Wednesday, February 27, 2013

Google gets the competition it deserves

At this week’s Mobile World Congress, the conference is focused on what happens next, now after Android has captured the majority of the world’s smartphone sales — and continues to gain share. Some distant clouds are on the horizon.

The WSJ this week asked whether the leading Android vendor, Samsung, is going to assert its buyer power against Google.

Google executives worry that Samsung has become so big—the South Korean company sells about 40% of the gadgets that use Google's Android software—that it could flex its muscle to renegotiate their arrangement and eat into Google's lucrative mobile-ad business, people familiar with the matter said.
The story said “Android head Andy Rubin … said Samsung could become a threat if it gains more ground among mobile-device makers that use Android.” The WSJ followed up with a blog posting asking “Can Samsung’s competitors catch up?” while Fierce Wireless reported a Samsung VP’s denial that Samsung’s success threatens Android.

The original WSJ story speculated that Samsung might ask for better terms, e.g. preferential access to technology.

In some ways, we’ve seen this story before. Symbian was supposed to be an open multi-vendor platform, but when Nokia accounted for 80%+ market share, it transformed both the Nokia-Symbian relationship and the level of interest and commitment by other vendors to Symbian. Yes, Google’s much richer and more independent than Symbian ever was, but it faces some of the same pressures that Symbian did. As it is, Samsung is making more profit from Android phones than Google is (an interesting reversal of the Microsoft-Dell exemplar).

(Google’s downstream vertical integration into Motorola is offered as an insurance policy, but since Motorola has been slowly dying for a decade, it’s not clear how credible a bargaining chip that is.)

Similarly, Samsung continues to support Tizen (the embedded Linux successor to LiMo, Moblin and Maemo), and plans on offering a new phone based on Tizen this summer. Samsung is using Tizen as an upward compatible replacement for its homegrown Bada, but it’s unclear how credible a bargaining chip Tizen will be — since it hasn’t offered a new Bada phone in two years.

The other challenge to Android comes with the introduction of the Firefox OS. Since handset OS makers — Google, Apple, Microsoft — are promulgating their own browsers, apparently the Mozilla Foundation figures they need an OS to put their browser into people’s hands.

The Firefox OS won support from LG, ZTE, Huawei and Alcatel, as well as serious interest from Sony (née Sony Ericsson née Ericsson) — but not from Samsung. It’s expected to ship from 18 carriers in nine countries, but not the US until at least next year.

There’s of course the question whether the world needs another smartphone OS, let alone another open source OS (remember webOS). After Android (69%) and iPhone (22%) together have 81% of the market, no other platform has more than 5% — with Tizen and Firefox starting behind Blackberry, WinMo and Symbian. But there’s no guarantee that the most popular OS in the US or Europe will be the most popular OS in China, particularly when China’s two largest handset vendors are supporting both Android and Firefox OS.

So based on recent history, Google’s concern in developed markets should be Samsung throwing its market power around (either within Android or to a rival platform), rather than having Firefox (or BlackBerry or WinMo) catch it any time this decade. It needs monitor Tizen or Firefox in the BRIC countries, but that could just be a matter of providing extra tech support engineers for Huawei and ZTE.

Monday, February 25, 2013

Correcting WIkipedia's distorted mirror on the world

We had friends over last week, and at one point the discussion drifted to Wikipedia. The husband is an adjunct and asked me what my policy was (after 14+ years of teaching) on using Wikipedia as a reference.

Wikipedia used to emphasize that it is a tertiary source — a compilation created from of primary and secondary sources — and thus is not an authoritative reference. However, today it appears to be less modest and more into bragging. Due to its price and comprehensiveness, Wikipedia has become the first (if not only) reference for the Internet generation.

Certainly Wikipedia is suitable for resolving a bar wager: the sort of questions that people used to argue about are now easily settled via Wikipedia (or via more specialized and authoritative sources such as IMDB).

It’s also well known that the Wikipedia mechanisms do not prevent the intentional fabrication of lies, whether it be libel (falsely accusing a political figure of conspiring to kill JFK) or false history (the so-called Bicholim Conflict). No process will ever prevent this — any more than a library can prevent cutting pages from books on open stacks — but the processes seem to eventually identify these problems and correct them. It also appears that true experts (e.g. actual scientists) will check key topics and make sure that the most egregious factual errors get corrected.

