Hey friends of Micro.blog.
I am a contractor working mostly on M.b curation for about 5 hours a week. So if you perceive I’m not doing as much as Jean did, you’re right!
I also haven’t been around for about a week because of a big family medical emergency. Everyone’s okay now.
💗
@KimberlyHirsh Glad everyone's okay. Replying here so others see it too... I tried to convey to folks when you joined that it wasn't an attempt to exactly replace what Jean was doing, but that's easy to miss. It's still a big help and also a chance to try new things in the future.
@KimberlyHirsh Good to hear that everyone is doing good, that’s the main thing. Also important information you’re sharing, I wish I could boost that.
@KimberlyHirsh I’m glad to hear that family are OK.
And thank you for clarifying your Micro.blog position.
@KimberlyHirsh Glad to hear everyone's ok. More clarification on your role from Micro.blog may help and some way of adding an "out of office" on the Discover tab will keep people from being too impatient 😊
@KimberlyHirsh I'm glad everyone's OK again after the emergency! And useful, as well, to know how much time you have to curate the Discover page.
@KimberlyHirsh I am so clueless I did not realized that discover tab is not algo-generated. But I assume behind the curtains there's a recommendation engine that you leverage?
@ronkjeffries There is not. Whoever is curating reads every post and manually selects what to feature.
@KimberlyHirsh This is an obvious opportunity for a little benign machine learning assistance. Just sayin'☕
@ronkjeffries I'll leave it to @Manton to address this further, but right now the process isn't burdensome, so I think sticking with human learning instead of machine learning will continue to be Micro.blog's M.O. for the foreseeable future. Maintaining Discover usually requires less than an hour a day.
@ronkjeffries I agree with what @KimberlyHirsh said, we like the human side of this as long as it's manageable. But I do think machine learning can be a help on the admin side, helping a human quickly find posts that could be featured or posts that should be flagged for community guidelines violations.
@manton I agree 1000 percent. you are so busy, have you considered hiring a bright creative intern who is interested in machine learning, then let them run experiments? MLs hould offer major upside for the human review process.
