My social medium of choice is, despite my many and loud gripes about it, Google+. It's not even nearly awesome (none of the major options are), but, for a while at least, it offered two compelling features.
The first is Notifications. Action on your own posts, posts with which you've interacted, and (selected) people mentioning you, will show up in your Notifications list (which ought be a stream in its own right). This is, if you think about it, one of the highest possible signifiers of relevance: something you've already indicated an interest in in which others are interacting, often directly with you.
The other was Search. After a couple of years, even the vast wasteland of G+ had accumulated a catalog of interesting posts, and via Search those could be found. Search was always a bastard child, and it's considerably worse now than it had been for a long time, but given a keyword signifier, it would, sometimes, return relevant content.
I actually wrote a really embarassing mash note to G+ management -- Vic and the project's then product manager -- highlighting this.
The response of Google was, of course, to kill both features, or attempt to do so valiently.
The problem with the algorithm is that at best it can guess or assume intent. But it never fucking bothers to ask you directly, or to allow you to indicate your preferences.
I've likened this to a good salesperson. These exist, and I've on occasion, despite my anti-commercial tendencies, encountered and appreciated them. The good one will size you up as you enter a store, compose a good idea of what you're looking for, consider elements (say size or measurements for clothing), and match these with stock. Then they'll ask you what you're looking for. They'll return with a selection of items and, often, a few others. They'll read your responses, even subtle, on what's been offered and take a "no" as a "no".
That is the one feature that is missing from virtually all current electronic information interfaces. I wanted, an hour or so back, to tell Amazon to show me only books published matching a subject from 1990-1999. Actually, filtering by cover colour would also have been useful. No can do.
For media streams, the thing I most want to do is filter crap. Again, G+ offers no such capabilities. There are websites I never want to see (say, WSJ at HN), topics or people I don't care to hear about, etc. The best option I've got is to uncircle or block crap, which actually works quite well. But finer controls would be hugely useful.
Twitter's experiment seems to be headed the same way. And it's not going to work.
My social medium of choice is, despite my many and loud gripes about it, Google+. It's not even nearly awesome (none of the major options are), but, for a while at least, it offered two compelling features.
The first is Notifications. Action on your own posts, posts with which you've interacted, and (selected) people mentioning you, will show up in your Notifications list (which ought be a stream in its own right). This is, if you think about it, one of the highest possible signifiers of relevance: something you've already indicated an interest in in which others are interacting, often directly with you.
The other was Search. After a couple of years, even the vast wasteland of G+ had accumulated a catalog of interesting posts, and via Search those could be found. Search was always a bastard child, and it's considerably worse now than it had been for a long time, but given a keyword signifier, it would, sometimes, return relevant content.
I actually wrote a really embarassing mash note to G+ management -- Vic and the project's then product manager -- highlighting this.
The response of Google was, of course, to kill both features, or attempt to do so valiently.
The problem with the algorithm is that at best it can guess or assume intent. But it never fucking bothers to ask you directly, or to allow you to indicate your preferences.
I've likened this to a good salesperson. These exist, and I've on occasion, despite my anti-commercial tendencies, encountered and appreciated them. The good one will size you up as you enter a store, compose a good idea of what you're looking for, consider elements (say size or measurements for clothing), and match these with stock. Then they'll ask you what you're looking for. They'll return with a selection of items and, often, a few others. They'll read your responses, even subtle, on what's been offered and take a "no" as a "no".
That is the one feature that is missing from virtually all current electronic information interfaces. I wanted, an hour or so back, to tell Amazon to show me only books published matching a subject from 1990-1999. Actually, filtering by cover colour would also have been useful. No can do.
For media streams, the thing I most want to do is filter crap. Again, G+ offers no such capabilities. There are websites I never want to see (say, WSJ at HN), topics or people I don't care to hear about, etc. The best option I've got is to uncircle or block crap, which actually works quite well. But finer controls would be hugely useful.
Twitter's experiment seems to be headed the same way. And it's not going to work.
Because the algorithm is an idiot.