Hacker News new | past | comments | ask | show | jobs | submit login

This has been my experience across a number of platforms other than Twitter (I've not used it in any significant fashion).

On Mastodon (and previously on Google+), my strong advice is to use lists, and to essentially treat those as priority rather than topical classifications.

Usually that ends up being three tranches, high, medium, and low interest, where the high-interest list is about 10--20 active profiles (with time there may be inactives which accrue, though my tier-I list remains about 40--50 profiles). The 2nd and 3rd tiers are less selective, generally if someone doesn't seem to rate tier-I I'll bump them to II, and if they're getting too annoying there, to III.

Mastodon has the option to limit boosts (retoots) and replies by profile, so toggle those if someone's a good primary source but is overly profligate in their amplification.

More on lists generally: <https://toot.cat/@dredmorbius/109305354126509106>

I rarely view either my own "Home" stream (all profiles I follow) or the local instance / global federated streams.

Fuckwits get blocked: <https://toot.cat/@dredmorbius/104371585950783019> It's AMAZEBALLS how much s/n improves when n is explicitly filtered.

On post/response type platforms (Google+, Diaspora* (yes, it's still around), good hosts are key, and what you're looking for is someone who posts interesting prompts and hauls out the trash in terms of low-quality comments. (HN's moderation team does a generally excellent-job-at-scale of that here ... present thread notwithstanding.)

Highly-voluable posters are almost never worth following directly --- what they lack is a self-editor, and you'll usually catch their interesting stuff through boosts or the equivalent. There's phenomenon I've noticed since the days of print newspapers where the highly-curated selected syndication of content gives a tremendously distorted view of the representative quality of a publication: when you're reading at the far end of a highly-selective filter, the crap and cruft has been cut and blocked. Prolific posters may achieve occasional hits and high notes, but the wheat::chaff ratio is often very, very low.

Similarly, there are profiles which post rarely but almost always with high salience. In a strictly streams-based view, these tend to get lost.

Another useful tool is directed search. Google+ actually had useful search at several points in its evolution (though it rather famously launched without any). HN's Algolia tool is a key value point to this site, and I mine both my own and others' content frequently, as well as search either articles or concepts as they appear or occur to me, with varying levels of success but at least the possibility is there. Occasional gems to appear. The Fediverse has taken a generally adversarial approach to comprehensive search (largely a futile effort as Alex Stamos has pointed out, and I strongly agree), though there are hashtags which can be searched, pinned, and/or followed, and some limited full-text search (your own toots, followers / local instance) which is quite useful. Platforms lacking effective search (Reddit, for comments, Diaspora* at all, Ello (back when it was more alive) in any useful sense) are severely handicapped. Yes, the feature can be abused, but it's also highly useful.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: