broadstuff

With Candide, Voltaire trook several hundred pages to satirise the silliness of unthinking optimism. Gillian Tett took a few hundred words in the FT. She was a young and enthusiastic Euro-optimist, her father was a grizzled Euro-cynic. Here is her apology:

...what is also clear is that my father was entirely correct to be cynical in the face of my optimism. So the question I keep asking myself today is, why did so many people (like me) get swept along by the hype – while others (like my Dad) did not? Why was his cynicism a minority view, in a sea of hope?

One easy explanation, perhaps, might be our relative ages. By the time we had our argument in Val d’Isère, my father was 58 and had lived through numerous economic cycles. I had only worked as an economic journalist for a few years. I had never stared a recession deep enough in the face to become truly wary, rather like those young mortgage traders who stoked up the subprime boom a decade ago.

Then there was the nature of my father’s career. Although he was a brilliant student of economics in the US and UK, he did not go to work in the City or on Wall Street; instead, he headed for the distinctly unglamorous world of British manufacturing, working mostly in midsized companies. So instead of worrying about complex financial products or public policy rhetoric, my father has spent his life thinking about making widgets, such as fire detection equipment.

In that earthy world, there is little place for punditry or wishful thinking: either your sums (and widgets) add up, or they do not. Stock rooms cannot be conjured up by rhetoric alone. Little wonder, then, that my father considered it madness to run a single monetary policy without free labour flows or fiscal union. If the eurozone was a business plan, in his eyes it simply did not add up.

There was also a more subtle factor. Precisely because my father worked in the (then relatively unfashionable) manufacturing world, he was less subsumed by fashionable policy group-think. By 1996, as the financier George Soros says (quoting the psychoanalyst David Tuckett), the euro had become a “fantastic object”: it was unreal but immensely attractive. Or to put it another way, political idealism had subsumed economic gravity. My father, outside the intellectual echo chamber of the eurozone elite, was never infected by the hype.

This sort of explains my complaint from 2006/7, when it was plain that there was a huge problem coming, and nary a peep from the Financial journalism community. I thought they were in the pocket of the Financial industry (as the regulators were), but being mainly "teenage scribblers" (as an earlier Chancellor asserted) who had never seen anything like it is alao a reasonable analysis. But teh real question is "what are the lessons" to avoid unwarranted optimism?

So what are the lessons? One is that it shows the value of retaining older people – or at least, a diversity of life experience – in the workforce. If nothing else, a range of ages helps guard against group-think. Another moral is that it shows the value of having some people from “real” businesses sitting in the corridors of power: handling widgets or running small companies is a good reality check. But perhaps the biggest lesson of all, which journalists, financiers and politicians should all have posted up on their desks, is how wrong we can sometimes be, especially when we are excited by a “cause”.

These lessons are as valid for Technology as Finance, of course - which is why I wrote about it. So if I may abstract, here are...

The 5 Lessons to Guard Against Irrational Optimism.

Run for the hills if:

1. If the issue is a "cause" with a "belief system" rather than based on any hard data
2. If the believers are all young and idealistic, and saying "this time it will be different"
3. If Old Fogeys who have been there before start to get sceptical about it
4. If the numbers don't add up without "heroic" assumptions, and requires everything to go right at all times
5. If the people calling for the New New Thing have never run a business in their lives.

Otherwise, your only hope is to pray "Apres Moi le Deluge"

Two trends...firstly, and predictably, the minute there is a new social currency, there will be the attempts to inflate it for commercial gain - Facebook Likes being the latest example - CNN:

Facebook has begun a purge of fake accounts and "Likes" as part of a set of site improvements announced last month. The result has been lower numbers on fan pages, including some of the site's most popular ones, but no actual loss of real followers.

On August 31, Facebook announced plans for improvements to its "site integrity systems." That, according to a blog post, included increased automated efforts to sniff out and delete fake accounts or Likes that had been illicitly sold or gained by otherwise shady means.

