Social Media - What Needs To Be In The Cloud?

Cloud I had an interesting email discussion yesterday with a friend. He suggested to me that media doesn't need to be in the cloud to create a great social media service. He said that the files can be stored locally and only the data needs to be in the cloud. I'm not so sure about that.

I've been scrobbling my music listening data to for a few years now (since Oct 2005 to be specific). So my data is in the cloud. But to be honest I don't get much social media value from that. 

The best experience I get from scrobbling my listening data is's neighbor radio which is a service that allows me to listen to a radio station programmed by my "musical neighbors". But it's really a solitary experience.

Compare that to what happens on Tumblr, for example. I post music on tumblr every day, one song a day. And I get feedback in the form of comments, likes, playcounts, and reblogs. I get to see what people think of the music I post. And I also get to see and listen to what other people are playing and listening to. I can do the same thing; comment, reblog, and like the music I find there.

The same is true of many other social music experiences out there; passing links to music tracks on twitter, listening at the hype machine, engaging at

So the question is if we can get a truly social experience with the files stored locally on our machines or networked storage devices and just our data in the cloud. Or do we need the media itself in the cloud?

My gut instinct is that we need both the media and the data in the cloud. They don't need to be in the same place or service however. I would love tumblr/listen (you have to be a tumblr user to use this service) to incorporate my scrobbling data to create a more personalized version, for example.

The primary reason I think we need the media in the cloud is that is the only way I know of to ensure that I can listen to the same thing you can listen to. If we continue to rely on file based media, then you might say you love I'm Turning Into You by Wheels On Fire but unless it's online (it is, right here), then I would not be able to listen to it and let you know what I think and more importantly share it with my friends. As an aside, I found this great song because I follow Luke on Tumblr and he posted it a few weeks ago.

The same issue is true of TV shows, films, photos, games, news, and other forms of media we consume both on the web and off of the web. If we continue to rely on files as the way we consume media, the social activities around our consumption will be muted. If we move to streaming all media in the cloud, we will see an explosion of consumption driven by social interaction.

There's clearly an issue of bandwidth versus storage costs in this whole equation and all things being equal it certainly would be better for media to be stored and consumed locally, but I think the power of social media and social interactions will change the balance in favor of cloud based media. What do you think?

Note: The photo in the upper right of this post is of Mt Shasta and was taken and posted to Flickr by SP8254.

Comments (Archived):

  1. joe essenfeld

    Keeping data in the clouds also makes piracy more difficult and enables users to access data on multiple devices without syncing them.

    1. fredwilson

      Good point about piracy

    2. Prokofy

      How does it make piracy more difficult when it’s all in one place, the cloud? Hack the cloud, you have everything. And the cloud sounds like it will be everywhere.

  2. Jean-François Noël

    I think we still talk about local versus the cloud, only because of bandwidth. I do not think this is even about the social experience. We move more and more of our data in the cloud and the people who are still stuck with the local version of this data are always amazed that I can reach my data from all my computers and my iPhone pretty consistently.There is a problem right now and you allude to it in your post. The data is not really in the cloud. The data is on the servers of ONE company and those servers are connected to the cloud. This means that for all the services that this company provide your data is reachable on the cloud, but for the services they don’t provide your data vanishes. This also means you have to thrust those companies with the security and privacy of your data. I would love to have all the data really in the cloud, under the sole ownership of its creator and have services compete to serve this data.I may not be clear so I’ll illustrate with emails. In this scenario my emails would sit in the cloud and I would give access to gmail to use their interface to manage them, they could cache the emails and do whatever they want while they have access, but they must guarantee they maintain the data in the cloud, this is the master copy. When I want to do a special function with my emails, I just give access to another company that provides this service for all emails. Robert Lefkowitz, once talked about something similar and he gave an example with music. The essence of it is, when you put me on hold and make me listen to some music, it should be my music (or the music I like).

    1. fredwilson

      there are companies that are allowing all of us to store “our” music in the cloud. but i think that still creates some of the same problems. it moves from access to permission and you may still end up not being able to really share media that way.

      1. Jean-François Noël

        Yes, the sharing can still be blocked even if everything is staying in the cloud. If this model was to become ubiquitous then the original producer of the media may end up being the one controlling the possibility to share the media. I was probably too focused on the ownership of data and not enough on the sharing part.

