The Seven Needs of Real-Time Curators

I keep hearing people throw around the word “curation” at various conferences, most recently at SXSW. The thing is most of the time when I dig into what they are saying they usually have no clue about what curation really is or how it could be applied to the real-time world.

So, over the past few months I’ve been talking to tons of entrepreneurs about the tools that curators actually need and I’ve identified seven things. First, who does curation? Bloggers, of course, but blogging is curation for Web 1.0. Look at this post here, I can link to Tweets, and point out good ones, right? That’s curation. Or I can order my links in a particular order. That’s curation. Or I can add my thoughts to those links, just like Techcrunch or VentureBeat do. That’s curation. Or I can do a video like Leo Laporte does and talk about those links. That’s curation. Or I can forward those links to you via email. That’s curation. The editor who sits in a big building at New York Times or your local newspaper that chooses what content you’ll see in your newspaper is a curator. So is the page designer who decides what story is at the top of the page.

But NONE of the real time tools/systems like Google Buzz, Facebook, Twitter, YouTube, Flickr, give curators the tools that they need to do their work efficiently. That’s why I’m writing this post, to try to get the industry to see that there’s an unmet need that — if they were met — would mean all sorts of things from better scrapbooks for family photos and events to better news systems like what CNN or Huffington Post are trying to build on the Web. More on that after I get through the seven things.

As you read these things they were ordered (curated) in this order for a reason. If you give me #7 without giving me #1 first your tool will suck and you won’t be used by curators. If you give me #1 without #7, you’ll be way ahead of some tool that gives me #7 only.

This is a guide for how we can build “info molecules” that have a lot more value than the atomic world we live in now. First, what are info atoms? A tweet is an atom. A photo on Flickr is an atom. A conversation item on Google Buzz is an atom. A Facebook status message is an atom. A YouTube video is an atom.

Thousands of these atoms flow across our screens in tools like Seesmic, Google Reader, Tweetdeck, Tweetie, Simply Tweet, Twitroid, etc.

A curator is an information chemist. He or she mixes atoms together in a way to build an info-molecule. Then adds value to that molecule.

So, what are the seven needs of real time curators?

1. Real-time curators need to bundle. We need to be able to bundle certain tweets together. What do I mean by that? Well, let’s say a news event, like an earthquake, happens right now while I’m writing this post. Which are the best 10 tweets that describe that event? Can we bundle those together easily? Bloggers can bundle, but making Tweets look like Tweets is actually pretty difficult for normal people and even for geeks like me. Gotta take a screen shot of the tweet, upload that, then build an image tag in WordPress, then link that image up to the original tweet’s permalink. Whew. What a lot of work for something that should be simple. This could look like tagging, but calling it tagging is pretty limiting because tags won’t get you to full curation. One question: why can we bundle Flickr photos together by applying a tag to them, but we can’t bundle Tweets together by tagging Tweets? For instance, here’s two photos I shot at Techcrunch’s offices showing their new TV team. How did I bundle those together? Simply by tagging them with “Techcrunch TV” tag. Now, what if I could bundle in Tweets about Techcrunch TV? How about a YouTube video? How about other people’s Flickr photos? How about photos on other services like Smugmug or Picasa? How about Google Buzz items? Now you’re starting to understand why we need bundling cross-platform so we can start pulling valuable atoms out of the real-time streams.

2. Real-time curators need to reorder things. Look at just those two photos. One is more important than the other. Now, imagine a bundle with dozens, hundreds, or even thousands of items. Why can’t curators put the most important ones at the top of the bundle, just like the New York Times front page editor puts the most important news at the top of the page? Or, even better, why can’t we organize them into sub bundles? During an earthquake, like the one in Haiti, some things happened on first day, other things happened on second day, etc. Why are they all in one flat stream? Or, look at Apple’s iPad launch. Some things are about the specs. Some things are about the people involved. Some things are about apps. Some things are about accessories. Why can’t we organize them all into sub bundles? All curated in order of importance?

3. Real-time curators need to distribute bundles. Let’s say I put together a report for my bosses at Rackspace about what is happening at YCombinator (they just had a launch this week of a new crop of companies). Let’s say I built a bundle of not just the Techcrunch article I just linked to, but the Tweets from the event as well as the reports from other tech journalists like those who work at GigaOm, who also had a report on that event. Now we need to distribute that bundle. Of course we’ll Tweet it. But that means a headline of less than 140 characters that must include a link to the permalink of the bundle. But what about Facebook? That can include a thumbnail. Google Buzz? That lets you upload items with longer headlines and multiple pictures. What about emailing this bundle around the way Chris Brogan emails his blog posts. Why can’t a curation tool be smart about distributing bundles and let you see and manipulate previews of how that bundle will distribute itself to the various places you need your bundles to go to get the right audience.

4. Real-time curators need to editorialize. So, now we have a bundle of Tweets, YouTube videos, Flickr photos, Google Buzz items, Facebook status messages, et al. We’ve seen a new pattern in the world and now we want to explain our view of that pattern. For instance, I was at the YCombinator event this week. What if I wanted to add my two cents into the patterns other people saw? I might want to blog like here. Or add a video of my own. Or a Cinchcast (audio recordings done on my iPhone). Or add a bunch of photos I shot, like this one of Paul Graham mentoring his startups at that event with what they did wrong and right. But why did I just need to click “img” and copy and paste a URL to do that? A curation tool would let me drag and drop on my new iPad that I’ll have next weekend.

5. Real-time curators need to update their bundles. When the Haiti earthquake happened, the news story changed over time. We had more information and many many more Tweets to bundle in, not to mention that the mainstream press started flowing stories into RSS and Twitter. If you can’t update a bundle then it will greatly limit the ability for us to communicate. Blogs are pretty bad at this. If I come back in two hours and update this post you probably won’t see the update. In fact, not only can I update this post, but everyone who leaves a comment underneath is really updating it too. Yet early readers won’t see the later comments. They are missing part of the story. Of course, once you update you need to redistribute. IE, let your Twitter and Facebook and Google Reader friends know that the story has changed and there is important new information on the bundle that you need to see.

