Daniel Lemire's blog

, 5 min read

We need more than spam filters: we need bona fide assistants!

6 thoughts on “We need more than spam filters: we need bona fide assistants!”

  1. Neil Conway says:

    There has been work on related topics in the AI community — for example, by Eric Horvitz at MSR. e.g.,

    https://research.microsoft.com/en-us/news/features/busybody.aspx

    https://research.microsoft.com/en-us/um/people/horvitz/lumiere.HTM

  2. @Neil Conway

    Though I was not aware specifically of MSR work on this topic, I do know that there has been work related to this problem… I have a friend who has worked on helping people write better emails through AI.

    My specific point is that it may not receive as much attention as it should. Not just from the AI community, but also from engineers.

  3. Guy Leblanc says:

    Au lieu d’étudier l’utilisation d’une intelligence artificielle pour les échanges avec les étudiants et les collaborateurs, je tenterais probablement d’utiliser un wiki. Ils permettent les communications asynchrones comme les courriels. Cependant, il est possible d’orienter les échanges vers la construction d’une base de connaissances utile pouvant aider à minimiser les communications redondantes. L’idéal, je crois, n’est pas juste de minimiser le travail de votre point de vue, mais bien de celui de tous les collaborateurs. Une telle approche pourrait permettre de souvent minimiser les délais pour les gens qui cherchent à obtenir une information de votre part. Évidemment, il y a plusieurs obstacles à une telle approche. Une analyse comparative des deux approches serait sûrement un beau sujet de recherche.

  4. @Guy Leblanc

    Maintaining a collaborative wiki is hard work.

  5. Mihai Christodorescu says:

    @Guy Leblanc

    The dominant characteristic of all wikis I have ever seen is that they are out of date. Then you end up emailing the maintainer/author to ask your question anyway.

  6. Rivest says:

    I agree it would be helpful. One thing that I found out for myself is that I tend to write more in the morning. So I avoid responding to emails in the morning. I rather use that time to draft papers or other things. If I answer the same email at the end of the day (say a bit before 17h), I will spend much much less time on it then at 8h in the morning. In short, WHEN I deal with an email impacts the time I spend on it.