  3. ErikSchwartz

    The primary reason I think we need the media in the cloud is that is the only way I know of to ensure that I can listen to the same thing you can listen to.Everyone having the access to the same media is the key to media being social of course.But you are underestimating the speed at which the price and volume of storage is becoming cheap. Storage increases about 2 orders of magnitude per unit price every 5 years and that has been the case for about 25 years. I can now buy at retail a 1.5 Terabyte drive for about $100. Assuming the same rate of increase continues in 2014 I will be able to buy 150 Terabytes for $100.150 Terabytes can store practically every recording of every song ever written by mankind. Pandora’s entire library fits on about 2 terabytesAt the same time the bandwidth situation is barely keeping pace with needs. The internet works well the way it is used now because it’s easy to run it in a contention based manner. While I’ve been typing this I have not used any bandwidth. Streaming doesn’t work that way. 100 people using email can easily share a fairly narrow pipe. 100 people streaming cannot.So not only does going from browsing to streaming increase the bandwidth needed in an exponential manner rather than a linear one but upgrading bandwidth is not simple in many cases. Bandwidth we need to break down into several areas. Backbone, local wired, and wireless. Backbone is relatively easy to upgrade because you can just add more redundant fiber. Local wired can do the same, but it is much more expensive and involved. Wireless is a whole other bucket of fish. Despite the protestations of some pundits I don’t see high bandwidth wireless connections coming soon in a way that can scale to the mainstream (for a lot of reasons).What caught my eye that led me to dropping Fred a note was that AWS is now allowing you to populate S3 by mailing them a disk instead of uploading via the internet (…. It reminded me of my youth (1989) when we joked the fastest way to get 40 meg cross country was a syquest cartridge and a fedex envelope. The same is true now but the numbers are bigger.At the end of the day it’s a matter of economics. Both approaches work. But the media business is in flux and clearly becoming one based on razor thin margins. Building the cloud out to support simultaneous high bandwidth for everyone requires huge capital expenditures. If huge upfront capital expenditures are required to be amortized it greatly limits the business models for media moving forward at a time when the media business needs maximum flexibility in biz models to survive.I’ll write some about the scrobbling side of our conversation in another comment.

    1. fredwilson

      this is a great point and perspective erik but it also suggests that we can all own all that media at an affordable cost

      1. ErikSchwartz

        I think that the business models around media moving forward are inevitably going to make the bits themselves meaningless and the data around the consumption of these bits the value point.

        1. fredwilson

          from your mouth to god’s ears!

          1. Prokofy

            Appropriate, because it is a religious belief.Bytes still matter, because they are a musician’s work product.Metadata about consumption only gives you metadata about a lot of freeloaders.You still haven’t figured out how to get musicians paid. They can have a million downloads on MySpace, and as Andrew Keen aptly points out, not even get enough revenue from the occasional gig to keep the band in pizza.Web 2.0 went astray by constantly trying to sever the metaphor between atoms and bytes, and not protecting that integrity to enable monetarization.

          2. fredwilson

            i’m an investor (personally) in a friend’s record label. he has broken four artists in the past year and each one was broken in the digital media and the vast majority of the music he sells is on iTunes (bits not atoms).artists are making a living selling bits, just not artists who are managed by luddites and are partnered with old school labels

          3. Prokofy

            Well, did you get a return on your investment yet, Fred?I’m glad your friend’s record label works. But that’s not a model for every musician, and there are zillions of them. Did your friend’s four artists give away all their music and then also try to tell on i-tunes on top of that? Well, why not sell on i-Tunes *more* then? Selling on i-Tunes is a good thing, as you’re acknowledging, but then why the millions of free downloads on MySpace first? Andrew Keen has simply found four *other* artists that *didn’t* make it work this way — anecdotes are not the way to have this discussion, however, because it’s not scientific.For example, Magnatune was touted by Wired and other Web 2.0 gurus as being the way to go. Well, how are they doing? Why do we never hear about them any more? They have everybody give away their music, and then broker licenses for those companies that want to buy a jingle for an ad or a sound track for a movie. But what a system, eh? It counts on capitalists being able to hang around forever to sustain your socialism, when you’re constantly downgrading the value — to zero. Why even pay for a licensed jingle when soon you can get that for free too? And what kind of cut does this Magnatune have to chomp out of your profits to sustain itself to do brokering? And so on.No one has proved that endless liberalization of content and giving away everything for free really works as a viable business model except for a few that…often have angel investors in the wings with an ideology that giving away things for free works — or who make their money giving lectures about how giving away things for free works at Web 2.0 conferences.To really work, this concept has to work for LOTS more people.I don’t feel the need to bash record companies that everyone feels simply because I have a longer memory, I guess, as I know that they sustained artists for decades, and whatever the belief in their rapaciousness, probably we could not have had the music culture we have today without them. Like building national roads and parks.Luddites are not people who were against technology just for the sake of being backward and unprogressive; they were against technology that cost them *their living*, and which enabled *other people* to take their living away from them and impoverish them.