6. Real-time curators need to add participation widgets. On some bundles you might want to ask your audience to take a poll. Some might want to add comments. Not everyone will. Seth Godin doesn’t have comments on his blog. Other bloggers might want to leave comments open for a few hours or a few days. Even here I’ve made it so you can only comment for 30 days on my blog posts. Why? Because of spammers and other bad actors. I can see a TON of widgets that would be available to get participation on widgets. These would be a great way for these systems to monetize, too. Would you pay $1 to add a poll to your bundle? I would.

7. Real-time curators need to track their audience. Look at this blog post. It has a TweetMeme button on it. That shows you how often this item has been retweeted. I would add such a button to every bundle I do. I’d also add Google Analytics and a few other things that would track where you’re coming from, what kind of engagement my items are getting, and even, how relevant you are based on your own participation in the system. Don’t think that’s already happening? Look at the curation system Spigit built for large enterprises. I met with them yesterday and their system does just that and is getting used by many of the world’s biggest companies like Wallmart and Starbucks.

Does such a curation system exist today? Yes, blogs, but blogs are HORRID for tracking this real time world. Just this post took me 30 minutes to bang out and that was after I had it in my head and I wrote it very quickly. Imagine I was talking about a real time event. The news is already 30 minutes old. We need a new system for real-time curation of what’s happening on my Twitter stream.

It’s interesting that no one has gotten close to even giving us the most basic curation tools. Why is that?

Why are companies ignoring our needs? In talking with CEOs at companies in the real-time space I’ve identified a few reasons:

1. Building-cross-platform tools is difficult. Each real-time feed has different APIs and isn’t set up to interoperate with other real-time systems. Twitter has no API to share its feeds with Flickr. Flickr’s tags don’t have any idea what YouTube’s tags are. WordPress is blind to all of it. Etc Etc.
2. Fear of platform vendors. No one builds these kinds of features because they are scared that Facebook or Google will build these kinds of APIs and kill their businesses. Not unfounded, either. Tweetdeck built lists into its product and then Twitter came along and added lists in a way that was far more useful than the ones Tweetdeck built. So, companies like Tweetdeck and Seesmic choose to work on things that Twitter will be unlikely to do.
3. Assumption that these features are only going to be used by “weirdos or professionals or both.” I hear this all the time “oh, Scoble, you need these features, but what about normal people.”

The first two I can’t do much about. I agree that these are features that would be best built in at a platform level and have told many of the players to do that. But the third is provably false if entrepreneurs would do some customer research (shocking, but many San Francisco area social networking companies do very little real customer research, which explains why they so often screw up around privacy and fail to find new features that dramatically improve our lives).

Let’s consider the mother who has a 1-year-old son. She invites 30 of her friends to a birthday party for her son. They take videos, do Foursquare checkins, one or two might blog about the party for their mommy blogs. Many take photos, but some of those photos end up on Facebook. Some on Flickr. Some on SmugMug. Some on Picasa. Lots of them Tweet about the event, or Facebook status messages, or put some Google Buzz items up, not to mention FriendFeed, Whrrl, Pip.io, or other systems where you can capture your life’s most interesting events.

Now, how does that mother build an online scrapbook of all the items that were poured into the system? Sure you can use a tool like Scrapblog but how do you get Tweets into that? It’s not a curation tool for the real-time web.

Let’s also take on what would happen once we move into such a molecular world:

1. Search would INSTANTLY improve. (I need a whole blog post on why this is so).
2. Trends would INSTANTLY improve. (You’d have real meta data about important events, look at just the ordering data that would be available to study).
3. Brands would be able to advertise on bundles. (CocaCola would love to advertise on bundles of movie feedback, for instance, especially on bundles curated by the best movie curators — they will never advertise on raw tweets because the risk is too high that their brand would be next to something nasty).
4. A new monetization strategy would INSTANTLY become available for platform vendors like Twitter and Google Buzz.
5. Location services like Gowalla and Foursquare would be able to add real value onto bundles (showing location trends would be a key part of bundles, where they have no real play in augmenting “atoms” like Tweets or Flickr photos).
6. A new form of relevancy, credibility, and authority data would be available for systems to automatically present the best news. Look at how Techmeme appeared after blogging did. Imagine all sorts of new displays of best bundles that would now be possible. Even Techmeme would be able to recommend the best curators on topics, which would greatly improve the real-time news available there.

Anyone feel the need for this kind of new curation tool? Join in, please curate this post and push it around your networks. Let’s see if we can find some companies who are working on providing this new kind of real-time curation system. I’d love to work with startups who are working on just this. +1-425-205-1921 or scobleizer@gmail.com or leave a comment here and let’s work together in public.

Comments

  1. I'd be interested in doing something like this. You know how to reach me, Robert. Are you free to meet up at the Ritz in Half Moon Bay tomorrow around dinner time? :)

    1. I’m the developer of a platform called Swift River which allows the realtime curation and filtering of data, primarily targeting Crisis response organizations and Journalists. I’ll be at CHIRP in San Francisco and I’d be more than happy to share our experiences actually doing this stuff in scenarios like the recent Haiti and Chilean Earthquakes.

      There’s a whole separate level of complexity when it comes to distributed trust and authority.

    2. I’m the developer of a platform called Swift River which allows the realtime curation and filtering of data, primarily targeting Crisis response organizations and Journalists. I’ll be at CHIRP in San Francisco and I’d be more than happy to share our experiences actually doing this stuff in scenarios like the recent Haiti and Chilean Earthquakes.

      There’s a whole separate level of complexity when it comes to distributed trust and authority.

  2. Ok, I've been forward through this, then backwards through it, and will likely go through it again.

    (Buzzword alert) First thought: where does semantic web/microformat hoohah fit into curation?

    Semantic web seems to be a sort of tidal concept sweeping in and out of fashion every couple of years. From what little I know about current technology, microformats could handle some of this problem. Or maybe not, I may be too far behind the power curve on this one.

  3. Robert, I am doing this. My system is called foldier sPress and (while is still in its early beta) it is all about “curation”.
    Would love to talk with you and show you more about it… You can contact me on Twitter @micurs

  4. 1) I believe we can help with our plans for opengard.in First draft notion is to enable user tagging of information. If several atoms are tagged “Earthquake” you have a molecule.

