new TLDs could balkanize the Internet

Dave Winer has some cogent critiques of the idea that companies can register their trademarks as new Top Level Domains (TLDs), from an intellectual property perspective. However I think the danger is more pernicious than that – allowing deep-pocketed corporations to create new TLDs at will risks the destruction of the Internet. In a nutshell, why would Google or Microsoft even bother with www.google.com or www.microsoft.com when they can simply use http://home.google or http://home.microsoft? Ultimately you will see entire ecosystems vanish behind these TLD-walled gardens. Forget about gmail.com; now you get redirected to http://mail.google. Take this further: these companies make browsers (Chrome, IE). So now if you’re locked into the walled garden of Gmail anyway and Google says “use Chrome, you don’t have to type http:// anymore” and IE users accessing Gmail see a moderately-degraded experience, then there will be forced migrations to ecosystems that don’t exist right now. Facebook is the worst offender already; imagine if they got into the same game with Opera or even worse allied with Microsoft/IE.

It can get worse. There are numerous limitations and flaws in the HTTP protocol since we have shoehorned all sorts of functionality onto what was originally just a hypertext linking platform. And support for HTTP starts at the browser. Today it’s already hard enough to write webpages for all browsers, and designers can’t code for the latest and greatest CSS/HTML spec and be confident it will Just Work. Imagine if Chrome decides to create a new protocol, g://? shorter, saves you characters on Twitter, built-in URL shortening, and much faster handling of video and pictures. Built right into Chrome! Interoperability between browsers itself is at risk here if the fundamental communication protocol itself starts to fragment; we’ve seen it happen already with HTML and CSS and browsers, but with custom TLDs the incentive to do worse will be irresistible.

The key is the ecosystem. Apps have shown us how companies move away from open protocols like RSS towards custom and closed APIs. TLDs will just accelerate and worsen the trend. Eventually your browser will run heavily customized and feature-extended HTML, with an optimized variant of HTTP that works best with the ecosystem it was designed for (be it Chrome/Android/Google or Facebook/Microsoft/Ie or Apple/Safari/iOS). Try to do anything outside that ecosystem and you’re forced back onto the “old” tools that will be slower and more unpleasant; sure, Hotmail will work on Chrome, but if you use IE it will be so much easier… switch! (to quote the Oracle of Pythia, “All this has happened before. All this will happen again.”)

Remember the old days when if you were on Prodigy or Compuserve, you couldn’t email someone on AOL without a complex extra header? We could be looking at the same thing, with the Internet. We will have to call it the InterIntranet.

Twitter doesn’t innovate

Twitter may be at the peak of it’s innovation. They haven’t really made any substantial improvements in user interface or functionality in a while; hiding replies on user profiles is basically a minor hack. What would be far more useful is marking a user (or hashtag) as “read” temporarily hiding their tweets from your stream (analogous to marking emails as read in your inbox).

But fundamentally, there really isn’t much more Twitter really CAN innovate on. It’s a micro-messaging service. They missed the boat on becoming an identity service; Facebook Connect beat them to it. They seem to have de-emphasized SMS as an interface (at least for the US market) – imagine if they had aimed at taking on BBM and WhatsApp? And they still insist on the 140-char limit, though they could easily allow for a “read more” type extra text (the way it’s done in blog software) or take other simple measures to alleviate the crunch, such as not counting http:// towards the char limit in links, or even allowing links to be metadata the way that photos and video are (the link would still be displayed in-line). You can’t even do simple markup like bold or italic. If links were meta like photos, you could even have a “Recent links” sidebar the way they do for your photos on your profile, but nope.

Twitter has no built-in emoticons, has no elegant way to show a conversation between more than two people, and when you click on the new “view conversation” link, doesn’t show it to you in the original chronological order. Twitter is deprecating RSS which means most bloggers use a plugin to embed tweetstreams; lists are usually not supported.

And of course, as Dave Winer has been saying all along, Twitter isn’t open. You can’t export your data and you can’t really even access older tweets (and did I mention that search is broken?).

It’s also worth pointing out that Twitter’s advantage of the network effect isn’t permanent. Look at what happened to MySpace and FriendFeed. Users will leave if they have a better option; you just need to woo the early adopters like Scoble and make a big splash at SXSW. Plus media attention will be lavish upon any company that has the balls to actually say, “we are out to eat Twitter’s lunch”.

Twitter doesn’t need to be beaten, it just needs to be threatened so it gets out of its comfort zone. Right now it’s chasing after NASCAR and trying to give users tailored content; that’s a fool’s game. Users will never warm to an algorithm’s suggestions – just ask Netflix (or better yet, ask a user).

And Twitter isn’t thinking big at all. What could they achieve if they wanted to? How about aiming for the moon – like becoming a defacto replacement for email?

The end of Facebook? not if it goes Prime

Is Facebook toast? I’m not asking because of it’s IPO, which despite whining from the tech pundits was perfectly calibrated. I’m asking because it’s basic business model is still such a clunker:

Facebook currently derives 82 percent of its revenue from advertising. Most of that is the desultory ticky-tacky kind that litters the right side of people’s Facebook profiles. Some is the kind of sponsorship that promises users further social relationships with companies: a kind of marketing that General Motors just announced it would no longer buy.

