for discussion threads and posts, forums could provide participants with hyperlinks or excerpts to contextually similar discussion threads in side-panel widgets. These contextually similar discussion threads or posts might be from the same forum or from other connected forums. As considered, forums’ administrators could select which other forums to connect theirs to, in this way.
Looking Ahead
Once having obtained contextually similar discussion threads from the same forum or from other connected forums, systems could look ahead in these, providing forum participants with subsequent posts from these, including posts marked as answers to questions.
For a concrete example, a forum instance for a computer-programming course could present its participants in question-answer discussion threads with a side-panel widget where contextually similar discussions and answers from services resembling Stack Overflow could be presented to them.
Summarizations
Once having obtained contextually similar discussion threads from the same forum or from other connected forums, systems could summarize these and provide these contextual summarizations in the form of excerpts in side-panel widgets.
Recommender Systems
Recommender systems on servers could filter these contextually similar discussion thread and post alternatives for individuals or groups participating in discussions.
Searching from Selections of Content
Forum participants could open context menus on selections of content in forums’ discussion threads, obtaining client-side embedding vectors for these, to use their preferred search engine to find contextually similar discussion threads and posts from a wide set of related forums.
Organizational Memory, Organizational Learning, and Collective Intelligence
Providing forums’ participants with opportunities to view contextually similar past discussions could enhance and support organizational memory, organizational learning, and collective intelligence.
Conclusion
Some possibilities were presented involving how server-side and client-side AI, specifically with respect to obtaining embedding vectors, could enhance the Fediverse. Thank you!