"A Like that doesn't come from someone truly interested in connecting with a Page benefits no one ...," the post read. "When a Page and fan connect on Facebook, we want to ensure that connection involves a real person interested in hearing from a specific Page and engaging with that brand's content."

It was ever thus....but the more worrying trend to me is the increasing belief that systems that measure this are in some way measuring something valuable - Klout has just got significant investment from Microsoft and it will be linked to Bing - TechCrunch:

You’ll begin seeing Klout scores — the combined measure of a person’s influence across Twitter, Facebook and other social networks — show up in the search engine today. The initial implementation will show Klout scores for friends in the “People Who Know” section of the right-hand column, alongside other third parties already in there, including Twitter and Quora. Search for a hot topic like “Facebook advertising”, you’ll see people with socially-proven expertise showing up. Mouse over an expert’s name, and their Klout score will appear, along with their Klout-determined areas of expertise.

There seems to be a mass amnesia about the GIGO concept, in that these reputation systems' parameters are half bake...sorry, "currently imperfect", and are measuring an easily inflatable and unstable social currency, based on the actions of a relatively small number of noisy vessels. Still, no doubt there will soon be a Klout Engine Optimisation industry (if there isn't one already).

Hopefully these Index systems will become more discerning over time, but i don't think that's where the payoff is today. To me its just another example of the online forces inexorably driving Greshams Law of Information - ie bad information is increasingly driving out good on the "Free" Web.

Recent bit of research implies Spotify may be ending its growth run....PaidContent

Music industry analyst Mark Mulligan, presenting at Future Music Forum in Barcelona on Thursday, sounded a note of caution…

“There is a natural ceiling of adoption of the people who are willing to pay $9.99 a month for music they don’t own. If you look at growth from launch, Spotify is – at best – on par with where we should be. The likes of imeem were the future of the music industry once, too. The most Vodafone (UK) got to was about 600,000 customers – Spotify’s (UK) paying subscriber count is about 600,000 to 800,000. This market should be much more dynamic than where we are now. It’s a niche proposition. The majority of mass-market consumers are still not interested in that pricepoint.”

So, is there a natural limit of about 800,000 people who pay for this sort of music? If not, where is it?

At £9.99 a month that's about £120 pa that Spotify charges. When we looked at the UK music industry a few years back (to help a client work out the viability of music on smart mobiles, oddly enough - and I'm too lazy to look up the latest data as I doubt things have changed a lot in numbers or propensity), it bifurcated into 2 markets - "singles" and "albums". These few years back, "singles" buyers were about 5m people in the UK and spent c £12 pa, so this demographic are not Spotify paying customers. There were about 15m "album" buyers who spend c £80pa on average, so the higher spending of them are the potential customers. But £120 pa is c 50% higher than their mean spend of £80, so how many of the 15m will pay it?

Allow me a quick blog-packet calculation. Assuming a standard normal distribution curve (a bell curve), only about 2% of all users lie outside the 2 standard deviation level, and about 16% outside the single standard deviation limit. Making the assumption that a 50% increase in the mean price is outside the 2 standard deviation limit one can estimate that the total UK market for Spotify customers is about 2% of 15m, or c 300,000 people. If its above one standard deviation there are c 16%, or 2.4m people, who are potential customers. This is not to say that all will be customers, of course - converting 50% of all these higher end album buyers to Spotify yields a market ceiling of 1.2m people (remember that Apple and various others are also in the hunt for this spend, never mind the people who just won't want to pay hard cash for music they dont own), at 1 standard deviation. (I find it hard to believe that a 50% higher price point above the mean is much less than 1 standard deviation). If you believe that they can only convert less that 50%, or that the price of £120pa is greater than 1 standard deviation from the £80 mean, then that number starts to fall quite fast (for example, 1.5 standard deviations is a mere 7% or c 1m potential customers in total).

