No-one actually writes it up, so you might average 150-200% more text than a great article - on the other hand some bias will be balanced out.
I'm not sure machine learning is (will ever be?) a good fit to fill in the role of journalist. But I think hobbyist/specialists dedicating a bit of time to commenting paired with a handful of people writing up the emerging concensus could be at least as good as much current mainstream media. I think that's true whether the experts knows the ins and outs of circuit design or schoolboard and neighbourhoods in a corner of Afghanistan - and who's gotten monetary support to build schools or something.
Exactly. The investigation, the debate and the plurality of viewpoints are easier to find in a reddit comment thread than a newspaper.