    2) A like feature goes a long way to helping order information. This will need multiple views, a chronological stream view, and a like ordered view (perhaps further segmented by my friends likes, or all public likes). I'm not sure how to deal with relationships with a stream based social reader yet… Stronger reorganization can be done at a local level but to project this new order outward you'll need “influence”? I'm not 100% on this

    3) This is gold Robert. I'd love to have a tool that does this type of packaging from multiple inputs easily. We'll have user crafted group streams (which represent multiple feeds) that are put together, but I'm not sure how this translates into a package or presentation format. I think you're stuck doing some legwork here as a curator. At work – my dayjob dataminer/engineer, we kid around about the automatic view graph generator. We create hundreds of view graphs with a series of programs, graphic tools, and powerpoint automations, but they're still uncurated and require organization, commenting with bullets, highlighting for results/summary.

    4) I think I understand what you want. High level editing tools to weave in your own outside content into a curated stream. Maybe something like this can be done by “refeeding” a group stream you tag/organize etc and republish? You're picky Robert, but you articulate your needs well. I can't help but think if I can build a couple of tools that you, Louis and Mahendra can use to make your lives a little easier I'd be doing something very useful for the rest of us. You want drag and drop editing between raw feed/stream data and a report you make. I think that's doable with open web apps, but I'll have to look more into it. My friend Dave Semeria is working on a utility like that for the LM Framework, the mixable web, I'll have to talk with him about it.

    5) Refeeding, the concept I talk out above should cover this need

    6) Are you describing customized feedback tools that you can append to your curated stream? If so I believe there are some open protocols that are made for that sort of thing.

    7) Yes yes yes, you need analytics on all your interactions. You need to know what portions of your report/refeeds gain the most engagement and interest, basically what your readers/users value most. Every button & div needs to keep track of usage. Although I haven't written anything like this, I know it's available and is important to maintaining relevance, my driving design goal for social/info tools.

    I think you need to talk with Tyler about his thoughts on a feed based social reader. He and I hashed this idea out, but he's the chief hacknical wizard making it happen.

  5. Great points Robert, I'm going to expand on this with the concept of the 'inbox' we were talking about last night. To all who are interested in developing something like this, please look into oembed, (http://www.oembed.com) and see if the service you're integrating with has that built in or not. it's fairly easy to take all flickr and youtube urls from the tweet stream and show them in my “inbox”.

    Also a simple thing like pushing the urls that show up most in my twitter stream higher and in a separate list/bucket can't be that hard to do.

    anyway, moar oembed please.

  6. I'm playing around, the first experiences I saw though, don't look anything like Tweets and don't look like they do what I need for the real time web.

  7. All of those things are important but use language that many curators find is impenetrable. :-)

    Microformats help get us there, but aren't enough to solve most of the problems I laid out.

  8. Howard: you are doing some important stuff with real time web. Of course the finance people get this better than anyone. Seconds matter when you're trading stocks.

  9. The problem/opportunity as we see it is you/techmeme, arrington etc….totally contribute to our sector and yet ignore it and pass it off as TRADING.

    It's information and it's money.

    If the finance people got it, we would not still be independent :)

  10. 1. User tagging gets you a long way, just like with Flickr photos. I'd rather keep calling it bundling because Tagging doesn't get you all the way there.

    2. Like features don't get you there. None of the liking systems let you order the tweets or info.

    3. I'll do an audio podcast about this and what it might look like.

    4. Feeds don't get me there. Bundles do. Bundle your info atoms with mine and then I want to explain the pattern I'm seeing. Needs permalinks, etc. Feeds don't get you there.

    6. Open protocols, but packaged up as widgets. Sort of like how you are using Disqus here to talk with me. I added Disqus to my blog globally. I would love to have a Disqus widget that I could either add globally to every bundle I create, or only to certain bundles. I'd pay a buck!

    Would love to see where you're going and what you do.

  11. Yeah, sorry for using poor language. It's just in the valley that's how we talk. I own and buy stocks too and my dad has a million-dollar portfolio he manages. We still call it trading even though that really doesn't describe properly what it is.

  12. 1) drag and drop into a unique bin? you want to hash/unique tag certain feed/stream info
    2) Agreed on likes, individually they don't reorganize information from your viewpoint
    3) I'll catch the podcast to get your audio take
    4) How does a bundle differ from an aggregate feed you construct? (the podcast will likely cover this differentiation)

    I think we have to get the minimum functional version of what you need in first, in order to attract users and then investors to get the cooler features in. So we're a little picky on our side of what we can bite off at first. If only I could code up my ideas faster, everything takes time, digging through documentation, understanding the nuances of what's needed versus what's extra. If we do it all openly we can work together with a bigger community. They're plenty of experts sharper than I in specific categories you're interested in.

    Take your list and slice it down the the top 3 and refeed it to us. Some combination of your fans and friends will get you there. It's taking too long in my opinion, that's why I started down this road late last year ( I started with matching relevant ads to tweets, now look at the problem we're working on )

  13. It is not only weirdos who need this. I need it. Sh*t, I can't even put my hands on all the hashtags for my own events unless I search and save the search. Then to update, search and save the search. It would be great if Twitter carried metadata, as Winer has always said, because almost all the other media (blogs, photos, video) already do. And then tags could be aggregated in Posterous-like pages.

    You can see that many people have broken their picks on this already: Posterous, Tumblr, NutshellMail, Brizzly, etc. They do the aggregating and leave me to do the curating manually.

    Hey, what about JustSignal, Brian Roy's effort? That's the best I can do.

  14. Awesome post…particularly in regards to the idea of bundles. Through the build out and growth of two startups, my team has been focused on solving the bundle idea. Do you have any interest in providing feedback? If so, we're on and committed to putting your ideas to work.