If you assume instead its a power law operating (say the 80/20 spread), then you believe that the top 20% will spend 80% of the money, so about 3m people will spend about £360 a head -so for them, paying for Spotify is a bagatelle. But how many of them wiill spend it? Again, assuming a 50% penetration only yields about 1.5m users. Now to be sure, there will be a few from the next 20%, but not many - power laws recede very fast, the next 20% spend only 12% of the money, so its about a £50 mean spend for this quintile - clearly most of them are not Spotify paid users. Now you can play with various distributions, but the fact remains there are not a lot of people who are going to spend £120 pa on music, no matter which you use, and much fewer of those who will buy it from Spotify (or any of the streamers for that matter, I merely use Spotify as it was the one commented on).

In summary, Mark is probably right, at this price point there is not huge room for further growth.

The other thing that came out of this piece was even more "interesting":

Spotify is having to acquire 1.9 million new customers a month in order to retain 400,000. It’s a huge, huge marketing problem. The average pay TV service would want to see churn rates in the low single-digit percent.

Never mind the cost of new customer capture.....

But to my mind the fundamental threat they face is from good old offset funding, from good old offset funders like Google (who is a fairly regular player in this offset funding game) with YouTube - as Mark says:

And all of their cases are challenged further by an uneven playing field. While all those music services have to charge for mobile access and have some gaps in their catalogues, YouTube provides unlimited access, on all mobile devices, with the world’s largest music catalogue, with video, for absolutely no cost at all to the consumer. As far as streaming goes, there is one rule for YouTube, and another for the rest. Until that anomaly is fixed, the rest will be swimming against the tide.

The only thing that has stopped YouTube from being the category killer yet is that no one has written a music playlist playout UI as good as Spotify's yet.

I was an early "Free" customer. Would I pay £9.99 a month? - No, but mainly because I know that its $9.99 (about £6.70) in the USA a and Euro 9.99 (about £8) in Europe. Comparison shopping ftw....my price point starts at the same deal everyone else is getting.

Readers of this blog will know we are following the intersection of robotics and the internet, especially flying robots (aka the Internet Of Flying Things). News in is that Iran has built its first drone is interesting (see clip above). The Atlantic magazine notes:

But how concerned should we be about Tehran's latest toy? According to the Council on Foreign Relations' Micah Zenko -- not much. So far, he told me, Iran hasn't been able to mount precision-guided weapons on the Shahed-129 like the kind U.S. drones rely on to perform targeted killings. And even before it can attack a target, the machine has to get there -- a task made more complicated by Iran's elementary command-and-control infrastructure. Even if everything went to plan and the drone escaped detection -- an unlikely event in any case -- Iran's military would still need to keep in contact with it to operate the thing.

"And," Zenko added, "since it is a country with a military budget under $10 billion and surrounded by perceived threats, they are compulsive exaggerators about what their military could do."

I think that is hubris of the highest order, frankly. The whole point of drones is thay are 2+ orders of magnitude cheaper to develop than human piloted fighting aircraft. It also shows a scary lack of the remembrance of things past. In 1914, aircraft were used for scouting and attacking enemy on the ground. By 1915 they were used to fight each other. I'd predict by 2015 we will have our first reports of drone attacks against other drones. The only other question is whether a drone will shoot down an aircraft (probably a helicopter) or another drone first

By its design (long, thin, wings) this drone is a glider, ie built for scouting, not fighting - but that is just how it started in 1914 too, and by 1916 proper fighter aircraft had arrived, in large numbers. Perish the thought that history will repeat itself here.

Attended the McKinsey event on Social Business in London Social Media Week yesterday, overall I don't think I heard anything new (the mere fact that McKinsey is sponsoring a session and pushing out a big report tells you Social Business is now over The Chasm, and is something saleable to the blue chips, not an early adoption technology). Even the Elephant idea is old now, we used it in the first "Social Business" talk we did at the first London Social Media week 2 years ago Most of the "ahas" I got, had to do with other implementation lessons (here are some of ours, from our own experience).

