Intelligence trainer vs third party clients?

Hey! So I’m not sure if I’m doing something wrong, or NewsBlur is.

I have a couple feeds that have content I really enjoy, and some complete garbage intermixed, which happens to use keywords in the title. Great! To set the stage, I have several items unread in the feed, I selected one, called Intelligence Trainer and set up a dislike. The matching items immediately disappeared.

However, the mobile clients I use see something different. NewsBlur’s own client sees one of the items (or did, I marked it as read and since I filter for Unread, I don’t see it anymore), and both Reeder and Fiery Feeds still see all of the items visible and marked as Unread.

I’m not really clear what should be happening, I have the sense that these clients may not support hiding items, but could NewsBlur at least mark them as read? – I tend to live out of the Unread view, so a read item is effectively gone for me, and even if not, at least it won’t be begging for attention.

Also, any idea why NewsBlur’s own iOS client still showed at least one of them?

Hi Dave,

Not sure what’s going on for you, but I can confirm that modifications made with the intelligence trainer are reflected in third-party clients for me.

I use Reeder, but train my feeds with the official client.

Yeah, I’m starting to wonder if the problem is client related, I’ve had Fiery Feeds mark something unread that I absolutely marked read on another client.

If I train something, it hides the older matches, but does it also mark them read in Reeder?

I’m already leaning toward going back to Reeder full time.

Jamie is correct, hiding stories in NewsBlur should hid them in other clients. I made a change a few years ago to correct for that. Some clients are able to show hidden stories, but by default they are not.

As far as I know, Fiery and Reeder do not support hidden articles directly, so the expected behaviour when I go through the trainer is that those articles should disappear (or at least be read) on the client?

Just to confirm, this will apply to unread articles already retrieved? Or does this only apply to newly posted articles (which does seem to work fine)

I’m unclear if my problem is my expectations, or just that some client is doing something weird.

I’ve just done a quick test. To me it looks like it applies to new articles only - at least on third party clients.