  15. Its all about the metadata.
    Applying Semantic Web (machine) entity identification, OpenCalais &/or Zemanta to the metadata; then applying (human intelligence) curatorial weighting; can achieve the sort of hugely scalable “real-time curation” solutions you are calling for, Robert.
    I was the original architect of such a platform Currently in development) for online public interest video funded by a $2.1M grant from the Bill & Melinda Gates Foundation: ViewChange http://www.linktv.org/viewchange/technology

  16. Yes, it's all about metadata, but it's also about UI and experiences. Just pasting metadata onto Tweets won't get us there. We need bundling and that goes beyond just metadata. This is why I don't call bundling “tagging.” Tagging is a piece, yes, but not everything.

  17. Phreadz can do most of what you ask here – supporting all the content types you listed. All that's needed is the tool for re-ordering – which would be fairly simply to do. Rather like an OPML editor can drag node/branches (bundles) around. (I did this for the rss feed manager on podcast.com)

    Every 'bundle' / thread / post / user / channel has an RSS feed, a JSON data feed (river) and also an OPML structure. The API has it all there.

    I bundle stuff all the time on Phreadz – related 'V.I.T.A.L' media (video, images, text, audio, links)

    Here's an example : starting with your tweet: http://phz.in/22vt

    Now see it's OPML http://api.phreadz.com/thread/GVFF30J61/opml/

    .. and its RSS http://api.phreadz.com/thread/GVFF30J61/rss/ aznd JSON http://api.phreadz.com/thread/GVFF30J61/json/

  18. Yes, I was a bit hyperbolic withe the “all” about the metadata.
    “Tagging” within the Linked Data model is very scalable, in that the “tag” refers to just one thing/concept, a semantic “entity.” Regular ol' tagging by humans is hardly scalable and introduces spurious noise – spamming/scamming/gaming. Thompson Reuters' OpenCalais service is a free tool of enormous benefit in disambiguating “tagging” and allowing for greater machine interpolation of meaning. Take a look at Zemanta too.

  19. Phreadz can do most of what you ask here – supporting all the content types you listed. All that's needed is the tool for re-ordering – which would be fairly simply to do. Rather like an OPML editor can drag node/branches (bundles) around. (I did this for the rss feed manager on podcast.com)

    Every 'bundle' / thread / post / user / channel has an RSS feed, a JSON data feed (river) and also an OPML structure. The API has it all there.

    I bundle stuff all the time on Phreadz – related 'V.I.T.A.L' media (video, images, text, audio, links)

    Here's an example : starting with your tweet: http://phz.in/22vt

    Now see it's OPML http://api.phreadz.com/thread/GVFF30J61/opml/

    .. and its RSS http://api.phreadz.com/thread/GVFF30J61/rss/ and JSON http://api.phreadz.com/thread/GVFF30J61/json/

  20. You're right, but Postling is expanding what you can do with Tumblr. Think of Tumblr as he output device and Postling as Social Media dashboard on steroids.

    They're already implementing many of the ideas you mentioned and I'd bet they would be open to integrating some of the others. You should get in touch with them. I'll pass this link along to them as well.

  21. Hi Robert. Thanks for this post. I would be very interested in working with you on that. In fact we are already working on this kind of tool. Our goal: provide a smarter way to publish Twitter (and make Twitter readable for non-Twitter users).
    We also provide analytics (the widget plugs itself into your Google Analytics) and a way for people to subscribe to the feed (filtered by keywords) by email (using the tool we developed earlier http://listimonkey.com).
    See for ex. http://beta.publitweet.com/scobleizer/most-infl
    http://beta.publitweet.com/scobleizer/most-infl
    Our office is in the Twitter building first floor. Come by anytime!
    My phone is +1 415 493 8403

  22. Great ideas. This was a really enlightening post to read. Thanks. Saw the mention about pip.io too. I just signed up today. There's something very appealing about that sight.

  23. Robert – You're describing a world where we'd get real news … no spin … no daily updates about Lindsay Lohan mixed in … hard news sorted by relevance AND monetizing it at the same time. I'm guessing you're going to get more than a few (positive) phonecalls. LONG overdue both in the US and worldwide. I'd LOVE to be tuned in to a REAL working project like that. Fascinating!

  24. Check out http://www.twavl.com/denver and check out the partner bar atop the page. The partner bar bundles Tweets from pre-selected Denver businesses. We're able to filter tweets, or alternatively, use only favorited tweets. For a different view of the partner bar, check out http://blog.connectme360.com – it features tweets from corporate social media strategists, a list I help Altimeter consultant Jeremiah Owyang manage.

  25. This has got to be, hands down, the best post I've read in a long time… not to mention, completely spot on. In fact, I was at the Twitter Tools panel at SXSW when you had mentioned all these points and wondered if you were going to elaborate, so thank you!

    I agree with all of your points. The thought of a real-time curation/bundling platform is very exciting. Not only would it be useful to the everyday publisher but also extremely useful to news and media sites. But to answer your “why hasn't this been done?” question, I would add that it's all of the reasons you mentioned plus the sheer fact that the real-time web/social media has accelerated much quicker than anyone would have ever expected. Companies in the real-time space are innovating and pushing out products at such a fast rate that we now need the type of curation platform you're talking about. Was there a need last year? Probably. Was there a need two years ago? If yes, nobody would have thought so.

  26. Hello Robert! Great post! Thank you!
    You've thrown down an impressive gauntlet for real-time content curating. At http://www.HiveFire.com we are exited stealth in Q1 and we're focusing on the need for B2B Executives to participate at real-time content curators. When it comes to bundling, you're right. It must be easy and it must be automatic. It can't be hard and convoluted. Curators spend the hard part of their time understanding the context of the information artifacts they are selecting. They shouldn't be spending time trying to put bundles together of the information they're collecting.
    One of our successful B2B customers has created “bundles” around entities that are automatically extracted from her curated and blogged content. Yeah, we analyze content and extract out the important bits and build real-time collections around those. Check out Verne Global's installation of HiveFire at http://www.greendatacenternews.org or Identity Force IdTheftDailyNews http://www.idtheftdailynews.com. In addition to bundling by categories, the Verne Global Team curates bundles of information around “entities” or “mentions”. Click any mentions and you'll be in a complete real-time category of curated content around that specific key term. Collections grow around their content while they curate and assemble tweets, pictures, news, and blogs. Their entire site is a collection of insight into relevant content, in real-time around a topic. These collections of curated content are meant to be
    Of course, we can do better. We can do your #3 and #7 much better and we're heading in that direction. Our distribution and sharing needs work as well. Each collection of curated entity information is its own feed, but we can do more to make it easier.

