Prismatic wants to build a social network around all of your interests | Hacker News Reader

Prismatic wants to build a social network around all of your interests

gigaom.com

17 pointsharper13 comments

I remember building the first gaming dashboard at facebook. The assumption is that things are better with friends (true) and that your friends can be a good filter for content (largely false). People are different, even your friends. I think Prismatic is really on to something here, a lot like the promise of hunch. A service that understands me (whiskey, zombies, startups, catan, nukes, ...) will eventually do a lot better than a friend ranked filter.

Exactly. I would much rather see the things I like and be notified when I'm digging the same things as my friends or possibly even friends of friends.

There is no better indicator of "Likes" than actual consumption. Consumption is a vote with dollars and time and carries a lot more meaning than clicking on a stupid little blue button.

What I find interesting about this idea is the way it crystalizes networks of interest. I think Twitter creates these networks informally, at least the way I use it, but having something more explicit offers new possibilities.

I have to add that most companies would have rested on their laurels with great UIs like the previous Prismatic apps. So, I applaud these guys for not settling on an already great tool, but taking risks to move forward.

I played around with Prismatic before but it just didn’t grab me and I found I didn’t use it that much.

This new version is a whole different animal. Not only is it much prettier (great design) but they seem to have seriously improved their relevance algorithms. I would be very interested to hear from their data team why the relevance is so much better now - anyone from Prismatic monitoring these comments?

Glad you're liking it! I'm one of the backend engineers who designed the system. We devoted a good chunk of time this year to improving our relevance algorithms, and I actually gave a talk on exactly this at Strangeloop in September. The video is at http://www.infoq.com/presentations/machine-learning.

Great presentation! Very fascinated by this world and Prismatic definitely appears to be at the forefront of achieving relevance. What I've been kind of baffled by are businesses like Taboola, Gravity and Outbrain. I understand their businesses are centered around placement of "related" or 'personalized' articles of boobs, bikinis, and booze just so you click so they get paid, but do you think behind the scenes they have relevance and personalization technology on par with Prismatic? They have a ton of money and people so you'd think they should be able to recommend genuinely relevant content, but publicly never appears so

I like Prismatic a lot and find it very useful and their relevance algos are top notch. I just don't understand why they cannot scrape and suggest high quality news sites like the NY Times. I mean if Google News can, why can't Prismatic as well?

Their feed was a lot more relevant last year. Also, it seems they're leaning towards iOS mobile app since there were lots of releases for it and none for Android. Pretty big bummer for me since I was expecting the Android one (though might not be so significant now that I found BaconReader for reddit)

Prismatic is the only app/website that claims to be relevant and actually finds relevant, fresh content. I don't know how, but it does work damn well.

Digging it. Seems pretty accurate so far.

I always, always seem to find something new and interesting when I open the app... bodes well.

They can start by building an Android app!

Looking awesome!

Reply on news.ycombinator.com