I tried to pick four trends and put them in a theme, and the theme real­ly is about people. 

People vs Celebrities

Rudolph Valentino in The Sheik

So the first trend is real­ly about celebri­ty, and if we look at this guy, he was the first real movie star. When he died in 1926, there was mass hys­te­ria and 300 peo­ple were injured at his funer­al. He was the pro­to­type of what it is to be celebri­ty today, and this was just when movies had real­ly been invent­ed. So in 3,000 years of act­ing, this guy sud­den­ly was more famous than any oth­er actor has been in his­to­ry. The rea­son for that is kind of obvi­ous. He can be seen by mil­lions of peo­ple in lots of movie the­aters, where­as most peo­ple can only be seen pre­vi­ous­ly in one theater. 

But then the Web comes along and this idea takes root about the Long Tail, and it’s say­ing exact­ly the oppo­site. It’s say­ing that the lit­tle guys win, not the big guys, not the hits, not the celebri­ties. And this is the ulti­mate cliché slide in an Internet talk, to bring up the Long Tail, but the rea­son I’m bring­ing it up again is that I think it’s over. And that’s because it only went half-way. It looked at real­ly what hap­pens when you get a shop like Amazon and they stock every­thing. So of course you’ve got a niche mar­ket prod­ucts and the niche stuff sells, and before you had a shop and they could­n’t stock every­thing so the niche stuff did­n’t real­ly sell.

So now you have infi­nite sup­ply and that’s real­ly what the Long Tail’s say­ing. It stops there. But celebri­ty does the exact oppo­site. When you’ve got infi­nite sup­ply, the finite com­mod­i­ty is the com­plete reverse. The scarci­ty is demand. And to cap­ture peo­ple’s atten­tion, which is the new scarce thing, you need peo­ple who can cap­ture atten­tion. And think about the peo­ple who can cap­ture atten­tion. They’re celebrities.

And so some actu­al con­crete exam­ples of how this affects us, the celebri­ty effect is the way we’re made. When peo­ple looked at MySpace, it was very dif­fer­ent from oth­er types of social net­work because some­thing new hap­pened, and they found that some of the nodes were not the same as the nodes on oth­er social net­works. There were actu­al­ly singers, basically.

And just like most song­birds are male and they will sing them­selves to death to attract females and to show how good they are at col­lect­ing food, that they can afford to spend extra time singing, rock stars real­ly are genet­i­cal­ly what song­birds are. So when you have a new­er social net­work like Twitter, Twitter has an asym­met­ric mod­el that’s dif­fer­ent from Facebook. If I fol­low Lady Gaga, she’s not going to fol­low me. So what that means is that every­one fol­lows Lady Gaga and you get this tremen­dous effect of celebri­ty, and you’ve got the fact that these racks and racks of servers… You can imag­ine a giant serv­er farm for Twitter, 3% of those racks are deal­ing with Justin Bieber.

“At any moment, Justin Bieber uses 3% of our infra­struc­ture. Racks of servers are ded­i­cat­ed to him. — A guy who works at Twitter

— dustin cur­tis (@dcurtis) September 7, 2010

That allows the sys­tem to be gamed, and it has­n’t real­ly been gamed yet, but peo­ple will learn how to game it. They will learn how to take what we’re cal­i­brat­ed to do when we evolved in the African savan­na walk­ing around in bands of 50 peo­ple, where the lead­ers of those peo­ple, the celebri­ties, were only com­mu­ni­cat­ing with 50 peo­ple. Now they can com­mu­ni­cate with a bil­lion peo­ple, so our risk mod­el, the way we’re cal­i­brat­ed, is flawed, and that can be exploit­ed. And that’s what mar­ket­ing is, really.

So as an exer­cise to look at this in terms of num­bers, Lady Gaga’s been very good at exploit­ing fame on the Internet. She’s not ​“Charlie Bit My Finger” or some­thing like that, she’s a real celebri­ty. And if you look at how much she’s been down­loaded on YouTube and you aver­age say, it’s 5 megs a video or what­ev­er, that equates [to] about 10 petabytes. Now let’s say we take that and we look at it charged the same as SMS which can be 15 cents for 140 bits, how much would Lady Gaga have to pay Google if she did actu­al­ly pay for the band­width that she used? Well, the answer is $10.5 tril­lion. It’s a very big num­ber. So the Internet’s a giant game of fol­low the leader in that sense. And the Long Tail starts to reverse as mar­ket­ing takes over. 