The remaining problem is the distorted mirror of American and other societies that Wikipedia presents to the world and to posterity. (By sheer weight of population. is clearly dominated by the American perspective, and the English articles outnumber the next three languages combined.).

By nature of its contributors — on average younger and with more free time than the average Internet user — Wikipedia has a social, cultural and political bias of its contributors versus the rest of society. In particular, these authors demonstrate a pre-occupation with contemporary popular culture over other (more enduring or important) aspects of human knowledge: not everyone can write about string theory, but just about anyone can summarize a Simpsons episode.

And thus Beyoncé Knowles has a 36-page entry, and the late Michael Jackson, 49 pages. John Lennon warrants only 20 pages, but the Beatles have 38 and Elvis 45. Charlie Chaplain merits 35 pages, versus 20 pages for Rembrandt, 19 for Beethoven and 17 for Mozart.

By way of comparison, the most famous scientist of the 20th century, Albert Einstein, rates only 29 pages. Other 20th century Nobelists include Niels Bohr (who gave us the atom) with 14 pages, Enrico Fermi (who split the atom) 16 pages, and Wilhelm Röntgen (who won the first Nobel prize for discovering x-rays) with 5 pages. Martin Luther (who changed the course of European history) had 32 pages, Charlemagne 36 pages (but only 24 pages in German) and Henry VIII has 30 pages.

Overall, the depth of coverage of major figures seems adequate. But, lacking limits on resources (either to generate content or in printing it) the coverage of trivial topics balloons far beyond all reasonable measure.

Exhibits A-Z are the coverage of cult favorite American TV shows such as South Park or the Simpsons. South Park has 28 pages, but then another 17 pages listing all the episodes. More significantly, there is a 3-10 page entry for each of 237 episodes aired across 16 seasons (thus far). One US TV show with a run of nearly two decades — which will be forgotten a century from now — merits 1000+ pages, more “ink” than all the major painters or composers of 500 years of European history combined.

If journalism is the first draft on history, now Wikipedia offers the first draft of a comprehensive encyclopedia that could, in the end, crowd out all other records of our contemporary society. The result is what one would expect if archeologists centuries from now tried to assess the 20th or 21st century from uncovered copies of US Weekly or videos of “Entertainment Tonight.”

I'm somewhat optimistic that the problem can be corrected in this century, because there are millions of people who have the knowledge to correct these distortions. It’s not some core problem of economics (Wikipedia demonstrates this) or scarcity, but merely a matter of incentives. Right now, people who actually know something that’s scarce have no incentive to give it away in an anonymous crowdsourced encyclopedia, but instead seek course credit, (less and less often) to sell it for publication or generate academic reputation through peer reviewed journals.

Google’s dead knol experiment was one attempt to create a new production community. Still, the sheer volume of excess available labor suggests that there will be other attempts. In particular, if there were a way that students or professors got credit for term paper-quality original contributions shared on the Web, we might have a more coherent (and representative) picture than what what Wikipedia presents to the world.

The MIT-spawned BioBricks Foundation is crowdsourcing synthetic biology components through its annual iGem competition, drawing on a far more skilled and specialized knowledge base than Western history . (Similarly, MIT’s Open Courseware has spawned a wave of free online university course content). This implies even one visionary university could lead us to a new model of knowledge generation.

Thursday, February 21, 2013

Strong US support for increasing unemployment

A USA Today poll reports that 71% of Americans back the president’s plan to raise the minimum wage: 87% of Democrats, 68% of independents and even a 50-47% plurality of Republicans. Only the tea party supporters oppose the move (64-32%)

The only problem with the plan is that — based on more than 100 academic studies — it will raise unemployment, particularly among the least employable workers (e.g., minority teens).

A 2006 NBER study by David Neumark of UCI (go Anteaters!) and William Wascher of the Federal Reserve — later published in Foundations and Trends in Microeconomics — reviewed nearly two decades of research on the impact of increasing minimum wages. Here’s the abstract:

A sizable majority of the studies surveyed in this monograph give a relatively consistent (although not always statistically significant) indication of negative employment effects of minimum wages. In addition, among the papers we view as providing the most credible evidence, almost all point to negative employment effects, both for the United States as well as for many other countries.