          4. fredwilson

            There’s always going to be ten times as many artists (maybe a hundred or a thousand artists) who didn’t make it for every one who did. That’s going to be the case regardless of the distribution technology and the business models.Giving music away on the net to stimulate demand is the new radio. Artists weren’t paid by terrestrial radio and had a compulsory licenseAs terrestrial radio declines, free internet streams are simply replacing it. That’s the new promotionMy friend’s record label is not proof, but it is an example

          5. Prokofy

            I look forward to hearing whether this venture is profitable for you down the line, Fred.I’m not talking about the ratio of attempts at success to actual success. I think you’d agree that this coefficient is wildly different than it was in 1960 or 1980 because of the ease which which you can use Youtube or Myspace or whatever to get attention to your music and find fans. So there’s really no comparison. But aside from those jillion wannabees, I mean to focus on those that really are the rough equivalent of someone worthy to be signed by the old-fashioned record label you hate. Such a band could have a million downloads on myspace without a penny in their pot — and that’s wrong. That’s not about “some not making it” — that’s about valuation, and the inability of the engineers of the web — for ideological reasons — not to deliver micropayment and other schemes to get people paid. I’m totally with Jarod Lanier on this when he says “it is so because we engineered it to be that way” — and they didn’t have to.You’ve made an interesting and valid point about “the new radio” as far as giving away content except your metaphor flounders when you consider that radios ran ads to get their radio station paid for, and to get the DJs salaries. And of course there was payola. Now, while there’s some blogola, again, there’s just no equivalent. Radio was a business model with a payment scheme as it “gave away” music and taping from radio was always inadequate and flimsy. Internet giveaway has no business plan! Call it “the new promotion” if you will and feel triumphalist that evil record companies and the RIAA boo hiss are all undermined by freebies but…who is really getting paid? Across the board? To make the entire industry thrive? It’s all just kind of busking, really.I noticed YouTube now puts up a VISA interface in front of your video to get you to pay, and on mash-ups, picks out one of the artists for you to go buy the song. I wonder if that works. I hope it does.The web has to get people paid, other wide it is an engine of destruction of value. In fact, it already is that, and only handfuls of people are profiting from services that are free to users.You’re one of them.I’m not.

          6. fredwilson

            I’m working on it. Targetspot is the leading audio ad network for streaming audio and I think there is great potential for monetization and royalties to artists. The radio model can work online. It just needs to be more personal, more relevant, and available on mobile devices

          7. ErikSchwartz

            The reason the music industry is so capital inefficient is their data sucks.The model historically has been throw small advances to a lot of bands to get them to sign their lives to a label. Throw them all against the wall, the ones that seem a bit sticky you market the heck out of. A few of those will succeed and pay for the whole thing. The income from those successful bands needs to amortize all the failures. This process created a lot of jobs at labels.The reason the bands agreed to this were two fold; first of all $50K seems like a lot of money to a bunch of 19 year olds who work at Mickey-D’s and live with their parents. Second, the upfront costs of production and distribution used to be prohibitive. Production and distribution costs are way down. There are no returns and cut outs any more. Do end cap promotions exist anywhere but walmart at this point?The real problem is the feedback loop to understand success is awful. The only way to find the gems was to polish a lot of pebbles most of which are junk. That is really expensive.So in the new world. People will pay for music, just not as much as they used to. Bands won’t get upfront signing money. If you can programmatically identify the hits then those successes can stand on their own because you are no longer needing to amortize the failures into the mix. What percentage of bands ever earn back their advance?The recording industry ends up much smaller. The labels in their role as music VC’s go away. There job becomes to monetize and curate their archive.ASCAP, BMI, SESAC and Harry Fox are actually in an interesting place if they play their cards right.