    Lastly, you didn't mention a big factor: Real-Time curating shouldn't take lots of time. It's already hard to process tons of data coming in real-time. I think any real-time curating application should watch out for curator time. Curators are experts and they shouldn't burn up their time organizing their curated collections of content. Any solution should show how little time it takes to do the job well.

    It would be a pleasure to connect and share more about what we're doing. Again, we're more B2B focused, but I'm sure you may have some insights and we'd welcome sharing more about how our customers are leveraging their organization intelligence to curate relevant content for customers and prospects.

    My email: taariq@hivefire.com. Would love to chat!

  27. I think another interesting benefit of a living news bundle would that it enables folks to add to a story without simply duplicating the work of others (even within their direct social circle. I touched on the almost urgent and uncontrollable need for many people to share a news story – even when the story has been shared a million times before – with their followers and friends in a piece called “We can't stop sharing news.”
    http://tweetagewasteland.com/2010/03/curation-n

    When I use digg, the system tells me, hey, we already have this link. On twitter and facebook, no one gives me the same warning so you hear the same story for the sixth time.

  28. Most of the comments have been from folks who can facilitate this happening. I can't.

    But as a creator/curator, I would use it. It feels full-up with potential.

    And obvious, in hindsight.

    I hope your seed flowers.

  29. I propose you call the Molecules 'Scobles' and the software/app 'The Scobleizer'.

    I really like this idea of Scobeling atoms together and can really see how it could be of use generally not just by super-curators like yourself and Mahendra. It would be like a blog-post on steroids or a hyper-scrapbook.

    This idea of atoms and molecules (Scobles :)) help explain the concept. I see it as much larger file/folder system. Why limit yourself to an application? I like the idea of collecting atoms in to buckets, drag and drop from streams, webpages etc…I thought Google Wave was heading in this direction and is the closest thing I can think of to what you have described but its not on the money yet is it?

    Feedly is also close to the canvas concept you describe except that exposes the need for curation/editing.

    I'm looking forward to seeing you harness the power of your network and see this great idea implemented.

  30. “Let’s consider the mother who has a 1-year-old son. She invites 30 of her friends to a birthday party for her son. They take videos, do Foursquare checkins, one or two might blog about the party for their mommy blogs. Many take photos, but some of those photos end up on Facebook. Some on Flickr. Some on SmugMug. Some on Picasa. Lots of them Tweet about the event, or Facebook status messages, or put some Google Buzz items up, not to mention FriendFeed, Whrrl, Pip.io, or other systems where you can capture your life’s most interesting events.”

    Having just been to a 6-year old's birthday party, I can say that the actual number of mothers likely to use more than a digital camera and Picasa is low – very low. As for foursquare checkins and mummy blogging!!… well may be in Silicon Valley, but not in the rest of the planet (yet).

  31. Wiki software can be used to do some of the things you listed. Each bundle would have its own wiki page.

    A tool to easily select an atom and put it in a bundle (wiki page) would be tricky to make, because there are so many kinds of atoms (but you could focus on the 20 main ones, I suppose). Imagine going to a webpage and getting a short list of all the page's atoms at the side. I wouldn't be surprised if there's already a Firefox add-on to do something like that.

    A wiki page can have arbitrary text (for editorializing), tables, social widgets, comments, tags, groups, categories (for more advanced taxonomies), a permalink URL, an RSS feed, a comments RSS feed, etc.

  32. Great stuff Robert – I think you are spot on – the collaborative aspect of this may create even more fascinating opportunities for “reality sense-making” from infinite alternative viewpoints.

    The critical change I see in collaborative real-time curation is also that, collectively we are now moving from individual gathering, collection and redistribution to collaborative making sense of reality at different levels. And this is a huge, powerful opportunity for everyone.

    Thanks Robert for writing this. :-)
    A big embrace.

  33. Gotcha.
    I can subscribe to feeds, an update (whether we pull or get it Pushed to us) is a bundle.

    I'll need to do some more research into a content creation format that is optimal for drag and drop + editing. Funny how even this simple concept hasn't shown up on our radar's yet. How far would painters have gotten without a brush?

  34. Apture can make #1 easier already like I wrote here: http://www.eenmanierom.nl/eenvoudig-multimedia-

    Videos can be embedded, even with timeframes. And if you paste in a link to a tweet it'll show up. You can see both in this post: http://www.eenmanierom.nl/te-tellen-hoe-vaak-je

    * my posts are in Dutch, Google Translate is available in the bottom left corner but to show you what Apture can do you can just scroll through the example post I mentioned above.

  35. Would Ellerdale.com do anything valuable in the direction you mention, mr. Scoble? I must hope they'll find your post, that I consider a nice one, considering that you write a roadmap to the semantic web, without even using that dirty word.

  36. http://listoftweets.com does some of this, it's a little side project I built last year. My main priority is to allow pastebin-style saving of lists, so you get a unique URL for each list (bundle) you compile.

  37. Thanks for bringing up such a great topic. I find what is also important is the characteristics of “curator” or sensemakers, from my observations and speaking with other “sensemakers” that in order to reach next level understanding and innovation we must be able to do things that you mention.

    What i would like to see is a real time mindmap being created in real time that compliments blog posts, i find after a 100 comments i can't capture the information that may be surfacing between the lines, a mindmap serves as a great way to compliment a blog that is getting momentum.

  38. Superpoint Spiro! it is the sense-making aspect the key core value of this process and yes!, mindmaps are a perfect vehicle for giving a comprehensive, organic view of a rich and granular object like a set of comments than the linear structure of a page.

  39. How does anyone go about choosing the “the best ten tweets which describe that event” without research of sufficient depth and accuracy to cut all the way through broken telephone-style opinion in order to get to the facts of the earthquake used in your example? Could someone do a word-by-word contextual comparison to the original story referred to earliest in the tweet thread? Good grief – that would be a difficult thing to do.

