The Common Sense Trap

Lindsay Lohan’s third best movie of all time.

I recently started reading Duncan Watts’s book Everything Is Obvious. (Watts was an academic sociologist who recently took up research positions at Yahoo! and then Microsoft.) One core theme of the book is the failure of common sense:

The same difficulty of reconciling what, individually, appear to be self-evident beliefs shows up even more clearly in the aphorisms that we invoke to make sense of the world. As sociologists are fond of pointing out, many of these aphorisms appear to be direct contradictions of each other. Birds of a feather flock together, but opposites attract. Absence indeed makes the heart grow fonder, but out of sight is out of mind. Leap before you leap, but he who hesitates is lost. Of course, it is not necessarily the case that these beliefs are contradictory–because we invoke different aphorisms in different circumstances. But because we never specify the conditions under which one aphorism applies versus another, we have no way of describing what it is that we really think or why we think it. Common sense, in other words, is not so much a worldview as a grab bag of logically inconsistent, often contradictory beliefs, each of which seems right at the time but carries no guarantee of being right any other time.

The same kind of aphorisms are found in tech commentary. For a fun exercise, I came up with two aphorisms. To make this more interactive, try to come up with some examples that go with each.

1. A good product does one thing and do it well.

Dropbox. Instapaper. 1Password. These apps and services succeed because they focus on a core feature and implement that feature better than anyone else.

Indeed, all the recent positive reviews of the note-taking app Vesper repeatedly emphasize the fact that, despite lacking many features present in comparable apps, it does what it does so exceedingly well. (See, for a small sample, reviews from beta-testers Marco Arment, Shawn Blanc, and Federico Viticci.)

It sure seems that the one-thing-well philosophy encapsulates what it takes to be successful in the increasingly competitive app space.

Take a break.

Ready?

Let’s go to the next aphorism.

2. A good product is adaptive and versatile.

Think about the products that Apple sherlocked. No one needs an HDR or a panorama photo app once Apple built those features in. No one needs a notification manager app any longer either.

Most recently, Harry Marks dismissed the death of all the flashlight apps beautifully: “The implication that Apple is hurting developers by implementing similar features into its OS means the writer has no concept of history, nor understands the difference between a feature and an application. Versatile apps survive. One-trick ponies don’t.”

It sure seems that the versatility philosophy encapsulates what it takes to be successful in the increasingly competitive app space.

Wait a minute.

Didn’t Dropbox get sherlocked by iCloud? Didn’t Instapaper get sherlocked by Reading List? And, with the iOS 7 announcement, didn’t 1Password get sherlocked by iCloud Keychain? Do these products survive because they do one thing well, or because they’re versatile?

Of course there are grains of truth in both aphorisms. Some products succeed because they do one thing well. Some products succeed because they are versatile. Some, such as Dropbox, succeed because they have both. Neither truism, though, encapsulates what it takes to be successful. They are at most partial explanations.

The problem is that, since both aphorisms sound so good, it’s hard to not hear them as complete explanations. It’s obvious that flashlight apps got sherlocked because they’re not versatile. It’s obvious that Vesper succeeded because it does one thing well. Nothing more needs to be said. Nothing more can be said.

In each case, we reach for whichever aphorism that is most convenient to us, pronounce it as the beautiful truth, and forget about its contradictory twin. We fall into the common sense trap.

How can we escape? For one, we can stop repeating and relying on such aphorisms. Even today, people are still constantly reiterating variations of if you’re not the consumer, you’re the product. Stop! For another, we can be more cognizant of our own fallibility. Have we got the whole story? Probably not. What are the other similar cases? Does the same aphorism apply there? If not, why not? Asking follow-up questions is a good way to not settle with aphorisms as be all end alls.

Common sense aphorisms are the bread and butter of lazy parochial tech commentary. If that’s not the kind of tech commentary you want to read or write, try to avoid the common sense trap.

PRISMatic Misdirection

Let’s get the obvious out of the way. PRISM is scary. Really fucking scary. It is especially scary to realize the level of cooperation that government surveillance agencies received from all our favorite tech companies — Google, Apple, etc.