    2. Jean-François Noël

      I see bandwidth as the major point here too, but I don’t share your pessimism about the augmentation of bandwidth capacity. Even you mention backbone as being the easiest to upgrade, this is the most expensive to use for ISP, the bandwidth inside their network is costly to install, but virtually free to use afterward. So the big problem for ISP is mostly their traffic usage on the cloud, but you eloquently said it is going to be almost free for them to cache aggressively the most popular content in their network. This takes care of all wired connections and wifi, but not wireless.Currently many wireless effort are concerned with range to limit the number of towers install. This means more people have to share the spectrum for every tower and the limitations are severe. I think it is safe to assume wireless bandwidth will stay a premium offering for quite longer. At some point the range concern will fade and someone will instead multiply the number of towers to limit the number of simultaneous users.What I’m saying is you are right, but only for so long.

      1. ErikSchwartz

        You’re inventing a really expensive solution to a problem that doesn’t exist in a world a cheap local storage.Save streaming bandwidth for ephemeral content. News. Sports. Live events. Stuff like that.

        1. Jean-François Noël

          Possibly. I think this comes from the fact I use too many different computers and get easily frustrated when one doesn’t have what I need. Even if I have all my data locally stored, I would stream it to where I am right now, so why store it at all? Having everything on my person could solve this, but then sync would have to be seamless too. This could work too.

          1. ErikSchwartz

            I was actually just confronted with this situation.We’ve got a house out on Cape Cod. All my media is ripped to a 1T drive that is hooked up to my machine in my home office. The easiest and cheapest way for me to get my media to the cape for the summer was to clone my external 1T media drive I and leave it up at the Cape.Bandwidth on the Cape is dodgy, especially when Nic across the street has got his torrents running.(The other great old saying was “never underestimate the bandwidth of a station wagon filled with magnetic tape”)

          2. fredwilson

            my approach to this problem is to have two music servers, one at home and one at the beach. the one at home is the master and the one at home is the slave. they sync over the public internet

          3. ErikSchwartz

            So you are already pretty much doing what I propose. Multiple copies of large cloned local storage.Now what is on your disks is largely different than what’s on my disks. But as the disks get larger and larger the differences become less and less as the libraries have more and more common points. At the limit where the size of the disk exceeds the size of all recorded music, the libraries become the complete recorded oeuvre of mankind (yours and mine and everyone elses) become identical.That’s the dream, and we’re only a few years from its technical feasibility.

          4. fredwilson

            I don’t think its a technical issue, I think its a behavior and business model issue

          5. ErikSchwartz

            I agree it’s not a technical issue. It’s about user experience and economics. For instance, 5 years from now you may have the choice between two competing in-car music services.Both services offer every song in every genre.Both services scrobble everything.Both services offer community recommendations and play list sharing from in, and beyond your peer group.Service A works everywhere. Cross country road trips. Tunnels. The far woods. It never hiccups.Service B has geographic limitations. If you drive cross country there are long stretches where it does not work at all. If you’re in a really crowded area (the LIE) there are also hiccups when there are a lot of other users in your area.Service B will add new music a day or two before service A.Service A costs $150 and a monthly charge of $5.Service B costs $125 and a monthly charge of $30 and a two year contract with a cell carrier.The consumer doesn’t care where the music is stored. They care about UX and price. UX can be identical. Local storage will beat streaming in price.

          6. fredwilson

            And guess who is going to offer service a?

          7. ErikSchwartz

            I think it’s wide open. Lots of people are so intellectually invested in the technology idea of streaming (just read the comment threads here) that they’re not thinking of different solutions that give the same or better UX with better economics.

  4. Michael Fidler

    As far as I’m concerned it’s a done deal. It’s been years since I’ve listened to my music collection, and a year for my DVD’s. Other than Blu-Ray, I consume all my media online now. It might seem unreasonable to compare my use with others’ but I’ve found myself to be a good barometer for what’s to come. However, it will take time for most people to get comfortable with this change, but they have some time to make the adjustment. As you mention, more bandwidth needs to come on-line, and companies like Microsoft, Apple, and Google are years away from rolling out an OS designed to fully take advantage of the cloud. Yesterday I posed this question on Louis Gray’s blog;”Is this the last OS we’re going to see; at least in terms of the way we’ve come to know them. I think we’re obviously heading towards a cloud based system. Will it be a thin client, which is partially embedded, (instant on finally) that downloads components as they are needed. Or, will there be an intermediate OS, which bridges the best of the desktop OS with a cloud OS?”I believe that companies like the ones just mentioned, along with others such as Amazon, and Netflix’s; will play a large role in how we consume media, based on the choices offered to us by them. It might be five years away, but the transition has already begun.Enjoy the rest of the weekend & have a wonderful Memorial Day!