Two other important conclusions emerge from our review. First, we see very few - if any - studies that provide convincing evidence of positive employment effects of minimum wages, especially from those studies that focus on the broader groups (rather than a narrow industry) for which the competitive model predicts disemployment effects. Second, the studies that focus on the least-skilled groups provide relatively overwhelming evidence of stronger disemployment effects for these groups.
While individual studies have reached varying conclusions, such comprehensive literature reviews are the way that science reaches a consensus on an unsettled question of causality (whether innovation policy or global warming). The administration wants to be guided by science when it supports its policies, but ignores inconvenient truths.

When our educational system turns out high school graduates that are illiterate or innumerate, they pay the price. When it turns out economic illiterates — leading to bad policies — others pay the price for those policies.

Friday, February 15, 2013

Sell the car, not the Kool-Aid

As promised, on Thursday Tesla CEO Elon Musk posted the logs from the disastrous NY Times long-distance test drive of his Model S sedan. Late Thursday, Broder himself posted his own attempt to reconcile his experience with the data.

At first, I was puzzled by Musk’s quixotic attempt to pit his famous money-losing startup against the country’s oldest and most valuable media brand. Yes, as a former engineer-entrepreneur, I get that engineers will scream “user error” at the top of their lungs rather than even consider that there might be problems with their baby.

There’s also the common sight of the rich egomaniacal founder — in this case, the highest profile multi-industry entrepreneur since Howard Hughes — assuming that a few business successes meant that he possesses the Midas touch, papal infallibility, or genius worthy of a Nobel prize.

Then I read the comment logs on some of these stories. Yes, the polarization seems worthy of a political campaign, but what surprised me was the vituperation against the NY Times reporter, in effect accusing him of being part of the evil Big Oil conspiracy to kill little ol’ Elon. Even one of my normally rational (except on politics) former co-workers has been tweeting that with a NYT reporter “you wouldn’t think guy would lie” and that the reporter was “unethical.”

The tactics being used by Tesla are straight out of a political campaign: Musk is not trying to convince the general public, he’s trying to protect his base, the true believers who’ll pay $80k or $100k for a car that makes them feel good about saving the planet but provides transportation equivalent (or sometimes inferior) to a $30k Toyota.

Psychologists tell us such an approach works due to “confirmation bias”: if you’ve plunked down a $5k-$40k deposit to get on the Model S waiting list, you will tune out messages that might suggest you’re getting a lemon. (I wonder how these buyers of this 400+ hp car will feel reading Tesla’s advice to avoid rapid acceleration and keep the speed under 60 on long-distance trips, let alone turning off the heater in the winter).

Perhaps Tesla’s near term goal is to keep those intended buyers from demanding their deposits back, since the company has been spending that money to keep the doors open. Or (as with a political campaign), perhaps Musk really believes what fawning reporters say within the Silicon Valley echo chamber. As was Steve Jobs in the Job I era, he’s likely surrounded by a staff who drink the Kool-Aid.

As with any product failure or business dispute, the truth probably lies somewhere in between. Some more neutral parties, such as Slate, Business Insider, Wired and particularly the Atlantic attempted to reconcile the new data with the original article by John Broder. And Broder himself wonders if some of the speed data was thrown off by using smaller wheels than normal.

If this were an actual court case, then Musk would lose — and lose badly — for two reasons. First, Broder is a more credible witness. He was there when the car was being driven, has witnesses and some of his explanations (like driving around a parking lot looking for the charger) disarm Musk’ accusations. He concedes that a caption (usually written by an editor) was misleading. In the end he’s willing to admit that if his sole goal was to maximize battery life, there are things he would have done differently.

The second problem is that, as Perry Mason or Matlock would be quick to point out, there’s no motive. Why would a NYT reporter who since 2009 has been “the Washington bureau reporter responsible for coverage of energy, environment and climate change” try to destroy America’s most famous electric car company here. We’re not talking Fox News or even the Wall Street Journal here. Meanwhile, with the survival of his company riding on this one product, Musk’s motives to present the data in the best possible light are quite clear.

Under other circumstances, Musk might able to win support from pro-business conservatives who distrust the NYT after years of editorializing on its news pages. (Some have compared Musk’s war against the press to that of Richard Nixon 40 years ago). But this audience is (for once) ready to believe the NY Times over a crony capitalist who sought $700 million in Federal loan guarantees (and won $465m) to keep his cleantech startup going.