People vs Robots

The sec­ond trend, real­ly big trend, is peo­ple vs. robots. And robots here are rep­re­sent­ed by Google. Google is quite a nerdy com­pa­ny, and it delib­er­ate­ly tries to be nerdy. That’s their image. They try and take the algo­rithm and make it some­thing that’s very core to what they do. For exam­ple, when I was work­ing on a new search engine, what Google did was they decid­ed they were going to copy it but they weren’t going to use edi­tors, and they spent an enor­mous amount of mon­ey to try and make a sys­tem that had no edi­tors that was just done by machines.

You’ve seen a mar­ket­ing ploy recent­ly. They cre­at­ed a car with­out a dri­ver. This is a dumb idea. Okay, it’s a mar­ket­ing thing, but it’s sym­bol­ic of Google [hav­ing] an ide­o­log­i­cal attach­ment to the algorithm.

So when you ask Google a very sim­ple ques­tion like, ​“Can you rec­om­mend me a good sushi bar?” You get basi­cal­ly human beings put through a meat grinder and an opin­ion churned out that—this might be via Yelp or might be whatever—but essen­tial­ly in the search results you’re get­ting an algo­rith­mic result. You’re get­ting some­thing that ulti­mate­ly dis­tills some­thing down to a number.

There’s a big exis­ten­tial threat for Google, which is Facebook. Facebook is about peo­ple, and it’s about a social net­work. And if you ask Facebook (you can­not do this at the moment; they don’t have a rev­enue mod­el that’s based around this) but it’s poten­tial­ly much more valu­able… Basically, you could ask Facebook, ​“Can you rec­om­mend me a good sushi bar?” I have a good friend who’s obsessed with sushi. If I ask him, he’ll tell me and it’s a nuanced opin­ion. I know exact­ly what he’s talk­ing about and frankly it’s not an algo­rith­mic result. It’s not fifty frat boys who’ve reviewed a piz­za joint and I don’t real­ly know what it’s about. It can be much more valu­able. It’s about real people.

And the rea­son that’s impor­tant is because I have a rela­tion­ship with these peo­ple that’s sep­a­rate, that’s offline, it becomes unspam­ma­ble. Google know this. They’re spend­ing mil­lions on it. They’re spend­ing mil­lions on maps and local. But they’re stub­born and they don’t own the social graph. So when they put blend­ed results through local results at the moment, because they don’t own the social graph, they’re kind of screwed. So the trend that I’m pick­ing out here is that I think rec­om­men­da­tions are start­ing to go to real peo­ple. The algo­rith­m’s going to have a tem­po­rary respite.

People pow­ered design

The third trend is a bit more abstract but it’s some­thing I know about because I used to be an archi­tect. It’s about how design for per­son­al tech­nol­o­gy actu­al­ly changes things and makes them bet­ter. This quote’s from Andy Warhol. He was look­ing at America and say­ing America’s dif­fer­ent. He’s say­ing, ​“Well, Elizabeth Taylor’s drink­ing Coke and I’m drink­ing Coke and the bum on the street’s drink­ing Coke, and it’s all the same thing.” For the first time in his­to­ry, mass mar­ket cul­ture has allowed us all to enjoy the same thing. This is not cham­pagne. The bum on the street can’t afford champagne.

But con­sumer tech­nol­o­gy goes fur­ther because it allows every­one to drink cham­pagne because of the way it works, and it cre­ates prod­ucts that are actu­al­ly bet­ter than the pro ones. A sim­ple exam­ple, and this is not a new one and it’s kind of been cor­rect­ed, but on Outlook you were search­ing your own email and it would take sec­onds to return the results. Then you search the whole world’s doc­u­ments, 18 bil­lion doc­u­ments, and it comes back in 0.3 sec­onds. And it’s free. This is a bet­ter product.