The only thing that has changed is that the market now has an £800bn price tag attached to it (a lareg anmount from the rebadging of Enterprise 2.0, I'll warrant). Anyway, here are my notes on "aha's" and "oh really's":

Implementation

- ROI is a hygiene factor. Early adopters are now getting benefits, but no one can cost justify it early up, so...
- Pilots are the best first step
- iPad was the game changer for Social Media in business, it combined power of notebook and ease of use of smartphone
- Social media business software has become far easier (and cheaper) to use (Lotus Notes to Wikis to Yammer)
- The "implementaion gap" is in middle management - juniors know how the tech works, seniors get huge support to make it work. Middle managers have neither. (Also, middle management gets the least benefit/is most at risk from its adoption)

The conflict between IP and getting Social benefits

- Social media "theory" today relies too much on trust, fine in consumer systems but motivation to cheat is v high in business, the problem of how to police this dichotomy is still in infancy. "Most people are good" vs "yes but the few who aren't can cause a lot of damage" issues still not clearly addressed, still a conflict between open-ness and IP protection.
- Key is to draw a hard barrier bewteen IP sensitive areas and those not, can't be at the business bariier.
- Is the user device in or out the enterprise now ( lock down vs user owned or shared resource)

"Oft repeated Theories still to be Proven"

- We are in "A world where abundance is possible". Saying something isn't possible is a clarion call to developing world developers everywhere. But what is abundant? Is the abundant worth having? See my notes on the declining quality of what is abundant in my "Gresham's Law of Information" posts
- "Open source has given a lot, it works" But look at who who makes money - the ones who can wrap up Open Source in their own IP
- 3D printing are next major assault on IP

Some Predictions from the Panel

Social Media Winners - Healthcare, Education, Consumer businesses, Electronics
Losers - Government, Consultants
Next Social Media company to to be acquired? Twitter
Next IPO? - Twitter
Most impactful next gen Social Media technology - Healthcare, User trust/rating

Good news was I saw quite a few faces I haven't seen for ages. And a Free Lunch

In late 1999 you couldn't move in San Fancisco's South of Market St area (SoMA) without tripping over a dot-commer. By late 2001 you couldn't move without tripping over derelict cardboard boxes and winos (I was there both times....). Now the Bubbletime seems to be in full swing again - San Francisco magazine:

In spite of the obvious urban warts, the word is out: San Francisco is the world’s leading tech paradise. At a rate eclipsing the dot-com boom of the 1990s, tech companies are setting up shop in the city by the hundreds, drawn by its beauty and livability, as well as the deep pool of engineering talent here and, yes, city hall’s increasingly tech-oriented policies.

Young entrepreneurs from as far away as Denmark, Singapore, and France can be seen with real estate agents in tow, roaming through converted South of Market lofts still vacant from when the previous bubble burst more than a decade ago. The city is currently home to more than 1,700 tech firms, which employ 44,000 workers, up a whopping 30 percent from just two years ago. And San Francisco has been the nation’s top magnet for venture capital funding for three years in a row. Consequently, the distinction between Silicon Valley and San Francisco has all but disappeared. It is us, and we are it.

I'm sure I could probably find the same breathless-yet-anxious prose in Red Herring or similar in 1999. Now there is probably a long term restructuring going on, as for a variety of reasons the Smart Young Things are gravitating to cities rather than campus complexes, but not this fast. Next will come the sky high rentals and deteriorating services for families moving in*. Those who cannot remember the past etc....

(*Hint. Property prices collapse as well as rise....)

Last month a naked Prince, this month a topless Duchess, where will we be by October at this rate

The really interesting thing has been watching the Meedja arbitrage and the angst going on because of it. In both cases, foreign media was quite happy to post pictures of bare Royal flesh, and anyone with a bit of nous and a browser can view the Royal bum, bollocks and boobs. But the British print media, still awaiting a spanking from the Leveson inquiry, is largely keeping its head down. Some, like the New Statesmen, are trying the "disgraceful, disrespectful" or "how childish" harrumphing play (good try, but no cigar). One paper, the Sun, broke ranks and published pictures of a body double in the Harry & Friend pose after they had gone viral overnet. (I wonder if they will put a Duchess double on page 3 now?)