    1. theschnaz

      @MichaelI agree. I’ve been moving more and more towards consuming media from the cloud (hulu, imeem, youtube, lala, and amazon.) I can’t recall the last time I bought a cd (for myself) or rented a dvd.I had a desktop computer, which held all of my mp3s. It died and now I can’t access them. Now I’m down to a laptop but I see no reason to repurchase those songs. For the most part works for me. They sell internet-only songs for less than $1. The price is right!I don’t use an iPod anymore, so I have little need for local media. The personal media players, iPods, zunes, psps, nintendo ds, etc., will be a big hurdle for cloud-based media to overcome, but it will happen.Greg

      1. Michael Fidler

        Thanks for the response Greg! I’ll have to check out For now, Blip.Fm, LastFm, and are doing the job nicely.

    2. Nigel Walsh

      michael – Im with you. I donated all my CD’s to Charity and sold all my DVD’s bar a few about 3 months ago. We moved house 2 years ago and all our music is digital and pumped around the house with Sonos. I actually found now we are digital, I watch less real TV and more stuff as I want – and usually in big chunks, nothing for days/weeks then 3 episodes of 24 or Lost …

      1. fredwilson

        We did the same things and have experienced the same behavior as a result

  5. LukeH

    To me it’s all based on utility and seemlessness. The social aspect is a nice to have, but doesn’t drive my behavior much. Currently my file-based iPod gives me the most utility when I’m mobile – planes, cars, exercising. That will change when mobile broadband is faster and non-limited. I also don’t use streaming apps much at work due to the restrictions most companies have on network usage. But if I’m in my home office I tend to use, hypem, and tumblr for audio streaming. I’m also starting to use cloud based apps for more home theater situations – netflix,, hulu. It’s only a matter of time until those streaming apps cover nearly all of my media consumption.

  6. loupaglia

    I think cloud will take hold first as it is already but as costs stabilize, it will be a solid combination of both. There is no reason, your social experience and related media can’t sit in the cloud with pointers to either the media locations at third party storage services or if you are set up with local storage (say in-house NAS type equipment), that those same pointers can’t point back to hardware at your house but connected to the web. For example, we see already a logical movement of where people are going to be able to hook their own hardware and storage space to be available and leverage-able to cloud-based services.Again, this will simply be a balance of the costs of bandwidth (both up and down) and the cost of storage overall. I just don’t see how there won’t be a healthy combination of both.

    1. Lucas

      You’re dead on. What we really need is a platform that can join up the media on your local machine, your external HDD plugged in to your xbox, external sources like Hype Machine or Amie Street, and the same for all of your friends. Universal content resolution gets around the major hurdle of any music service which is we cannot assume an infinite catalog with access rights specified by each user unless we spend millions of dollars and years negotiating agreements with individual providers. This is why Playdar ( is so exciting. It has the potential to solve that resolution problem on the technical front.The next problem to solve is the monetization/business end. What can we come up with that is simple enough and still takes care of the obligations of the content providers? Should there be a master resolver out there that any start-up can come to, pay a fixed fee per stream/user/etc, and the resolver will handle all of the reporting and payments necessary to make this work? That way, start-ups will be able to innovate on the front end, have a business plan that’s easily visible as viable in the very short term, and not have to worry about all of the licensing minutiae.

  7. Jeff Jarvis

    I thikn your friend is making a corollary argument to what I’ve said: The internet IS the social network. The winner in social will not be another network we join but he or she who brings that existing social network of the internet the elegant organization Mark Zuckerberg talks about. That is the meta layer – or stack, as you said in the videos Seth put up – where the greatest value lies. That allows us to have our stuff anywhere and still to organize ourselves through that stuff.