Still, I think Gabe Weatherhead’s knee-jerk response is misguided:

The danger of ad companies like Google and Facebook isn’t that they are selling aggregated personal data. The danger is that they are aggregating unprecedented amounts of information. I applaud Google for fighting against the Justice Department but the problem, as The Guardian has shown, is that Google will lose much more than they will win.

The problem is with the prominent mention of “ad companies”, which suggests that the scariness has to do with ads versus no ads. It doesn’t. Collection and analysis of user information is nothing new. Yes, Google and Facebook do it to serve targeted ads. But Amazon and Target do it too to sell more things. Indeed, one of the scariest part about PRISM concerns the data from phone carriers. Phone carriers are, of course, user-supported and not ad-supported.

How Not to Argue Against Google

The impending death of Google Reader predictably set off a wave of fury amongst infovores. Anger is, after all, an important part of the five stages of internet grief. I get that; I’m angry most of the time. Unfortunately, frequent byproducts of anger are unreasonable charges and illogical arguments.

Among these are: Matt Alexander calling what Google did with Reader in the RSS space “anti-competitive”, Marco Arment saying that Google engaged in “predatory pricing”, and of course all the usual people pointing to the fact that Reader was free to be the root of all evil.

The predatory pricing charge is strange. As the Wikipedia article Arment himself linked to explains, the predatory part of predatory pricing comes from the fact that a company uses low prices to gain a large market share… so that it can then raise the prices to an unreasonable level.

Clearly, that’s not what happened here. Yes, Google Reader did essentially own the RSS space. No, Google Reader never charged anyone a cent. The fact that Google is shutting down Reader makes it clear that predatory pricing is not even part of what Google wanted to do with Reader.

So what, you say. Arment may be a lesser Wikipedia economist than me, but Reader had a monopoly on the RSS market anyway. And everybody knows that monopoly = anti-competitive practices — even if “predatory pricing” turns out to be the wrong word.

Not so fast with that equal sign. Monopoly denotes a kind of outcome. Anti-competitiveness denotes a kind of practice. Not all monopolies result from anti-competitive practices. A natural monopoly is a kind of monopoly that results from an efficient market rather than any anti-competitive practices.

As far as I can see while scanning Wikipedia’s list of anti-competitive practices, Google Reader did nothing of that sort. Its pricing, as I said, was not predatory. Users were free to choose other services. There was no collusion. Users weren’t compelled to use Reader if they used other Google products.

Instead, Reader’s dominance of the RSS space looks to be to be a textbook example of a natural monopoly.

It’s so easy to forget now, but as Khoi Vinh reminded us, Google Reader was (and still is) a damn fine service.

Google Reader didn’t beat every other feed reader purely because it was free. Google Reader won because it was an extremely well-executed example of interaction design.

The technological superiority is even more apparent under the hood. Google Reader was (and still is) fast, powerful, and good at solving not-so-obvious problems like getting feeds from sites that don’t obviously have one. IT SYNCS. As many developers are discovering, Google Reader’s background plumbing is not easy to replicate now. It certainly was even harder to do then, when Google Reader came to its RSS dominance.

Finally — and this is the key — arguably it’s inefficient for the market to have many competing RSS plumbing services doing all the things that Google Reader does; it’s much more efficient for developers to build apps and secondary services on top of one. Indeed, that’s arguably what actually happened.

Be angry if you want. But let’s not distort the economic terms or revise history just to make irrational arguments that fit an agenda.

Feel free to talk to me on Twitter: @RagingTBolt.

What is Pragmatically Bad about Tech’s Diversity Problem?

Jamelle Bouie has an excellent piece in the latest issue of The Magazine about the source of tech’s diversity problem.

He points out that the lack of diversity in tech, broadly construed, has less to do with people’s explicit attitudes, and more to do with their implicit attitudes. The problem is perpetuated by little things we don’t pay much attention to, like network effects. As anyone familiar with the research on privilege is keenly aware, it is hard to acknowledge privilege — let alone to fight against it — because it is largely invisible, even to our conscious minds. (A good companion piece is Tom Morris on geeks and privilege.)