The Model S is not the first cutting-edge product to fail a product review: this sort of thing has been happening for PCs, software and smartphones for decades. Sometimes, it’s because the rough edges are still showing: Broder was clearly misled by the onboard computer’s range estimates. In other case, the reviewers (again like real customers) use the product in ways that the engineer would not or would not even imagine.

In the end, Tesla will have to recognize that they made a mistake in providing their car in the middle of winter to an independent reporter, and telling him that he could drive up the Eastern seaboard using only the company’s “Supercharger” stations. The company will need better software, an onboard charging station locator, and 7/24 tech support equipped with better scripts for dealing with range anxiety in cold weather and other adverse conditions.

As with anything else in life, we learn most from our mistakes. Tesla will have a much better product (and make Elon Musk even richer) if they make a Model S that anyone can use rather than trying to sell the Kool-Aid to true believers.

Monday, February 11, 2013

Bluster as a substitute for execution

Cross posted from the Cleantech Business blog.

In the ongoing search for electric car nirvana, the Tesla Motor Company has enjoyed an unusually charmed existence. Perhaps it’s the Silicon Valley mystique, perhaps it’s the Midas touch attributed to its co-founder Elon Musk — who became a centimillionaire from selling PayPal to eBay, and then used his funds to start a car company, a rocket company and a solar company.

After discontinuing its $100k niche toy, the Tesla Roadster, the future of the company depends on producing and selling its $60-100k Model S sedan in volume. The latter effort was dealt a major blow Sunday when the NY Times reported the very real problems in an actual test drive:

Stalled Out on Tesla’s Electric Highway

Washington — Having established a fast-charging foothold in California for its electric cars, Tesla Motors has brought its formula east, opening two ultrafast charging stations in December that would, in theory, allow a speedy electric-car road trip between here and Boston.

But as I discovered on a recent test drive of the company’s high-performance Model S sedan, theory can be trumped by reality, especially when Northeast temperatures plunge.
The problem was that — after several close calls — the car ran out of power shy of the next charging station, requiring a complex and time-consuming flatbed tow. Perhaps it was the effect of cold upon the battery life, perhaps it was the power consumed by the heater, perhaps it was bugs in the software or hardware.

Still, there’s no reason to think that the problems didn’t actually happen. In response, one would presume that Tesla would both improve its products and add additional charging stations to enable long-distance recharging.

Instead, the notoriously thin-skinned Musk tried to smear the messenger, both on a CNBC interview and on his twitter account:
@elonmusk: NYTimes article about Tesla range in cold is fake. Vehicle logs tell true story that he didn't actually charge to max & took a long detour.
In responses to major media outlets, the NYT stood by its story:
The Times's February 10 article recounting a reporter's test drive in a Tesla Model S was completely factual, describing the trip in detail exactly as it occurred. Any suggestion that the account was "fake" is, of course, flatly untrue.

Our reporter followed the instructions he was given in multiple conversations with Tesla personnel. He described the entire drive in the story; there was no unreported detour. And he was never told to plug the car in overnight in cold weather, despite repeated contact with Tesla.
Apparently the attack was an effort to prop up the stock price, which fell 4% in response to the NYT story. (That’s about $175 million in market cap — more than any of us mere mortals will ever see in a lifetime).

Despite the stress on the company and its stock, this is a textbook example of how not to handle a PR crisis. But it appears that within a NASDAQ-traded public company, no one can tell the emperor of Tesla to put his clothes on, or to listen to professional advice. As The Atlantic summarized its media report: “Elon Musk's Crusade Against The New York Times Isn't Helping Tesla.” The WSJ wonders whether this sort of concerted effort to intimidate reviewers will discourage coverage in the future.

Of course, this happened the same week that Musk — an expert in all things everywhere — was offering advice on Boeing 787 batteries. As Seeking Alpha dryly put it:
Tesla has burned through $1.25B in free cash flows in order to develop the company, and we expect that terrifying test-drives of electric vehicles from Tesla Motors will continue. We find it amusing that Elon Musk is willing to help out Boeing's Dreamliner due to the battery issue. We would like to remind Elon Musk and his team that they need to first fix their problems with their products before trying to be a superhero with the products of other companies.
Cruising range is an inherent limitation of the current generation of electric cars, and thus “range anxiety” will be a major obstacle to adoption. Musk has done himself — and the industry — no favors by helping to call attention to the article, rather than (as his employees apparently were trying to do) work with the reviewer to understand and correct the problems.