    Crowd-sourced information is inherently inaccurate because too many originators lack the reticence or experience to tweet accurately about events in which they're directly involved. That lack of accuracy (people directly involved in events tend to tweet continuously, gradually moderating or making their tweets more accurate as they slowly acquire new information, or skew their impression of real-time events to suit their political or social mores), keep most twitter users following, effectively wasting enormous amounts of everyone's time. If I had the time, I might consider it. If there was a real curator (computer, computer-aided human) comparing the absurdities, fantasies, inaccuracies and eventual conclusions generated by tweet originators to hard reporting by trained observers or scientists and so on in order to produce a definitively accurate thread/collection, then crowd-sourced event information tweeted by inviduals living within the context of an event might have some value. Until that happens, tweeting and twitter itself are just another colossal waste of our time and attention. Twitter is most accurate when it disseminates information about individuals tweeting about their own personal activities.

    All of the ways in which curration is defined in English dictionaries relate unerringly to maintaining and preserving a collection. Play with definitions if you like – no one should stop you – but please don't warp a perfectly decent word into a catch-all for a fantasy about yet another manifestation of technology which serves us poorly because of how it fragments our time with largely absurd or otherwise pointless distractions. Twitter will likely never be a reliable source of event information beyond “Band X tickets just went on sale at TicketMaster.” Useful indeed, but of short-lived usefulness, and not worth the efforts of any Curator.

  40. The content DJ (aka curator) is definitely such an opportunity. Especially if you can put a unique spin on it. My philosophy is success comes from content.

    Either content you:

    1. Create
    2. DJ or Curate
    3. Sponsor

  41. Hi Robert- As usual, you keep pushing the envelope, and that's great for everybody. Indeed, “Curation” is arguably the most misused and abused emerging term.
    We already have several of these features, and more upcoming ones. That's how we put together this portal on you http://bit.ly/chVwPw, but we haven't yet exposed the Curation Admin dashboard which is to be used by business, not technical users. I wanted to talk to you about this when I was in SF a few months ago.

    Btw – How much do you expect to pay for these capabilities? And do you prefer a one-time license or a lower cost recurring SaaS subscription?

  42. I agree that this recently popularized use of “curator” is off track. I think when we bring together these atomic bits of information we're operating much more as filters or switches. In fact, I greatly prefer either CONDUCTOR or IMMEDIATOR (immediate mediator) as far better metaphors for the role you're describing. Immediator is quite accurate actually, whereas all other words I have seen used are borrowed from other (analog) disciplines which don't have the web's agility or basic properties. And yes, I think semantics matter in this case. Just ask any well-respected fine art curator whether they think what you're talking about has anything to do with “curation” and see what kind of response you get ;)

    Nonetheless, whatever terminology we agree on, the fact does remain that there are tools missing to carry out this work, and for that I commend you on your analysis on some of the problems which we encounter.

    You and your readers might greatly enjoy reading a white paper I self-published some weeks ago about a process which I call IMMEDIATED AUTODOCUMENTARY which attempts to take a harder look at how self-reflexive media practices and the proliferation of cheap/fast digital video cameras are changing the face of audiovisual communication in marketing and storytelling.

    http://ks12.net/files/GabrielShalom_imautodoc_1

    Greetings from Berlin,
    Gabriel

  43. Robert,

    Don't act too surprised when I tell you that this has already been thought of and people are working on it, just not on mainstream. While it is true that it would scare most people to try to be curators in this model, or to use these tools, the future is exactly as your describe (plus minus a few smart people making it different).

    You described how the semantic web should work.

    Nice summary, well done.

  44. Having watched the web develop and become what it is today, over the past 5 yrs, I can appreciate what you mean. I have conducted small experiments(mostly with respect to making blog posts which draw from different things) and I have found the changes which you say. Maybe I should get into this…

  45. Thoughts while planning some curation:

    Curators need to answer questions definitively — it's not enough to produce tons of content, it has to come together to address a real need. How many questions does your curation answer, better than any other site on the web?

    Curators need to ask good questions, too — what's the ideal way to present and interact with this information, if anything is possible? What if other sites are doing it inefficiently or just plain wrong?

    Curators need to discover their own data sources — don't expect everything to come from Twitter. There's great untapped sources of semi-structured data that only need a litte TLC to be fully-structured.

    Curators need niche focus — I think the tendency is to go too broad. Generally, when this happens, users lose.

    Curators need systems — probably for aggregation, definitely for participation. A solid layer of curated content is usually begging for more layers to built on top of it.

    Curators need heart — for when systems don't solve the “last-mile problem” of content.

    Real need is more important than real time — maybe they intersect, partially… maybe not.

    At least, those are the things I'm taking into battle. Appreciate the topic and the discussion.

  46. It's open source, and I have the code on my desktop now. By all means let's see if it can be “liberated”. I'm talking with an expert on wave bots, and creating external sites with wave content now. It'll take me a little longer to hit the ground running otherwise.

  47. Robert,

    Great points and observations. We are working on a similar project on http://socialpulse.comto engage members to create conversations about any topic, curate them themselves and with other members and create social capital for all participants.

    For example this is our social profile for Haiti Earthquake http://www.socialpulse.com/Haiti_Earthquake. It gathers real time news and tweets and anybody can post on the wall using their SocialPulse login or Facebook or Twitter account. As soon as the post is made it gives a prompt to post the item on your twitter and facebook wall which positions the link directly to the position of the post on SocialPulse.com.

    The peer to peer wall and blog system can be installed on any site and the members accumulate their content and social capital (reputation, insight, wisdom) in one place.

    We are Beta testing now and hope to launch it by April 20. Like to hear your (including vthe visitors) comments and suggestions.

    Thank you
    Sincerely
    Tom Vellaringattu (tomvell@gmail.com)

  48. Robert

    Our http://www.socialpulse.com wall system can be a good solution to your ideas above. You can see it working at the bottom of http://www.rahulgandhi.us . Users can use a socialpulse.com login to build social capital or use other logins like Facebook and Twitter as guests to post on the wall.

    I like to talk to you more about our ideas for students providing an activity journal system to document all their academic, social and community services for presenting to college admissions and at job interviews!

    thank you
    tom vellaringattu @tomvell

  49. Such timely thoughts. I am building a new project built around content curation without automation. I hope to launch later this month with a network of 30+ highly targeted sites all curated by and for women.

