Agreed. My problem (and it may just be mine) is that my 100 or so friends on Facebook (mostly IRL friends) are not that active on Facebook groups. They will update their status and post photos but will not share content on groups. I am part of a number of groups around my interests (tech, photography etc) but tends to be me and maybe one or two others that contribute content. Very little discussion occurs. I wonder if Google+ will change that with the ability to add in non friends around an interest (ala twitter) but without broadcast to all.

What I would like is the circles feature to encompass this, so that you would have the equivalent of Twitter wide spread for instance a circle on Node-JS, a music circle which I share with friends as previously mentioned and another wider circle of friends which is more like facebook, a more general taste circle. I think this makes a lot of sense, and I want these groups to be explicit, I don't particularly want any groups to be "We thought you might be interested in Penguins seeing as you are in the Linux group"

I agree... this may hurt/help Google group making. Either people will add non-contributors to a group and get bored with others' lack of contribution, and by extention Google circles. (To combat this, it'd be great if Google helped curate or edit circles, like with listservs. Non-content posters would be flagged for me and I could opt into auto-removing non-contributors or just being notified that they are diluting my circle.) Or, it will result in better curated lists of people -- similar to twitter, where you really only want to follow someone (and up your following count) based on the quality of their content.

But wouldnt this enhance the bubble-ification? I like that on twitter I often see other sides of personality and new things from people.

Bubbles have nothing to do with this. You're talking about a list of people you chose to see updates from. This has nothing to do with algorithmic surfacing of content. It's your own fault if you make a circle called "Liberal Bros" and only ever visit that one.

That's not what he means. Hes talking about making a new friend at a mutual friends party, lets call him Joe. You look at Joe's profile, and all you see on Joe's profile are photos of him drinking.

Meanwhile, Joe has been sharing a ton of geocaching and hiking posts with his "hiking" circle, but you don't even see it since he didn't know that you liked hiking, he only knows you from a party. His alternate interests are hidden from you because you weren't in the right circle on his end, completely out of your control.

Isn't that what the Sparks component is for? (Connecting based on interests?)

I'm not on Google+ so I can't really say for sure, but I understood Sparks to be more like a global stream that you subscribed to. I didn't understand that Joe would write his hiking posts into the Hiking Spark unless he was sharing something he thought strangers would be interested in.

It still demonstrates the issue with circles based on what content your friends would be interested in. Circles seem to make perfect sense to me for security (you don't want to see your mom posting photos of you partying maybe), but for interest based valuations it seems like you pretty much always want it to be public (at least public on your profile page, even if you don't particularly want it to show up in other peoples 'feeds')

Oddly enough, at the end of Pariser's TED talk, he called out facebook and Google to allow users to control what gets through their 'filter' and what doesn't. This is definitely bubble-ification, though at first glance it's the user who's in control of it.

Isn't that sort of what Google is doing with Circles?

The Circles idea would be good if it auto generated people who should be included in Circles based on use patterns(mentioned here "It is a shame that there aren't any recommendations for people that ought to be grouped together automatically into a common Circle.")

It wouldn't be creepy if the interface picked up on behavior similar to how Gmail suggests people to cc on emails based on past email use patterns. People are ready for that.

It actually does that too! https://twitter.com/#!/albertlai/status/85954355496370176