Facebook’s answer to its critics is: pay no attention to the carping. Sure, grunt-like advertising produces the overwhelming portion of our $4 billion in revenues; and, yes, on a per-user basis, these revenues are in pretty constant decline, but this stuff is really not what we have in mind. Just wait.

It’s quite a juxtaposition of realities. On the one hand, Facebook is mired in the same relentless downward pressure of falling per-user revenues as the rest of Web-based media. The company makes a pitiful and shrinking $5 per customer per year, which puts it somewhat ahead of the Huffington Post and somewhat behind the New York Times’ digital business. (Here’s the heartbreaking truth about the difference between new media and old: even in the New York Times’ declining traditional business, a subscriber is still worth more than $1,000 a year.) Facebook’s business only grows on the unsustainable basis that it can add new customers at a faster rate than the value of individual customers declines. It is peddling as fast as it can. And the present scenario gets much worse as its users increasingly interact with the social service on mobile devices, because it is vastly harder, on a small screen, to sell ads and profitably monetize users.

The basic problem is that Facebook’s major innovation is to facilitate social interactions, but unless you charge people 1 penny per like you can’t actually monetize those interactions (and any attempts to do so would act like a brake).

But there is an obvious way to monetize Facebook that I am surprised few are talking about. Consider the numbers: Facebook is valued at $100 billion, has about a billion users, so each user is “worth” $100. But Facebook only makes $5/user annually in revenue from ads. So, why not offer users a paid option? If Facebook followed Amazon’s example and offered a “Prime” service, they could charge users $75/year (or $8/month ongoing). In return, that user could get a pile of perks:

  • no ads anywhere, of course
  • free digital gifts and an expanded menu of “pokes” (bring back the sheep!)
  • a “premium” version of the Facebook app with built-in Skype functionality
  • more search filters and automated searches for friends (akin to LinkedIn’s subscriptions)
    the ability to track who views your profile

this is just a basic and obvious list but I am sure there are other perks that could be offered. For example, given that Craig’s List hampers innovation in the classifieds space, Facebook can and should leverage the social graph and offer it’s own (as well as compete with Angie’s List, or buy them outright). Facebook Prime users could be rewarded with better access or free listings.

And then there’s the coupon space – Facebook has all the data it needs to outdo Groupon or LivingSocial. If Facebook acquired the latter in fact it would have a headstart, and again Facebook Prime users would benefit with specialer-than-special offers or early access to deals.

People have already compared Zuckerberg to the next Bezos, but unlike Amazon’s profligate revenue streams, Facebook remains stubbornly focused on one thing. It’s time to diversify and leverage that social data in ways that people actually use. And let the users pay for it!

mind your b’s and K’s: the arcane art of measuring download speeds

I’ve just upgraded to the 30 MB/s internet plan at Charter cable (and added HBO so we can watch Game of Thrones), so here’s the obligatory speedtest results.

It occurs to me that the units for download can be incredibly confusing. Charter advertises the download speed plan using units of Mbps. So, the question naturally arises, how long should it take to download something 18.3 GB in size? (and a related question, if I am downloading something at 300 KB/s, am I getting my max download speed?)

1 GB refers to a gigabyte (10^9 bytes) in this context, since we are talking about file sizes and network speeds. If we were talking about RAM, a GB would actually refer to a gibibyte. However, 1 Mb is a megabit (10^6 bits), not a megabyte (10^6 bytes), because of the small-case b. So 1 Mb is actually 1/8 MB (since there are 8 bits per byte).

So 18.3 GB downloading at 30 Mbps should require:

(size) / (speed) = (time)

(18.3 x 10^9 bytes) / ( (30 x 10^6 bits / sec) x (1 byte / 8 bits) = 18.3 x 10^9 * 8 / 30 x 10^6 = 4880 seconds = 81.3 minutes

Wolfram Alpha gets the answer right, too (and I like teh natural language query – very intuitive).

Now, suppose I’m rocking 300 KB/s according to a certain beta software download client. How am I really doing? The capital B means it is kilobytes, so that’s actually 300 x 10^3 x 8 = 2400 x 10^3 = 2400000 = 2.4 Mbps. Wait, what??

I’m only getting 1/10th my actual download speed for this??

This is why it’s important to do the math. Of course, the download speed may be limited by a lot of other factors, most notably how fast the server at the other end can deliver the data. I clocked almost 40 Mbps doing a speedtest with some local, low-ping server somewhere, but for downloading this big file I’m probably going a lot further and their server has a lot more to do than humor my ping requests. I guess I should be satisfied.

(But, I’m not. grrr….)

renting bytes: the case for digital non-ownership

Cutting off the DRM nose to spite the reader's face?
In the course of my search for free ebook content, I found an advocacy group called Librarians Against DRM. I found the existence of this puzzling, because if not for DRM then the free-lending program of ebooks by libraries wouldn’t exist. In fact note that Macmillan, which is among publishers that refuse to give ebooks to libraries, is one of those moving to DRM-free ebooks. These facts have a powerful relationship to each other that I think is being ignored by most of the DRM activists. The knee-jerk reaction to DRM (it’s always bad! cheer it when it’s gone!”) misses the point on how its absence might have negative consequences of its own.