But why should the Great British Public not see their Royal nekkidnesses? After all, everyone else can. Don't we have an even bigger right- as one wag on Twitter put it, as we the British taxpapers pay for the Royals, their assets are by rights ours. The real issue exercisng the great and good is not this particular boob, but the general case. Lord Mandelsohn pretty much summed up the overall issue, writing in the Financial Times:

"The bigger question is how the domestic media market can be made economic and subject to any form of regulation in an era when, a click away, there is access to information that respects no national boundaries and the laws of no single national parliament or the basic standards of conventional journalism," he wrote in a letter to today's Financial Times. He said it is difficult to see a future for newspapers when the internet has "ransacked" their business. "It is hard to see how some of the best-known sources of quality English-language journalism – the Times, New York Times, the Guardian spring to mind – will ever make money again," he added.

But lets just look at the end to end media value chain operating here to unpick where exactly the problem is.

- Firstly, content capture is not really an Internet problem, its an awareness of cameras problem. There were paparazzi and zoom lenses long before there were PCs and browsers. The lesson of several decades has been that people of public interest who get their bits out are going to have them captured, thats a given. And you'd have thought the lesson had been learned by the Royals of all people after the Harry episode. To misquote Oscar Wilde, one naked Royal photoset is an accident, two looks like carelessness.

- Secondly, content aggregation and distribution is not an internet problem per se, the Royal bitz were published in Olde Fashioned Print Media PDQ, just not (except for The Sun, eventually) in Britain. So its not as if no-one gets to keep abreast of the naked reality, its just the British who can now see what everyone else can

So in reality, the "internet issue" that is occurring is that of User Access by The Wrong Sort Of User. There has been a loss of limitation of access by those with power, to content that already exists, via the web. Getting around these authorities is not a new thing, by the way - in 1759 Voltaire published Candide in 3 European capitals at the same time so he could get enough copies sold before both the pirate media copied it and the various authorities banned it. But today, once its on a server somewhere, the authorities can't pull it down. (Plus ca Change, as they say in France).

And that, as Mandelsohn notes, has put the "Old Media" in an impossible position, as if they keep the old wary symbiotic dance with authority going they lose reader credibility and lose money, but if they also show their readers what many have seen already, then they will get taken to task and lose money. But his solution, unfortunately, seems to be more Old School. He feels that there is a problem with Free Speech:

"we come to grips with the fact that the internet is giving public access to uncorroborated, undigested and unmediated news, all in the name of free speech, is becoming one of the defining issues of the 21st century".

This is a Canutian view of the world unfortunately, unless one believes it is possible to censor the whole Internet. (Actually, Canute was trying to point out to his staff that even he had limitations to his power - now there was a wise British monarch). But it is worrying that the blame is being laid on the perils of free speech, and the huge risk that the general public should know things that previously only their betters (and sundry foreigners) did. That this is an absurd proposition in the first place seems not to occur.

So, as we await the Leveson enquiry outcome, two emerging New Media principles are - in my view - becoming clearer:

(i) Media that hacks into the public's privacy is going to increasingly be seen as a Bad Thing, unless the public wants it*, but....

(ii) Media that hacks into people who are paid for by the public in one way or another will be seen differently (and virally).

And once its all hanging out there, it is senseless to stop the media in the UK reporting on it. Any new laws, regulations, modes of operation etc that don't admit this reality will just be seen as an ass and a load of bollocks.

*Whether they know the risks they are taking is moot, so hopefully these sort of episodes are instructive.

It will be interesting to test our prediction of the iPhone 5 (made about 18 months, end of 2010/early 2011 - see above) ago vs reality. From what we have gleanes so far it has:

Processor Speed - Apple Twin Core A6, estimated speed unknown but rumours are it will be similar to the iPad 2, c 900 mHz
RAM - 1 GB
Memory - 64 Gb
Screen resolution - 326 ppi, but screen area size has increased by c 20% by increasing the length

(We have been arguing that smartphone screens will get to banknote size for quite a while, all the work we have done says the bigger the screen, the better the functionality, so eventually it will be as big as can be easily carried in a pocket - look at our iPhone 6 predictions, its hard to iamgien a device like that constrained in a c 2 x 3 inch screen)

