Friday, 4 July 2014

Changing the "Culture" of Silicon Valley



One of my friends on the social networks posted an article last night (I saw it this morning - she lives in San Francisco, California, and I live in Paris, France, nine hours ahead).  The article isn't really unique, following a rash of "why is tech so hostile to women/ethnic minorities/people who expect to have an actual life" links that have sprouted like mushrooms after a week of rain.

The basic theme of them has been that the Valley, which purports to be an objective meritocracy is in fact, a sort of echo chamber dominated by white men.  They report rampant, but unfortunately, intangible misogyny of the sort you might find in a college fraternity.  Albeit, a frat filled with nerds who would make Lambda Lamda Lambda look like a bunch of well-adjusted dancing smoothies.

Previous articles talking about how the valley is dominated by white men are actually factually inaccurate (the most risible was reported by CNET under the headline "Yahoo Workforce Mostly Male, Mostly White", and supported its findings with stats that showed that among some of the top firms - Google, Yahoo, LinkedIn, Facebook, about 50% of the workforce is white, in a country where two thirds are, so the truth is, from a mathematical angle, tech is under-represented by whites.)  You won't get much traction with articles about how there are too many Asians in tech.

For obvious (and in my opinion, legitimate) reasons.

I expect that the US president, diversity "activists," and others with certain political benefits at stake will fudge, distort, or simply fail to understand mathematics.  CNET, an ostensibly tech oriented journal should not make such egregious, obvious mistakes.

The blog post I read today, written by a man called Carlos Bueno, was better than those and not subject to the same laughable gaps in logic and reasoning.  Bueno, who spent time as a programmer at, among other places, Facebook and Google and is himself now the author of a book called Lauren Ipsen, a play on the vacuous filler Latin that anyone who has used publishing software will immediately recognise, described problems in the hiring process at tech firms under the rubric of culture.

The two are of course related, but not the same thing.  It's a minor fault, and perhaps not nearly as obnoxious as using a stock photo of a Google office that is obviously in Singapore as a stand-in for "Silicon Valley" (copy editor, clean up on aisle six).

Bueno's real argument boils down to the lack of professionalism in the 'interview' process at many start-up companies.  Using anecdotal examples, he illustrates that typically, the "process" is not actually a set of real interviews, but rather, sequences of unwritten hoops the candidates must hop through - going out for a casual coffee, a round of after-work drinks, playing ping pong (!) with some potential candidates.  Bueno rightly argues (in my opinion) that this process is not likely to identify who the best programmers are, but rather, to identify who "fits" with the biases of the team.  And, surprise surprise, it leads to what he calls a "Mirror-tocracy," or, a sort of band of brothers (and to be sure, 70% or more of those in tech are male).

This result is undeniable.  But I wonder, is it "bad?"  And by that, I mean, bad for the businesses?  Bad for the economy?

Bueno seems to imply that culture and bonding and sameness are intrinsically bad for companies, as they result in a lack of differing views, and a difference in perspective.  As evidence, he cites a study done by economics professors at the Harvard Business School.  The findings of that research indicated that the more affinity that existed between venture capital funds co-investing, the less likely the venture was to succeed.  Success was narrowly defined in the studies as "IPO."  Interestingly, two funds that were run by two investors who had worked at the same company were 17% less likely to 'succeed,' that if they had been to the same undergraduate school, 19% less likely, and of the same ethnic background, 20% less.

These facts cast serious doubt on the wisdom of having your company too similar.

But then, a few problems if you look at the fine print.  The most obvious is that the success metric was "IPO."  There is a good many startups who are very successful without going IPO (YouTube comes to mind.  It did not have an IPO, but who would call its multi-billion dollar takeover a failure?)  Then, there is the issue that the study is not about tech, but rather, investment.  VC provide the money, but they are arguably not closer to "tech" than Goldman-Sachs, save for the fact that Sand Hill Road is in Menlo Park and not New York.  And finally, the sample space, if one reads the article, is mainly about Chinese and Indian investors.  Thus, a venture where one investor is Chinese and the other Indian is more likely to succeed than one in which both are Indian.

There is a host of reasons why that might be true, and not one of them is related to the culture in a start-up.