  8. MissXu

    Hmm, great post. Interesting questions. There are 2 things going on here. One is technical. The other is the bigger issue.To address the technical portion… straight storage of media by owners/companies running ‘the cloud’ would be expensive – always more expensive (2x more the last time I checked) than the consumer pricing per Tb of data since we have to factor in management costs for reliability to however many 9s. And storage on local drives might not give people the immediacy to other people’s ‘stuff’ as they supposedly want (or the stability for that matter, you’d know what I’m talking about if you’ve ever had a HD crash!).There are ways of getting around this. Folks like Riverbed (and others – I have no stake in these guys btw) have done some pretty cool caching technology to look at content that is actually accessed/needed so the system can be optimised. This can be looped with consumer friendly storage policies + appropriate pricing policies to minimise overhead costs for both consumer + provider (Kodak Gallery is a top of mind example).That said, the location of the files/data/whatever is an implementation detail.The bigger issue lies with creating a successful service. If this service is one that relies on social behaviour, it has to allow its users to interact gracefully with the system and with each other. Depending on the type of service, the target demographic of users, and the main goals of the service, this experience differs.It’s also arguable that we don’t really need to interact with each other synchronously to encourage social interaction + community. There is something beautiful in finding goodies that have been left when we are away to know that sometimes, even when we can not be physically present, we are still ‘there’. The rest time in between the interactions is as important as the interactions.To build a great social service / product / offering, it’s a balance of user experience and company goals. The best UX keeps a company afloat/thriving when there is a market / product roadmap / pricing strategy to back it up.And if we want to know more about how our users might interact, contextual ethnography, sociology, + anthropology goes a long way in providing insight.M

  9. ErikSchwartz

    Big storage is a paradigm changer. When you can easily afford enough local storage to store everything then replicating that everywhere becomes an incredibly efficient way to make it available.CDDB has 7 million unique waveform fingerprints. 7 million tracks fits on ~25T of storage.25T of storage is well under 2 orders of magnitude from where we are now. In 3 years 25T of storage will be commonplace and cheap.

  10. MattCope

    It’s all well and good to give Flickr posters credit.But don’t do it because a few jerks were out there criticizing you.You’ve got too large of a community of *constructive* critics to let a handful of assholes influence your behavior.

    1. fredwilson

      true, but i think it was a valid critique

  11. Lloyd Fassett

    I think you and your friend are both right and the question is a red herring. Where files are stored and how they are delivered is a back-end issue. The economics of the cloud will pull it’s use into the mainstream in the same way A/C electricity begat centralized power generation (The Big Switch / Carr). The purpose of listening is I think the issue. I don’t know how you got through the post without the word ‘copywrite’ because that’s still the main issue.I think when you wrote “My gut instinct is that we need both the media and the data in the cloud.” you meant ‘accessible’ instead of ‘in the cloud’. Yes that’s the ultimate.It also seems this is the same as multiple ways to use a search engine, the two leading ways to be discovery and navigation. Comments and access are for discovery and local file access is for utility, just like LukeH’s comment about having files on an iPod.I agree with Jeff Jarvis about ‘new stack’ too and look forward to Disqus being a link for comments between multiple platforms to fill that role within commenting systems.The recommendation engine though, is to me a meatier part of your post. Now we have a lot of media, an emerging concept of a ‘stack / meta data’ to link them, the recommendation engine is something I’d get real utility out of. Personally, I get enough out of iTunes Genius because my music budget is file sharing and when something is discovered I want to navigate back to it. The service is kind of like a local / cloud hybrid too becaus it does recommend music in both places with an offer to pay for the songs from their ‘cloud’.You are right in your assessment months ago that it’s not so great, but you are much more a pro at depth and interest of listening and I have something to discover in my 8,000 files. I can see where you would prefer or both deeper and wider services. The key point anyway is better discovery through a lot of data and a discovery process is a replacement for previous navigation processes. This seems the same as the crumbing of the newspapers is in part because they were a navigator to content they don’t own.Today, copywrite holders do hold the upper hand but eventually the social power behind open cotent will prevail. It could take years though. I wonder if it will be like your post of investing in the free-pc business where you thought it was ahead of it’s time.And I enjoy your channel too when I come by.

    1. fredwilson

      thanks Lloyd. is so easy because it’s just a stream of the songs i post every day on tumblr.

  12. Jason

    one thing the cloud does provide us with now (unless you have access to an oc192 + cdn of your own) is virality. a video on youtube and other video streaming services can be distributed to the masses instantly. thank you cdn technology!