So why is that? Well, one rea­son is low mar­gin­al cost. When you look at this, this is not to do with the hard­ware, because the Vertu is basi­cal­ly a cheap phone in a gold case, and the iPhone is about the soft­ware. There’s very low mar­gin­al cost. So five hun­dred bucks for mil­lions of peo­ple is a greater amount of mon­ey than twenty-thousand bucks and the world pop­u­la­tion of gangsters.

So how does that mech­a­nism work? Well, from architecture…this is a bit nerdy but I think cru­cial to the way it works, is that an archi­tect fil­ters fea­tures. They take you brief, you say, ​“Oh I want a build­ing and I want a room that does this and this and this,” and the archi­tect throws out some of those fea­tures. And the dif­fer­ence between archi­tec­ture and build­ing is that a build­ing, you just pass the fea­tures to the con­trac­tor and he builds it. There’s no inter­pre­ta­tion of that. And when I went to Silicon Valley I was hor­ri­fied that I realized—I thought, ​“Oh, I’m going to call myself Chief Architect,” and every­one looked at me as if I were a los­er that I want­ed this title because it’s not a stan­dard title. You were called VP of Marketing or CTO if you were doing a startup.

And until recent­ly, soft­ware was­n’t designed. The prod­uct mar­ket­ing guy would pass the stuff straight to the engi­neer­ing depart­ment and they would build it. But the Internet’s chang­ing that. And it’s chang­ing that because peo­ple fun­da­men­tal­ly have more per­son­al­i­ty than cor­po­ra­tions, by def­i­n­i­tion. It leads to bet­ter products.

So that vision on the left is basi­cal­ly the vision we had of com­put­ing that was dri­ven by small busi­ness soft­ware. It’s a near-Victorian work­sta­tion. The vision on the right of course is what we have now. It’s fabulous.

So it’s not about fea­tures. That’s a cheap hi-fi with lots of fea­tures. [slide not vis­i­ble] It’s about design. That hi-fi is 40 years old. It has very few fea­tures, but it sounds great:

Image via Hi-Fi News & Record Reviews

So per­son­al tech­nol­o­gy allows good design to apply to the process itself by sep­a­rat­ing the mission-critical from the per­son­al. The Space Shuttle has a 386 chip. It’s decades old. It’s com­plete­ly bug-free now. But it’s not actu­al­ly use­ful for what peo­ple have to do up there, so they take their own lap­tops up there. That’s fine. They’ve sep­a­rat­ed the mission-critical from their own per­son­al stuff. You’re not going to want to run the life sup­port sys­tem on their lap­top because if you get a Blue Screen of Death, that’s a prob­lem. But, you can run the oth­er stuff that you might be doing up there.

Image via David Galbraith/Flickr

And that hap­pens clos­er to home. If you look here, this is the ATLAS con­trol room at CERN and the guy’s actu­al­ly look­ing at Facebook, prob­a­bly like quite a lot of you now. But it does­n’t mat­ter because they’re sep­a­rat­ing the mission-critical from the per­son­al, and this is dri­ven by design, fundamentally.

Public vs Corporate Networks

The last point is about net­works and the way they’re made, about why the hell is this bizarre thing, this super fab­u­lous science-fiction‑y thing which is hav­ing a video con­fer­ence… This is what was in the film 2001. It’s free. Why the hell is that free, and why when I phone my wife does it cost up to fifty cents a minute? Something is wrong there. 

“The Internet 2003” via The Opte Project

And it’s to do with the way the infra­struc­ture of the Web works. That map there is actu­al­ly a dia­gram of— It’s a bit like the London Underground map. It shows how infor­ma­tion routes but it does­n’t show actu­al dis­tance. Whereas the actu­al infra­struc­ture of the Web looks some­thing like that:

Chris Harrison, Internet Maps, ​“World City-to-City Connections”

And that’s why, large­ly, you can unplug Egypt like you unplug a vac­u­um clean­er. Because there are very few ISPs going there and they can change the rout­ing tables for a very few providers. You can dis­con­nect that. Burma was just dis­con­nect­ed before. Nepal was dis­con­nect­ed before. All for civ­il unrest. That’s why the Chinese gov­ern­ment con­clud­ed it’s not a threat. The Internet’s not a threat to them. For all we talk about and want to believe this pos­i­tive sto­ry that there’s going to be a Twitter rev­o­lu­tion, etc., the rev­o­lu­tion either will or will not work, and yes there will be a role with that. But also remem­ber that the Internet is not FidoNet. It’s not a tru­ly unstop­pable net­work. Governments can unplug it.

