Gordian Knots

While there are plenty of theories about why Google launched a web accelerator project (since many don’t see how it relates to search), Tristan Louis has submitted his own theory, saying that it has a lot more to do with search than most people think. His argument is that the Web Accelerator is just a back door way for Google to use everyone’s computer as a distributed part of the Google grid. Since Google gets to record every webpage you go to, it’s not hard to see how that data can be processed back into the overall search engine crawler data for indexing — or for Google to just build an indexing system directly into the Accelerator. Considering their earlier comments, it might be a stretch at this point to think that’s where they’re going with this, but it’s at least a theory worth mentioning.

Meanwhile, people are freaking out after realizing that Google’s web accelerator is doing the one thing you don’t want a caching system to do: caching private login info. So, people are finding themselves logged in as someone else on pages that require a login (and probably realizing that others are going to be logged in as them). This looks to just be an extension of an earlier problem people noticed with Google’s search engine caching — meaning, they should have known about this already.

But wait, it gets even worse. Various sites are reporting that the “pre-fetching” option basically clicks all links on a page, including ones that say things like delete this account. Ooops

“The programming challenge that occupied two Russian PhDs for three years was to schedule the most profitable usage for the three seats available in 40-300 DayJet aircraft operating in a region about 500,000 square miles in size, and to do this all within five seconds. Because the FAA part 135 air taxi rules require a price quote for each customer, that’s the basis of the computational task. Using a few questions (Where do you want to go? When do you need to be there? How early can you leave? How many seats do you need?), DayJet software has five seconds to map a best-case route and generate a detailed price quote with the goal of 85 percent of the time making the potential customer an offer they are unlikely to refuse. This may not seem like much of a routing task, but it is a vast increase in complexity over anything else the aviation industry has tried before.

A big airline like Southwest, for example, has to make the best economic use of 417 aircraft flying 18 hours per day, which is network optimization problem with at most 7,200 variables. But DayJet, scheduling shorter routes on a per-seat basis that involve potentially vastly larger number of destinations, has a 15-hour day that includes a minimum of 13,500 variables before you even get to the possible destinations, and the DayJet network has to be re-optimized dozens of times per hour if the system is going to be profitable.

The result of this three years of schedule development, along with two years of complexity research on agent-based modeling of regional business travel, makes Iacobucci think his company can make money IF it has at least 40 aircraft per region. ‘Our greatest limiting factor turns out to be the rate at which Pratt & Whitney can make jet engines,’ he said.

While DayJet is a great idea and ought to be a lot of fun to do, its impact on business travel will really be modest. Iacobucci estimates that if the company is successful the mix of regional business travel in the U.S. southeast will go from the present 87 percent by car and 13 percent by scheduled airline to 84 percent by car, 14 percent by scheduled airline, and two percent by DayJet. Notice that the scheduled airline component is calculated to go UP, not down. That’s because DayJet expects that many customers will use the service in only one direction, mixing it with scheduled flights that may be cheaper. So airlines, rather than being threatened by DayJet, ought to help promote the service.

We think we know how the world operates, but we really don’t, write the authors of Freakonomics. And in that ignorance we make choices that often have significant, if not catastrophic, results. Take the decision made by the parents of an eight-year-old, who decide that little Allison will not be allowed to play over at Mary’s, because Mary’s dad keeps a gun in the house. But they have no opposition to Allison visiting Monica down the street, whose family has a swimming pool. Parents can be terrible assessors of risk: The likelihood of death by pool is 1 in 11,000; by gun, 1 in 1 million-plus.

Freakonomics uses the science of economics and concrete data to challenge our assumptions about everything from teenage crime to the motives of real estate agents. And what the authors find is that incentives, not a sense of right and wrong make the world go round. “Morality, it could be argued, represents the way that people would like the world to work, whereas economics represents how it actually does work.”

Levitt, who teaches economics at the University of Chicago, and Dubner, a New York Times writer, provide page after page of research demonstrating that:

* Incentives are the cornerstone of modern life.
* The conventional wisdom is often wrong.
* Dramatic effects often have distant, subtle causes.
* Experts from criminologists to real estate

You’ll walk away with not only a few good party tidbits (Did you know that popular baby names start with well-to-do parents and work their way down the socioeconomic ladder?), but also a more critical eye to many things presented as fact. (Timber cutter is the most dangerous job in the United States, the government claims. But the government evidently hasn’t compared that occupation to that of crack dealer.)

Watched two Sci-fi movies this weekend; ‘Close Encounters of the Third Kind’ and ‘The Day the Earth Stood Still’

Gilder’s Law: Winner’s Waste. The futurist George Gilder wrote about this a few years ago in a Forbes publication. The best business models, he said, waste the era’s cheapest resources in order to conserve the era’s most expensive resources. When steam became cheaper than horses, the smartest businesses used steam and spared horses.

Today the cheapest resources are computer power and bandwidth. Both are getting cheaper by the year (at the pace of Moore’s Law). Google is a successful business because it wastes computer power–it has some 120,000 servers powering its search engine–while it conserves its dearest resource, people. Google has fewer than 3,500 employees, yet it generates $5 billion in (current run rate) sales.