    1. fredwilson

      exactly. this is the point i was trying to make

  13. Chris Phenner

    I have almost 30gb uploaded into, I stream music far more than I download it, and rely on a combination of Rhapsody, imeem and Pandora for music discovery — I love the advantages of cloud-based access as a consumer.But in thinking through next-generation services as a professional who works for a digital media firm that serves over 2,000 mobile handset types, I see more complexity than ‘cloud vs. local’ in terms of providing service that people can rely upon. I think a blend of local+cloud is already finding its way into several of the best services, music and otherwise.SugarSync keeps locally-stored files on PCs sync’d in the cloud, but ensures my latest document versions are already on my PC as I work on airplanes. Google Gears/Sync offers a similar hybrid. Delivering ringtones locally to handsets is a requirement, but as consumers change or lose handsets, it’s not just a requirement to have a back-up copy, but that said back-up copy is properly trans-coded and delivered to the new handset, and the file type may well be different on the newer device. Slacker provides ‘streaming’ radio, but caches files on an ephemeral basis for disconnected listening. Amazon’s WhisperSync delivers local content copies with sync that’s aware of where readers last stopped reading (and so on).I think the most current and most useful definition of ‘cloud-based’ is a service that is aware of the device and context in which content is being accessed, and that doesn’t insist on connectivity being available. If consumers seek the immediate gratification of a lower bit-rate file on their handsets for (i) faster downloads and (ii) offline listening, services need to go there. But if a ‘permanent,’ more pristine copy is sought for one’s PC-based or cloud-based storage, that’s a likely value-add.’Cloud’ translates for me into ‘reliable,’ which means that regardless of (i) device capabilities, (ii) connectivity or (iii) life’s unexpected events (e.g., device loss, hard drive crashes), a valuable service is there to serve regardless.

  14. Ilia Uvarov

    I was an avid user of, Pandora and hype machine, but since my move to UK 7 months ago I have given up on all of that. Why? Check out It pretty much does everything Fred has on his wish list for a social media music service/app. Yes, you can create and share playlists, find new music, etc. What’s even more important, you can replay the same music over and over. Essentially, it is just like Rhapsody only it’s free (obviously, ad supported).

    1. fredwilson

      i used to use spotify when it was in private beta. i stopped because it didn’t have all the music i wanted to listen to in it. i guess it does now because i’m hearing a lot of people talking it up now.

  15. niallsmart

    I think the importance of physical (or file based) media depends on the context. For personal consumption, I think it’s essential. There are so many contexts in which you listen to music but don’t have connectivity to the cloud (commuting, holidays, traveling, gym, etc).In the context of a social conversation, the location of the media can be a lot more fluid. Clearly it’s more bandwidth efficient to just push the metadata – the problem is connecting the dots to a reference (streaming?) source for the other participants in the conversation.Architecturally it seems this model requires two layers – the top layer being the various consumer-facing music discovery / recommendation / discussion services that drive and facilitate the conversation (Tumblr, Blip.FM, Hype Machine, individual bloggers, etc) the second being industry supported brokerage service(s) that power the actual media streaming (scope: resolution, semantic search, content streaming, catalog maintenance, metrics, infrastructure, licensing, etc)The obvious problem is that this second infrastructure layer doesn’t yet exist 🙂 (And in that absence: SeeqPod emerged). iMeem seems to be evolving towards being this legitimate streaming provider (but it is not quite an open platform, and its public APIs don’t seem rich enough yet). Perhaps Playlist’s acquisition of the TotalMusic assets indicates they are also moving that way. Alternatively, P2P driven engines and ad-hoc content resolution services like Playdar will mature. Who knows.Until this platform emerges, it seems the consumer-facing services will have to continue to manage both media + metadata, with all the associated bandwidth + legal + other headaches.

  16. LukeG

    Song and pictures are both good examples of media as social objects, right? When I think of all of the interaction that occurs around a given song (say – the listens, the discussions, the sharing, the relationships it affects – it seems to me that it’s a web around a single object, not around the millions of copies floating about in the cloud and stored locally on people’s machines. The social object itself, the heart of that network, is a halfway interesting example of a platonic ideal.I don’t know that the relationships surrounding a given song (or picture) are more valuable than the song itself, but I do think that there is massive value in the dynamic, emergent context in which these objects live. History seems to suggest that the new walled gardens surrounding music in the cloud, for example, will eventually be rendered obsolete by a more open platform. Services fight to make *their* copy of a song the dominant one; the closer to truly open we get, the closer we are to a single shared song.I don’t have the faintest what it will look like, but I’m pretty fascinated to see how the networks around these types of objects will be illuminated. In some senses they are actually more interesting than our current person-centric social networks because of the greater variety of node types. As amazing as a song can be, alone, I think that it’s becoming easier to recognize that the social object is perhaps much greater than the original object itself.