So back to the main point which is this play­off between net­works and what it’s going to mean for us, is that the Internet’s like a road net­work. Most of it’s free and you make mon­ey along the side like Google, etc. But the phone sys­tem is like a rail­way. You paid as you used it. And they’re real­ly in con­flict with each oth­er, and the prob­lem is that they run over the same thing. So you’ve basi­cal­ly got a car run­ning on rails. You’ve got some­thing that’s incongruous.

And what’s hap­pen­ing is that’s been ignored. That’s why your Skype cal­l’s free, because it’s been ignored for a while. But video is cre­at­ing a rea­son for this not to be ignored any­more. And it’s a big problem. 

So going back to the Lady Gaga exam­ple, just to show you the dif­fer­ence in video to mes­sag­ing, what is the dif­fer­ence in the val­ue per bit, the rev­enue per bet, that’s gen­er­at­ed? If we say ten bucks CPM impres­sion for a video, and it’s 40 megabits per impres­sion, 140 bits for an SMS mes­sage again (okay, SMS, let’s say ​“tweet” for that), well the rev­enue is about 10^-9 or 10^-10 for the video bits, but for the SMS bits it’s about a thou­sandth of a dol­lar. So you’re talk­ing about a 25 mil­lion dif­fer­ence in the value. 

You’ve basi­cal­ly got bits that are fly­ing around where the car­ri­er is told in net neu­tral­i­ty that they can’t charge a dif­fer­ent amount for the bit, but there’s a 25 mil­lion to 1 ratio dif­fer­ence in the val­ue that there is in that thing. 

So what hap­pens of course is mes­sag­ing, we’ve been fleeced for years. With mes­sag­ing, with SMS, it’s ridicu­lous it costs so much. But at the same time, video is real­ly crip­pling these guys. And they’re not get­ting the rev­enue for it. So you can see both sides of the argu­ment. But it’s a train­wreck. Something real­ly bad is going to hap­pen here because fun­da­men­tal­ly, two busi­ness mod­els are colliding.

So what are the options? You have gov­ern­ment con­trol, so you sub­si­dize the Internet. It does­n’t need to be gov­ern­ment. The gov­ern­ment might super­vise non-profits, etc. or they tell the ISPs what to do. Or you have two Internets, because what will hap­pen is if the video’s swamp­ing the nor­mal Internet, cor­po­rate mes­sage net­works will say, ​“Well, we’ll put just our mes­sag­ing stuff with­out the video stuff and we’ll build a sep­a­rate Internet.” And the third thing is that then they decide, ​“We’re going to pri­or­i­tize bits dif­fer­ent­ly,” in which case net neu­tral­i­ty’s dead, and that’s actu­al­ly a real­ly big deal.

So these things sound very nerdy, but they’re super impor­tant. This is the back­ground of how things like unplug­ging the Internet in Egypt plays out. It’s to do with the way the net­work works. Who owns it? What does it evolve like in future? So this trend seems real­ly banal. It’s the last thing I’m say­ing and it seems real­ly triv­ial to say, ​“Well, there’s no more all-you-can-eat data,” but real­ly that’s what’s impor­tant. We’ve talked about the fact that you can do what­ev­er you want on the Internet and it’s all you want. That’s real­ly impor­tant. So if what’s hap­pen­ing… Most of you if you have a phone you might know that if you start watch­ing video movies on your phone, the band­width starts get­ting throt­tled, it becomes a real­ly real­ly big deal if that plays out into the sphere of what’s hap­pen­ing around the world with not being able to voice data for­ev­er. So, that’s it.