    Each site will be targeted by interest. From Fashion to Finance to Startups, I am recruiting women who naturally find quality content and are able to add relevance and meaning to the information as an editorial component. Because I am building the sites in my usual way (bootstrapped) I haven't thought of adding Tweets/ Updates/ Checkins to the mix. The focus at this point is sharing and editorializing articles and information through a human filter.

    I think the best curation will always be done on the human level. It is the individual curator (or group of curators) who has the depth of knowledge to quiet the noise and curate what is truly relevant for the reader.

    Participation widgets! Now that is a tip I am going to go explore. Thank you.

    And on a side note, your pics from SXSW- http://www.flickr.com/photos/allisonworthington:)

  50. Sounds like GP or your Mesa incubation lab need to get on something like this. It would be like a customized, real-time newspaper.

    This is really what the AP should evolve into.

  51. To summarize the seven pieces above, they are:
    1. Aggregation and Clustering
    2. Ranking
    3. Dissemination
    4. Editing
    5. Updating
    6. Polling
    7. Analyzing

    A service like Google News handles #1 Clustering and aggregation, #2 Ranking, #3 Dissemination (via Website, Email and RSS) and #7 analyzing by incorporating that feedback into the machine-learning ranking algorithm. However, steps 4, 5, 6 are yet to be automated.

    At HiveFire (http://www.hivefire.com), we're taking a similar approach to enable B2B marketers effectively curate content efficiently by automating as much of the process as possible. At the end of the day, our aim is to allow curators to focus on curation while automation takes care of the rest.

    As my colleague Taariq mentioned above, we'd love to chat.

    Pawan Deshpande
    CEO, HiveFire.

  52. Lots of people on this list say they or someone they know is working on this:-) GP would be the one, but they are busy with a hundred other things.

  53. Well that's all nice and curmudgeonly and everything, but unfortunately, playing the contrarian here doesn't make you look smart. It makes you look like the guy claiming computers will make us look dumb because we won't learn how to use a slide rule.

    I'll get off your lawn now and go continue wasting colossal amounts of time and attention doing what I do for a living.

  54. I think what Mark is trying to say is that Wave is both a protocol and a client. The crappy UI was just one of many possible clients, as the twitter homepage is just one of many possible clients (and certainly not the best).

  55. Howdy Tom. At VM we're looking to use a social capital/point system as well. How have you find that helps encourage behavior in tagging/filtering content in their streams. Obviously points can't match getting a good core community, but I'd love to hear about your experiences. Hit me up at messel at victusmedia dot com if you want to chat anytime.

    The code we're developing will be open sourced, so perhaps there are pieces we can help out with that you may desire.

  56. At http://favit.com we work exactly on a tool like this and not surprisingly we are meeting some of the features your are requesting above now – the future is happening as we speak :) We are preparing a new big update that will be released shortly and some of the new things that we will be rolling out that aim to bring us closer to the molecular world are described at our blog: http://blog.favit.com/en/blog so stay tuned – the real-time curation engine is closer than you think :)

  57. Yes, certainly. However, interpretation seems to be a part of the editing phase (whereby the curator interprets third-party content and adds their own interpretation in the editing process). Interpretation is also at the heart of curation. A curator has to decide what content to curate and what not to curate and is only able to do so by interpreting every piece of content.

  58. But Calais tops out at 50k hits per day, for a web service this number of updates is very small (certainly one that wishes to help curate the real time web).

    A single user could leverage it well though it's still noisy (lots of false positives from a single post), although non-uniform humans are better daggers of information by group.

  59. I've got a detailed plan to implement this from a free agent who's very interested (I've used their work before). But this will take an external path for low angel level funding to ensure it gets built fast, and according to the specs you are asking for (they can't tolerate startup risk/timelines atm). This is one of those times where we can build something in a couple months on our own, or in a few weeks with funding and the right help.

    I'm excited about the opportunity either way, and look forward to what other startups are building. How about a contest? The winner(s) pony up the tool and support for potential seed funding support.

  60. Hmm, bear with me, curating my thoughts while typing this… means trying to focus. Move important items to where they are easier to find and grasp. Bundle as kind of category, possibly loose, like friends of friends (which get you jobs and partners, maybe).

    Mind thinks in similarities and differences. Intelligence as the ability to assign relative importance. We all do, and it is part personal, part shared in agreement. Depends on personal viewpoint and context. It is about self-determined direction of attention, made easier if others keep score. The kind of value a secretary or an editor adds.

    What feels like progress, what gets results? For many, it is live energetic discussion, augmented with clever non-linear recording (mindmap, concept map, comment stream, wiki for example). Watching a video stream is nice if you can jump in from the transcript, TED style or read, that is 3x faster, for the loss of entertainment value.

    Write so you don't have to read it all (Robert. E. Horn on structured writing).

    I see here a community brimming with toolmaking power. Hi, I'm new. Not a programmer, but I get processes running for people. Sometimes. How to close in on Robert Scoble's vision? For me, focus and augmented understanding gets to ample flow in personal discussion. Live. Skype, or better. With chat and a, um, self-curating backchannel that does not distract but augments the discourse, respectfully.

    This means two classes of roles, not egalitarian:
    1. the partners in discourse, free to improvise, no burden of keeping record and score.
    2. loving real-time curators who focus the message which may be recorded or not
    Sounds like it may take some discipline and practice to get good enough for near real-time.

    Steps I imagine to get there: design experiments to prototype real-time curation, as objects and processes if you will. Act them out – first in slow motion with people acting the parts. Talk-through, walk through, run-through. Repeat as needed. Then we may see what bears modeling in software and new components may come into view.

    This is inspired by Venessa Miemis' Junto concept. Its value would come with real-time curators or immediators as Gabriel Shalom prefers to name it. http://emergentbydesign.com/2010/03/22/junto-di

  61. Hi Pierre, how would you distinguish between what Robert calls the single Atom item and the more complex constructions of one or more Atoms, or Molecules when you convert things to social media goods nomenclature?