  17. csertoglu

    This really depends on how the content is managed. In P2P networks, the media is local but accessible. The cloud includes part of your lcoal environment, as well. My guess is there will be further integration of local enviornments to the cloud.

    1. fredwilson

      Good point about p2p. But you can’t get instant play in p2p

  18. RichardF

    I use Orb which streams video/music/video from my home media server (a pc!) to my iphone or my laptop. I can also share that media if I so wish via my Orb public website. So to answer your question I don’t think both need to be in the cloud.I like the fact that I store and own my media and am free to do what I want with it and I think a lot of people will feel that way. It is basic human nature.The blended solution is the way forward for now. Maybe when we all have instantaneous access to always on high bandwidth and a storage solution in the cloud that you can trust (not to go bust, trawl your data or prevent you from deleting it ) then that’s probably the way forward.The start of the blended solution exists already with the likes of Orb and the Pogoplug being just two solutions you could choose from others and it will not be too long before it is in the mainstream rather than just early adopters.

  19. bijan

    you are absolutely right.local media = an isolation chamber.that’s how i describe my photos in iphoto.on flickr and tumblr they are freed up and so much more is possible.same with my blog.same with music, video & in the cloud solves other problems like sync, backup, search, etc.

  20. hannahh kelly

    Nice post. I really liked it. I came across another article similar to your post. Would like you to go through it once.Hope you will like it -…

  21. bankownedbyamerica1

    When I was counseling an acquaintance who had just been laid off, I said, “one has to have a back up plan.” She poignantly said, “that job was my back up plan.” I don’t mean to be flippant here but competition is now rampant in America, whether for jobs or even for a comfortable retirement. That is what the bankster scandal was all about, see an opportunity (streets paved in gold) take it. No matter what! As we haveinteresting …Facebook worth $40 billion? Twitter Worth $15 billion?

  22. hubrt

    The case for storage: A couple of years ago, I used to scare music execs at Popkom with this calculation. – Pretty soonish, a top of the line hard disk will offer you something like half a petabyte (or 500 TB) of storage. – You could preinstall every song ever recorded. And would still have left enough space to record your monthly dose of HD tv (plus any bloated OS update you want or need)Storage is getting exponentially cheap. And for the next foreseeable future, portable storage is, well, anyway the way to go. I like the radio app on the iphone (despite the huge gaps in its cloud based repertoire, despite its battery guzzling usage of wireless). Thing is: network coverage won’t be anytime soon near the point where you can convert any given music player into a cloud accessing device. And don’t try to subsidize the infrastructure invest by milking music: at their best time, the EBIT of a single carrier like vodafone surpassed the complete turnover of the whole music industry.And, not to forget: the carbon footprint of permanent cloud access must look insanely scary (compared to local plays).But still: the case for cloud accessibility is pretty valid, too. Cutting off the social from media will only hurt media. Because media tends to talk the value of creative talk. But it is defined (and monetized) by its use of the technical representation of creation, the technical medium. The new medium is not bits (been there, done that, has been called CD), but networks (technically and socially, leveraging technology). Non-networkable media will look like pre-Gutenberg books: nice scribblings, no future.I guess we should be talking hybrid. Play local – point to the cloud. Where you can play, too, of course – but how will I fill up my 500 TB HD?

    1. fredwilson

      Ooh. Good line. The new media is not bits, its networksThat’s gonna get reblogged when I get to a computer!

  23. Michael Fidler
    1. fredwilson

      There are so many great web services for music lovers out there

  24. hypermark

    In social media, the conversation is endemic to the medium. My expectation is that to the extent that we agree that media of all kinds (music, movies, photos, posts, threads) will continue to be conversational, the underyling media needs to be reliably accessible.Whether that becomes an iTunes ala carte subscription or something akin to basic cable (i.e., buried in your DSL or cable modem service cost), it seems that this kind of Library of The Commons is inevitable and necessary.

    1. fredwilson

      A library of the commons is what we need. And a business model for those that contribute to it

  25. Bruce Warila

    It seems as though energy consumption has to be an important part of this debate. Smart metering of energy is going to tell us how to consume less of it, and access to renewable energy to power our own local ‘kits’ will probably factor into the decision. Otherwise, it seems as though the ‘local’ option (your own on-demand computing and storage) * 300,000,000 people versus solar charging a smart phone and accessing what you need when you need it, is going to be better for the planet? It will be interesting to see over the next ten years how tax policy, energy costs and energy infrastructure improvements shape this debate.

    1. fredwilson

      Good points