It’s an article of faith that DRM is bad and that when you buy something, you should own it in the digital realm just as we do in the physical. Ebooks are probably the most vibrant front in this war against Big Content and the End User. And I have to admit that I do prefer it this way, in an entitled sort of way. The idea that I’ve have to pay $9.99 every time I wanted to read my digital copy of Reamde is of course utterly absurd and offensive – I should be able to read it whenever I want, precisely because I paid so much for it. Ditto the MP3’s and videos I buy on Amazon or iTunes.

And yet there’s an assumption here that we do reuse our content. Obviously with music, we do – and that’s facilitated by the low price of the media. MP3 tracks don’t cost $9.99. But what about books and video? How often do we really rewatch or reread? The answer is, it depends. I don’t think I’m ever going to re-read 90% of the books I physically own, and that percentage will only increase for digital copies. For video, especially long series like Game of Thrones or Battlestar Galactica, the main experience is watching it without foreknowledge, and the value of rewatch is low (though there are exceptions, such as Farscape or Firefly). Movies have the lowest rewatch of all, apart from a handful of favorites (Star Wars, LOTR, Princess Bride etc).

It’s worth noting that the rewatch potential is inversely related to the price. But as you move up the chain from music to ebooks to video to movies the production cost of the content also goes up, which is why cost of ownership of that media also increases (obviously, not all of these forms of digital content are truly DRM-free such that we fully own them outright).

So, let’s factor in the rewatch potential, and ask ourselves, is ownership really useful to us? Are we getting our money’s worth? I don’t really think so. As a consumer, what if I had more choice, and paid accordingly?

The scheme I propose is just a starting point for a thought exercise of course, but imagine if (using ebooks as an example) we paid significantly less for “first read” and then paid a little more for each reread until some threshold after which we “unlocked” the content. So instead of paying $9.99, what if I paid only $1.99, which earns me one complete read through. The second read through would cost $1.07, and subsequent read throughs would be $.99 each, until I’ve read the book 9 times at which point the book unlocks to unlimited further reading – and I’ve paid $9.99 total.

The advantage of this is that the barrier to purchasing a book is much lower, such that publishers will see more sales. Not of the individual title, perhaps, but of more titles overall. It wouldn’t be hard to do some numerical estimates based on reasonable assumptions: suppose that number of people purchasing a book scales with price according to a Zipf distribution for example.

The bottom line is that maybe we as consumers should stop focusing on theoretical rights and instead focus on actual expenses and cost-benefit, the same way we do for toilet paper and cereal, when it comes to digital goods. If we think of them as consumables rather than goods, it would be more in line with our actual usage.

the singular implication of uploading one hour every second to @youtube …

This is an astonishing statistic: Youtube users now upload one hour of video every second:

The video (and accompanying website) is actually rather ineffective at really conveying why this number is so astounding. Here’s my take on it:

* assume that the rate of video uploads is constant from here on out. (obviously over-conservative)

* the ratio of “Youtube time” to real time is 1/3600 (there are 3600 seconds in an hour)

* so how long would it take to upload 2,012 years worth of video to Youtube?

Answer: 2012 / 3600 = 0.56 years = 6.7 months = 204 days

Let’s play with this further. Let’s assume civilization is 10,000 years old. it would take 10,000 / 3600 = 33 months to document all of recorded human history on YouTube.

Let’s go further with this: Let’s assume that everyone has an average lifespan of 70 years (note: not life expectancy! human lifespan has been constant for millenia). Let’s also assume that people sleep for roughly one-third of their lives, and that of the remaining two-thirds, only half is “worth documenting”. That’s (70 / 6) / 3600 years = 28.4 hours of data per human being uploaded to YouTube to fully document an average life in extreme detail.

Obviously that number will shrink, as the rate of upload increases. Right now it takes YouTube 28 hours to upload teh equivalent of a single human lifespan; eventually it will be down to 1 hour. And from there, it wil shrink to minutes and even seconds.

If YouTube ever hits, say, the 1 sec = 1 year mark, then that means that the lifespan of all of the 7 billion people alive as of Jan 1st 2012 would require only 37 years of data upload. No, I am not using the word “only” in a sarcastic sense… I assume YT will get to the 1sec/1yr mark in less than ten years, especially if data storage continues to follow it’s own cost curve (we are at 10c per gigabyte for data stored on Amazon’s cloud now).

Another way to think of this is, in 50 years, YouTube will have collected as many hours of video as have passed in human history since the Industrial Revolution. (I’m not going to run the numbers, but that’s my gut feel of the data). These are 1:1 hours, after all – just because one hour of video is uploaded every second, doesn’t mean that the video only took one second to produce – someone, somewhere had to actually record that hour of video in real time).

Think about how much data is in video. Imagine if you could search a video for images, for faces, for sounds, for music, for locations, for weather, the way we search books for text today. And then consider how much of that data is just sitting there in YT’s and Google’s cloud.