Hal Varian, Google’s Chief Economist and University of California at Berkeley professor, talks with Russ Roberts about Google, the role of technology in our everyday lives, the unintended paths of innovation, and the value of economics.
Size: 18.2 MB
Right-click or Option-click, and select “Save Link/Target As MP3.
0:36 Intro. Grandfather story: who wrote the poem with the line “yet the strong man must go.” Finally jumped up and yelled “Robert Browning, Prospice. Now we just Google it. But some are threatened by that. Cultural behemoth. Should people be worried about Google? Inevitable that large, powerful company attracts scrutiny. In many ways, healthy phenomenon. Public relations. [July 22, 2008 podcast taping on site at Google.] Article: Google Making Us Stoopid, on technology shaping our brains. Short attention span. Calculators were also once accused of making us stupid. Frees your mind to pursue creative processes. Using GPS vs. reading a map; shoeing a horse. Hard to read long books. Easily distracted by email and other things. Being a University Dean destroyed attention span. Time divided up into 10 or 20 minute chunks. Similar: 850 word limit for NY Times articles. Robert Frost: Like playing tennis with the net down. NPR commentary, 475 words. In Washington nobody pays attention to anything that doesn’t fit on a bumper sticker; in business, 1 page memo; academic, 40 page document with footnotes and appendices.
6:46 What is the future of searching? We have become incredibly spoiled. Most people think the cool thing about Google is that it searches a lot of sites; but the cool thing about Google is that it “knows,” yields, produces a set of pages we’re likely to want to read. Extraordinary achievement. Is it going to get better or different? Herb Simon, “a wealth of information creates a poverty of attention.” Information retrieval: precision and recall. Precision: Of the documents retrieved, how many are relevant? Recall: Of how many documents there are, how many have been retrieved? Can always retrieve hundreds, thousands; challenge is to make sure they are the most relevant to the query at hand. Over time users have become much more skilled at entering queries; gives search engine a much better chance. Art of searching. Machine-aided cognition: nothing wrong with using a calculator; important thing to know is which numbers to type in. Google is an aid to retrieval. Is technology itself going to get smarter? The technology is ultimately based on statistical relationships. Google search algorithm in English applied to another language it will work remarkably well. As we get more data, more experience, it will get better; but interacts with the user’s getting more skilled. Research used to be sitting in a musty library; now it’s sitting at a screen. The car in 1910: car’s gotten better, but drivers have gotten better, too. Mobility search: iPod used to search. What we want to do with our cars also changed. Heated and cooling cup holder in your car. How could we live without it? New car with radio key, push button start, no fumbling, blinks hello. Remote car door opener.
13:30 Hardware and the Internet itself are going to change. IPhone: people use it to get directions. Research and convenience. Where is it headed? Internet makes information more democratic. Democratization of data. Will it change politics–policy–how the world gets molded by political processes. Pretty rapid fact-checking via Google, verifying what a candidate is saying. The best lies of politicians aren’t factual ones. Spreading of education, not just facts. EconTalk. No better time in life to be an aspiring economist without going to graduate school. True in every subject. Blogs. You can dip your toe into intellectual activity, almost without charge. Blogs, solipsism, talking only to their own group; freedom of entry makes possibility of challenge. Global warming, huge problem of our time; debate, what is cost, what are the impacts, how are things developing, what is the magnitude–incredibly healthy to have that kind of debate. Concern that news is becoming partisan. May not be a bad thing. It’s what people want, to confirm their own biases, enjoy rant by a fellow thinker. Political science friend: Best predictor of your views are your neighbor’s views. Now the world is a neighbor.
19:31 IPhone: different technologies for accessing information. Open and closed: computer is sort of open; phone is closed. Different networks, cable TV closed network; mobile networks proprietary. Books, newspapers, more competition. On the Internet everybody owns a printing press. Internet is a lab experiment that got loose. Not designed to be a popular communication mechanism but for elite scientific community. Those people are still surprised that it became so widespread. First Xerox machine was thought to have no market–how many copies would someone want to make? Made to replace carbon paper, carbon copies, cc in email from “carbon copy.” One business model for the telephone was to pipe music into your house: picture of orchestra playing, 3 cents a minute for pop music and 5 cents a minute for opera. Early type of iPod, prior to radio. Killer app for radio was ship-to-shore communication; nuisance that others could overhear your communication. Might not be a bug, might be a feature. Challenge was how do you pay for it? Chris Anderson book Free, recent podcast.
24:44 What role will the government or competition play in closed or open systems? Genie is out of the bottle; push for open model.
Europe: standards for document exchange. Business model to support open network. Open science, publication. Column at New York Times: whether information technology matters. Nicholas Carr suggested that IT didn’t matter. Issue: has IT become somewhat generic. More generic now that 20 or 30 years ago. Systems that were only available to the largest and richest companies are now available to many. Wal-mart, K-mart; now smallest Mom and Pop store can have same inventory tracking technology. Data side: infrastructure is reproducible across companies but expertise is not. Have to make sense of the masses of data. Key to success: scarce factor of production that is highly complementary to something that is ubiquitous and cheap. Monopoly on left shoes when right shoes are free. Data is ubiquitous and cheap; analysis is complementary scarce factor. Telling a story and translating it into practice. Statisticians, sexy career of the next decade. Major in business, learning to write. Course on data bases and statistics requested by employees. In defense of the MBAs, the way the topics get taught in a business school are not always the most useful, partly because they have to be general but also because the people teaching them tend to be high-minded, ivory tower. Hard to teach people things in universities that don’t easily lend themselves to exams. Wisdom is not easily tested. Test facts, proofs; miss out on subtler forms. Gap even in economics, apprenticeship model. Important to learn the facts; but also transition to apply them in an intelligent way. Econometrics: didn’t learn it till had to teach it. Professors are required to communicate what they know.
34:03 “Use of Knowledge in Society”–read it but didn’t understand it till had to teach it. Comparative advantage. Statistics: apprenticeship is right model. Chief economist at Google, used to be the kind of job that involved a large scale macro model. What is that role now? Dean at Berkeley, time off for good behavior. Look at ad auction, on side of page. Did game-theoretic model, simple techniques offered insight; econometrics analysis. Model within 4-5% error does good job of explaining what goes on there. Google traffic fallen through the floor; look at the data. Plot in log terms. Percentage drop was the same as the percentage drop last summer and summer before. Looking at absolute numbers was worrying. Simple things can go a long way in generating insight. Became chief forecaster.
[originally published in econtalk]