We assumed weight would stay about the same, in fact its gone down c 20%, from 140g to to 112g - quite a drop. It would be interestimg to know where its gone down
Battery life we also assumed would be similar to the 4S. The iPhone 5 and the 4S both have eight hours of talk time, but the iPhone 5 can browse the Web for 10 hours instead of eight

Looking at the specs it seems that the processor is a genuine jump forward, but its other systems are more like an iPhone 4 with a bigger screen and a bit better battery life, as you'd expect it runs the latest comms systems (4G not 3G, latest WiFi etc) and a Moore's Law-esuqu increase in toy performance. No doubt the 5S will move to a 128 Gb memory and prove our prediction right

We don't think price has much to do with cost, when the costimgs come out it will be interesring to see where the weight was saved - I hope its not in the casing.

Now this is interesting - Reprieve:

Apple has for the third time this month rejected an iPhone app which alerts the user to a drone attack and to the number of people killed. Drones+ enables those concerned about the CIA’s illegal, unregulated use of these remote-controlled weapons to track the strikes to their handset.

This is no doubt an uncomfortable prospect for the US authorities, whose use of drones extends to Pakistan, Yemen and Somalia, where no war has been declared. Such drone strikes have killed more than 3,300 people in Pakistan alone since 2004, according to reports by the Bureau of Investigative Journalism. Reprieve is working to expose and challenge the programme, through investigation on the ground and representing victims in legal action across various jurisdictions.

Seems like Apple is squirmning a bit with such "difficult" content:

Strangely, Apple’s three rejections of Drones+ have each been for different reasons. On the first occasion, the company said its content was “not useful” (betraying something of an ignorance of the number of people interested in this issue, and the value in raising awareness of it); next, Apple said it was due to the inclusion of a corporate logo within the app.

Now, the company has told Josh Begley, the app’s developer, that the content is “objectionable and crude”. This is despite the fact that the app doesn’t display images or provide graphic details of any deaths, but simply aggregates news of drone strikes from various publicly available sources.

So Drones+ is more objectionable than, say, an app providing daily updates as to new weapons (including drones) developed? Or one teaching users to accurately shoot a range of 27 rifles, mimicking the effect of real guns? Both of which are available in the App Store.

The reason the Drones are attacking over the Afghanistan border is because thats where the enemy is, and everyone knows that, but nobody wants to say they know that, as the ramifications of officially admitting it are just too uncomfortable for all parties concerned. And Apple is a business, biting hands that feed you is just not going to happen. But its two of the biggest modern trends at loggerheads - mobile robotics vs social media. On the flipside, within the US and other countries there is increasing worry about private drones being used as spies, in which case the various invested interests here would be in alignment.

New technology giveth, and it taketh away. One to watch......

Interesting piece in the WSJ about an Urban Drfit for technology companies:

For as long as many of us can remember, high-tech industries have flourished in the suburban office parks that are so ubiquitous in Silicon Valley, North Carolina's Research Triangle and other "nerdistans." But in recent years, high-tech has been taking a decidedly urban turn.

Silicon Valley remains the world's pre-eminent center of high-tech industry, of course. But even in the Valley, denser, more mixed-use and walkable places, like downtown Palo Alto, are becoming the preferred locations for start-ups and smaller firms. And many other start-ups—Pinterest, Zynga, Yelp, Square and Salesforce.com, to name just a notable few—are taking up residence in downtown San Francisco.

New York City's Silicon Alley—after a false start during the tech bubble of the late 1990s—is now home to more than 500 new start-up companies like Kickstarter and Tumblr, not to mention the gigantic Google satellite in the old Port Authority Building on Eighth Avenue between 15th and 16th streets.

I recall at the height of the dotcom boom, you couldn't move in San Fran without tripping over a a Blazer'd and Chino'd dotcommer, but Silicon Alley was still mainly Chinatown - so is it just a new Bubbletime, or is it a real shift?

