Sunday, April 22, 2012

Finished (E-Reader) - "In the Plex" - Steven Levy

"In The Plex: How Google Thinks, Works, and Shapes Our Lives" - Steven Levy is an excellent history of Google, which covers key points in the history of the company, how it started, grew and became huge.

The interesting aspect of the book is how Google struggles with the "do no evil" mandate that they try to live by, even when they are now huge enough that some consider them in the "evil" role formerly occupied by Microsoft and earlier by IBM in various periods of recent computer history.

Trying to imagine the pre-Google world is difficult.  When I was in graduate school, Windows was not yet the standard on the desktop (Windows 3.0 was the first really big Windows product) and there were DOS, UNIX and other machines operating.  E-mail was pretty new - grad school was my first e-address.  Most of the internet was vague to most people, and was primarily file sharing point-to-point, with newsgroups and bulletin boards providing some of the services now taken fro granted.

Google was originally a search algorithm, which provided better results than competitors, primarily by considering websites somewhat like academic articles - keywords may serve a role, but it is the central usage of key players that determines importance.  In academia, this is citations - the more your work is referenced and built upon, the higher "importance" that work can be thought of to have.  Google put this to work and was able to determine importance of websites by visit and linkages in a similar manner.  By itself, this innovation would have been important.  The bigger impact was to apply this lesson to advertising - to let the "value" of the ads to consumers (e.g. us) determine ranking, and to link ads most of interest to customers rather than to link ads randomly.  This allowed Google to do what many start-ups couldn't do - make tons of money, and do so in a tri-mutually beneficial way (Google made money - it was happy; advertisers were only charged if people paid attention to the ads and clicked - they were happy; customers were less harrangued by ads that had little or no interest to them, and in some cases, found the ads were as meaningful as the search they were undertaking - they were happy).

Thus, Google was able to live up to the "do not evil" idea.  Free e-mail and much cloud-based applications and storage are all due, in large part, to Google, who can afford to provide free products to customers, as it lends itself nicely to the advertising model which provides they money - they aren't required to profit from the application(s) and service(s) themselves, they help feed the information and advertising model that drives the profits.

Living to this standard became more difficult when they introduced free e-mail, but decided to search the mail files to target advertising (considered a breach of privacy - particularly at the time), or to make forays into China, a country that wants to screen and censor information.

All in all, a great read - easy to follow, compelling stories, interesting personalities (not always in agreement) and a contemporary timeframe (all the stories, issues and concerns are recent enough to resonate with most readers).

I come away from the "In the plex" with a new respect for Google, and the innovations that made the company successful, though I can't avoid the feeling that much of the "best days" from a "do no evil" or even from a "wow, that's cool" perspective are behind them (a belief shared by some employees who left for smaller start-ups, to regain some of the early-Google feelings).  This is not to say that Google has become evil, it is just much more difficult to play that role when everything you do is magnified to the scale that the current Google corporation is - every step, or mis-step is going to make waves - some waves are beneficial, oxygenating and re-vitalizing the water, others grow and become tsunamis.

Finished (Audiobook) - "The Science of Good and Evil" and "Why People Believe Weird Things" - Michael Shermer

I listened to "The Science of Good and Evil" by Michael Shermer and was, surprisingly, quite bored and disappointed in the book.  I've recently read many books on the grey area between science and religion (e.g. Dawkins and Hitchens; some Creationist stuff), many better than this one.  However, I should acknowledge that many of the discussions/examples suffered by being similar to issues discussed by Dawkins and Hitchens, so Shermer was starting at a distinct disadvantage - I might have appreciated the book much more had it been earlier in my reading sequence.

However, "Why People Believe Weird Things" is a great review of the goals of science, and the distinction(s) between science and pseudoscience.  Nicely laid out, interesting, with great examples of psychics (cold reading), dismissing astrology as it failed to predict disasters that killed thousands (e.g. 9-11 and any of several tsunamis) and basically lays out skepticism (not cynicism) as the basis for investigating relationships - also known as the scientific method.

I'm not sure where I got it, but I had a recording of Mark Bell (Coast to Coast am) radio show with Shermer as a guest.  The show is a call-in show, right of centre politically, playing to a primarily religious audience, though likely not an extreme one.  One of the callers (prior to the start of the recording) called in to say his wife was quite ill with an aneurysm and was critically ill, asking for listeners to pray for her.  Shermer, of course, doesn't support the interventionist-God hypothesis (don't think he considers himself an atheist, as does Dawkins, but he doesn't believe in the power of prayer).  At some point he was asked about the caller and the request for prayers.  To his credit, Mr. Shermer did represent his views, bringing up issues such as asking whether or not God was not aware of the situation prior to prayers, or if people who have loved-ones who die should somehow feel responsible, as either they didn't pray, or didn't pray enough.

The discussion that followed was very interesting, particularly for having been "off the cuff" - it involved whether or not psychics or others who are entertaining and may provide solace to grieving individuals (e.g. by allowing survivors to talk or communicate with the deceased) are doing damage or a service.  The host was taking the "service" perspective, as folks felt soothed by the "communication".  Shermer took the "damage" perspective, as treating entertainment as science reduces the general understanding of "science" and that people aren't served by being tricked.

The discussion didn't go deep enough to sway any believers in either side, but was a very nice example for skepticism and its role in problems solving, and an entertaining philosophical discussion.

All in all, my overall opinion of Shermer is unchanged - I like his viewpoints and the vast majority of his writings (including the articles he used to write in Scientific American).  I would even recommend "the Science of Good and Evil", particularly as an introduction to the religion-science border regions.

Finished (Paper Book) - "Are You Smart Enough To Work at Google" - William Poundstone

As a book of clever problems and puzzles, I liked "Are You Smart Enough To Work at Google" by William Poundstone.  Many of the problems presented would have left me in a coma, but going through the answers, there did seem to be certain ways to step back and examine the tenets of the problem, some of which you can actually take strides to solve.

However, as a book about HR and hiring practices, the book actually left me depressed.  I've never had a great love of HR interviewing techniques, as they all seem so idiosyncratic, an often are based only on applicants proving that they respect the hiring process.

From my view, from both sides of the table, I consider interviews an examination of whether or not the applicant and the corporation are a mutual fit - a two-way street.  As such, there should be a level of openness and honesty in the conversation, with the HR role, of course, making sure that the applicant has the required background skills/knowledge (largely gained from the resume and work history), but primarily focusing the interviews on the much more important "fit" characteristics.  Excellent candidates who expect to work in a team environment won't work out well in an isolated environment and vice-versa.  Highly political environments favour some candidate profiles over others.

Doing problems of the type explored in Poundstone's book merely shows that HR folk don't have any science to fall back on.  Their real value of assessing candidates on fit/personality and making sure that processes are fair to candidates (e.g. not swayed by friendships or biases) is a key role, but doesn't seem to be one that HR folk hang their hats on.  Doing "trick questions" may be fun, and in some cases may provide some illumination on particular personality characteristics (e.g. frustration tolerance, ability to deal with incomplete instructions), but these can be dealt with more honestly and directly through more conventional means.

I fully understand that Google has many more applicants than it can possibly deal with, but doing an unreliable selection methodology seems quite "un-Googly".