    Anthony Blow (AnthonyGadgetX )
    http://www.gadgetx.com

  62. [...] The Seven Needs of Real-Time Curators – I keep hearing people throw around the word “curation” at various conferences, most recently at SXSW. The thing is most of the time when I dig into what they are saying they usually have no clue about what curation really is or how it could be applied to the real-time world. blog comments powered by Disqus var disqus_url = 'http://www.newmediaphiladelphia.com/bookmarks-for-march-29th/ '; var disqus_container_id = 'disqus_thread'; var facebookXdReceiverPath = 'http://www.newmediaphiladelphia.com/wp-content/plugins/disqus-comment-system/xd_receiver.htm'; var DsqLocal = { 'trackbacks': [ ], 'trackback_url': 'http://www.newmediaphiladelphia.com/bookmarks-for-march-29th/trackback/' }; (function() { var dsq = document.createElement('script'); dsq.type = 'text/javascript'; dsq.async = true; dsq.src = "http://newmediaphiladelphia.disqus.com/disqus.js?v=2.0&slug=bookmarks_for_march_29th&pname=wordpress&pver=2.33"; (document.getElementsByTagName('head')[0] || document.getElementsByTagName('body')[0]).appendChild(dsq); })(); [...]

  63. Robert,

    The Pearltree I created at the Swagapalooza event last week which I shared with you (see it http://bit.ly/aN81dI) is an example of real-time curation but your list includes many features we could add in the future.

    I've been writing about Curation as the future of media for many years now, long before this term became popular (see my blog post from 2008 – http://bit.ly/18Kpor)

    Curation is a very complex topic – I personally think it is about pulling out and telling great stories, about capturing data (perhaps in near-realtime) but then more about putting order & focus to the streams and to content that is all too easily lost. Curation is about selectivity and focus – it is about pulling out stuff & highlighting it.

    It is also all about context – presenting information within a context and adding context to individual bits of content. Here the new media curators are doing what curation has always been about – presenting a point of view & telling a story through the highlighting of works (curation is as much about what isn't displayed as it is about what is displayed). Curation is also all about a particular point of view – the curators.

    And like in the case of art or museums not all curators are for all visitors – here in SF, for example, while I like the building, I don't like the point of view or the choices made by the curators at the de Young.

    Curation is also about a lot more than just media reinventing itself (though to address a point made above – the future curators are more like old-school Newspaper editors than individual reporters). Curation is also the future of many types of businesses – online as well as offline.

    Great curation is a very viable reaction to the Walmarts & other big boxes of this world. Amazon.com's “lists” are an example of adding curation on top of a nearly infinite set of choices & objects for sale (their A-stores are another somewhat similar layer). But so to are amazing online successes such as Woot.com, 20×200.com and many other sites which sell only a small number of carefully chosen objects.

    If anyone wants to follow this topic further I'm building a bunch of Pearltrees on curation see http://www.pearltrees.com/rycaut)

  64. i really like you blog. i learned a lot reading here today. Keep up good work. I will make sure i visit this blog daily

  65. Thank you, Bernard, for mentioning that piece. Your related ReadWriteWeb post really helped me review my process for creating that. During that review, it became apparent that we simply don't have the tools to do consistently do this type of work in real time.

  66. I’ve been thinking about the issue of realtime curation as well. It started on this year’s SXSW while keeping up with tweets during Clay Shirky’s talk to make a blogpost of it. Cumbersome work with tweetshots.com got me thinking, this can be done easier…I made a small mockup of my thoughts, you can check it on my (dutch) blogpost “In three steps to real time curators? yes please!” http://bit.ly/9ZLfnx
    The idea is simple: Collect URL’s with a bookmarklet, shuffle them in the way you want them viewed and make a slideshow of it. This slideshow can be embedded, emailed, shared and re-used for new slideshows.
    You can see the screenshots from my mockup in the above mentioned blogpost

  67. I've been thinking about the issue of realtime curation as well. It started on this year's SXSW while keeping up with tweets during Clay Shirky's talk to make a blogpost of it. Cumbersome work with tweetshots.com got me thinking, this can be done easier…I made a small mockup of my thoughts, you can check it on my (dutch) blogpost “In three steps to real time curators? yes please!” http://bit.ly/9ZLfnx
    The idea is simple: Collect URL's with a bookmarklet, shuffle them in the way you want them viewed and make a slideshow of it. This slideshow can be embedded, emailed, shared and re-used for new slideshows.
    You can see the screenshots from my mockup in the above mentioned blogpost

  68. After a few days of thinking, it's worth it for me to finish up hRecipe. If it helps solve even a little bit of the problem, it's worth it the time. Thanks.

  69. I like this article. I come from a technical communications background, and I’ve been arguing that technical writers need to become content brokers for their communities. Curator is a better term, if only because it is sexier, and technical writers need all the glamour they can lay claim to.

    I’m interested in seeing how this all plays out within the intersection of enterprise content and communities of practice. Most of social networking practice arises out of marketing, and I’m trying to help migrate these ideas into a traditional yet overlooked space in which there is much potential.

  70. I like this article. I come from a technical communications background, and I’ve been arguing that technical writers need to become content brokers for their communities. Curator is a better term, if only because it is sexier, and technical writers need all the glamour they can lay claim to.

    I’m interested in seeing how this all plays out within the intersection of enterprise content and communities of practice. Most of social networking practice arises out of marketing, and I’m trying to help migrate these ideas into a traditional yet overlooked space in which there is much potential.

  71. You've lost me at #2, Scobe. Who's to say that I believe in what someone else believes is important to me? I should be able to make that distinction myself, and aggregate content thusly, without anyone else's help. Give me that tool.

  72. Robert,
    Your comments about the need for both curators and curation tools are right on the mark. You've provided a handy to do list that outlines ways we can improve Loud3r. That said, we should set up a demo of where the product is today. I think you will find it touches on most of the items in your wishlist.

    Lowell

  73. Robert,
    Your comments about the need for both curators and curation tools are right on the mark. You've provided a handy to do list that outlines ways we can improve Loud3r. That said, we should set up a demo of where the product is today. I think you will find it touches on most of the items in your wishlist.

    Lowell