Across the Atlantic, London's once- derelict Shoreditch district—now known as Tech City or Silicon Roundabout—has been transformed into a thriving high-tech district housing 3,200 tech firms and 48,000 jobs, according to a recent report from the Centre for London.

Except that's not really true, the action is more spread out - Soho, London Bridge, and other areas all have their enclaves, but they are still mainly within the inner city Circle line, it is true. So asssuming it is a real trend, what is driving it? Part of it is avoiding the increasingly traffic-jammed commutes to beltway business parks, says Mark Suster (now in LA, but he was a London boy when I knew him), and part of it is other amenities:

There are also the distinct lifestyle advantages of setting up shop in the hurly-burly of real urban districts. Compared with previous generations, today's younger techies are less interested in owning cars and big houses. They prefer to live in central locations, where they can rent an apartment and use transit or walk or bike to work, and where there are plenty of nearby options for socializing during nonwork hours.

"It's not that young people wanted to live in Mountain View in the past," Mr. Suster blogged. "In fact, so many did not that companies like Google & Yahoo had free buses with Wi-Fi from San Francisco to their Palo Alto and Sunnyvale headquarters."

I suspect there is also the shadow of underlying modern economics there, in that I am reliably informed that you can no longer afford a Mountain View house and car on startup money, so if it's to be a bijou pad and bus it may as well be somewhere with decent amenities outside.

But SV and New York and London have long had overpriced housing, so why did people go to out-of-town campuses in the first place?

The answer was cost - companies used cheaper land business park campuses to minimise their costs, while also externalising the inconvenience cost by forcing employees to bear the cost of commuting and car ownership. However, when the majority of employees own the companies (as is the case in a small startup) then that economics is turned on its head and it's far more rational to live as near the shop as you can (or even above it, or in it). Also, its an attractor. A company baesd in central London is easy to get to for a large number of people all around London. Put it out on the beltway, and its becomes inaccessible to 3 quadrants out of 4 without a bad commute. And if they have other options, you've just lost 75% of your potential staff recruiting base.

Also, it would seem the startup and market economics have changed:

An even bigger part of the story is rooted in the changing nature of technology itself. A generation or so ago, the fastest-growing high-tech companies were more like factories. They developed proprietary software systems, designed and manufactured chips, built computers and created the infrastructure that made the Internet possible. Whether it was Microsoft or Apple, they deployed big engineering teams—and they needed big suburban campuses to house them.

The changing nature of technology—cloud-based applications in particular—enable new start-ups to succeed more quickly, with smaller teams and much smaller footprints.

The speed of technology has also accelerated. The companies that succeed are the ones that stay in the closest contact with their end-users and first adopters, as MIT's Eric Von Hippel has shown. When a company is located in a city, many of those end-users can be found right on its doorstep.

At the same time, high-tech products and industries are more multidisciplinary than they used to be. Success often requires excellence in more than one field of technology and in other lines of business. East London's tech scene is led not by tech firms per se but by "digital creative" companies that combine computer technology with music, art and narrative—and musicians, artists and writers cluster in cities.

I'd discount the London thing as an overall trend, London bats way above its weight in media (online and off, it always has, it houses the UK's Madison Avenue, Hollywood, News Media and the BBC plus its multiple spin outs/offs/ins/etc), just as SV does in new technology infrastructure, so I'd say that's a London cluster-specific issue. The rest of the economics of firm size, speed to market and close to consumer issues do ring true though.

The other thing about cities is that, well, they are just fun places for smart people to live in:

Cities are central to innovation and new technology. They act as giant petri dishes, where creative types and entrepreneurs rub up against each other, combining and recombining to spark new ideas, new inventions, new businesses and new industries.

Once you are tired of London (or New York, San Fran, Berlin, etc etc) you are tired of life.

(Incidentally the article is written by Richard Florida, whose major thesis is that liberal urban areas attract the most groovy, smart, creative and interesting people. As a Londoner, I have to agree . As an aside, my experience is that having a big University (or even better, a few) in the town is a huge boost to that vibe as well )