The larger point that the author seems to want to say that programming would necessarily benefit if programmers were more 'diverse,' and by this he means that there are more women, more Latinos and/or blacks, and that fewer middle-class young people were employed.  All of these might be true; but are they necessarily so?

Like Bueno, I worked in Silicon Valley for many, many years. Part of that time, I was in a small startup.  And like, Bueno, I put in many hours writing "code," which is the mot juste for "programming" these days.  For better or for worse, the overwhelming majority of this work is commodity labour.  The "problems" being solved by programmers, with a few exceptions, do not require any particular genius or insight or skill.  People in the valley do an enormous amount of self-promotion about "disrupting" this or "changing" that.  But in reality, what the tech industry is now is mainly about slick packaging.

"Tech" is now about 90% marketing and 10% engineering (warning: estimate could have wide confidence intervals, but you get the point).

This was highlighted recently in a piece on LinkedIn about the millions of dollars being bet on an iPhone app called "Yo," which does nothing more than broadcast a text message to a friend or friends saying "yo."  That's it.

Writing the 'code' for such an application does not require any particular genius.  Someone has to come up with the idea.  Someone has to promote the idea.  Writing the software to do it is fairly simple.  I seriously doubt that a solipsistic interview process is going to overlook any necessary 'talent.'

Other 'hot' start-ups in the news include Tinder, which is a sort of internet dating site for those not mature enough for Match dot com.  It's in the news partly because of the buzz, but also because of the repulsive behaviour of the head of marketing.  Again, the company is not presenting any revolutionary new technology (Match dot com has been around for 20 years), but rather, it is offering more or less the existing world, packaged with a different marketing spin.

In short, 'tech' has evolved - it was once making transistors, then silicon chips, then the software to run the computers.  Now, it's about cobbling together a marketing vehicle disguised as an app and selling it.  Quickly.  Silicon Valley got its name, after all, due to hardware.  The last fab closed a decade ago.

In the start-up I helped to co-found back in 1997, our product was a mathematical analysis package.  It was developed not in some 'hot' programming language, but instead, in Visual Pascal, a language as old as I am.  The value we presented was in the maths behind the packaging, and we were rewarded with multiple patents in data mining.  Some of what we did in terms of programming was challenging, but few problems required particular software genius.  I recall clearly one example where the approach to a problem required some code insight rather than a mathematical one - I had to develop a UI that would dynamically place nodes in a decision tree on a screen that allowed users to click and change the analysis.  It was a bit tricky, and I recall debating with the head of engineering about my approach, which relied on a recursion algorithm.  He thought it wouldn't work, and I was sure that it would.  It turns out, I was right, but I suspect the disagreement was more because of my inability to articulate the method rather than his lack of vision.

But for the most part, programming requires a description of a problem (not, itself a 'tech' issue), for which a set of instructions can be coded to tell the computer how to execute (this is the 'tech' part).  The former can require vision; the latter almost never actually does.

What was a big part of our success, and contributor to our failures when they happened, was the cohesion of the team.  And this is where I think that Bueno undervalues the importance of 'culture.'

He may find it trivial that the programmers like to go out for a beer together.  The author of this famous piece describing the "ping pong theory of tech sexism" may ridicule the 'boys playing ping pong all day long."  But one thing I can tell you is undeniably true is that in a start-up, with very few exceptions, you will have too much work and not enough resources. That is going to mean that you will spend a lot of time with your colleagues, and relatively little time with your friends, your family, or your pets. And thus, like it or not, if people on the team do not fit in, there will be friction and bickering, and not of the constructive sort envisaged by the Harvard professors.

In Silicon Valley, people consider lack of a real life to be a badge of honour - people at times compete to boast about how little time has been spent with friends, at home, or even on basic hygeine and sleep.  

In any case, I left "Silicon Valley" many years before I physically moved away from San Jose.  I have been employed in creating models and clinical research for more than a decade.  Part of that is because what I found attractive in my 20s and single was considerably less appealing as a marrie 30-something.

There is a particular reason people choose to work in a start-up an not Microsoft.  Part of that is the culture.  And I suspect that, if the culture were to be changed, the industry itself would change in ways that are not necessarily 'good.'  The French are currently trying to incubate a 'start-up' culture in a massive former warehouse on the edge of Paris.

I say good luck to them, as culture in this sense cannot be grown in a petri dish, and it is going to have elements that people don't like.





Post a Comment