Tech’s diversity problem is still here. Most recently, Matt Andrews pointed out the existence of an all-male (and mostly white) tech conference. Shockingly, one of the organizers responded to his inquiry with: “I don’t feel need to defend this, but am happy with our process”. A firestorm ensued, and you can catch some sparks by clicking on the links that Andrews added to his post, at the bottom. (Don’t miss the contributions in the comment thread.)

I am less interested in stoking this fire than in tackling a common attitude that I see in the discussion. Basically, the attitude is something like: let’s grant that the lack of diversity in tech is morally bad; what is so pragmatically bad about it? In other words, from the perspective of a conference organizer who is neither interested nor disinterested in issues of social justice, why shouldn’t there be an all-male (and mostly white) panel speaker list?

I want to suggest that we can begin to see one way in which tech’s diversity problem is pragmatically bad through the recognition of the false-consensus effect. (Since the effect is well-documented, the Wikipedia entry, which is based on an academic review, is actually quite good.) This effect, roughly described by Wikipedia, refers to

a cognitive bias whereby a person tends to overestimate how much other people agree with him or her. There is a tendency for people to assume that their own opinions, beliefs, preferences, values and habits are ‘normal’ and that others also think the same way that they do.[1] This cognitive bias tends to lead to the perception of a consensus that does not exist, a ‘false consensus’.

Here is the causal pattern that I see, without the fine prints. First, implicit biases drive the lack of diversity. Then, the the lack of diversity drives the false-consensus effect. In the end, the tech world loses out on perspectives that would have added significantly to the conversation — perhaps by challenging our conceptions of normality and guarding us from overgeneralizing from our own experiences. If something like this is going on, then you should still care about the practical losses for tech even if you don’t care about the social injustices (but of course you should care about that too).

The matter, as always, is more complicated. In this context, a further complications comes from the fact that implicit bias and false-consensus effect can interact in dangerous ways. In a recent article in Journal of Applied Social Psychology, researchers found a direct relation between holding an implicit prejudice and perceiving a consensus about the acceptability of the prejudice. They then noted some consequences of this interaction:

[...] the illusion of strong support among prejudiced people can be expected not only to make them more willing to act on their views, but also more forthright in expressing their opinions, less prepared to compromise, and less likely to modify their attitude. Furthermore, consistent with spiral of silence theory, we can expect that people with nonprejudiced attitudes will be less likely to express their attitudes if they perceive themselves to be in the minority, leading their attitudes to be underrepresented in the environment [...]

Together, implicit bias and false-consensus effect generate another form of pragmatic badness from the vocalness of the minority who fail to recognize the moral badness in tech’s lack of diversity. Taking this into account, it is all the more important for us to speak out and fight against the problem. Even from a selfish perspective, there can be a more fun tech world if there exist a wider range of voices.

Update: At this point, a common follow-up is: well, what about merit? People’s judgments of merit turn out to be significantly gendered and racialized too. Amongst the earlier studies that convincingly demonstrated this is Steinpreis et al. (1999), which found that people rate a male CV more highly than an otherwise identical female CV. Most recently, Moss-Racusin et al. (2012) got a lot of press. But really, the effect is studied extensively and proven to be quite robust.

Feel free to talk to me on Twitter: @RagingTBolt.

You’re the Product: A Logic Lesson

It’s the internet’s favorite refrain:

If You’re Not Paying for It, You’re the Product

The latest person to repeat it: Andrew Sullivan, who is now asking for $20/year for his blog. Although he never quite says this explicitly, the insinuation seems to be that if you pay for his blog, then you won’t be the product. Sully is probably not the only one who thinks this.

Unfortunately for Sully, logic disagrees with him. Yes, it logically follows that you’re the product for things you don’t pay for, like Facebook and Twitter. But it does not logically follow that you’re not the product for things you do pay for, like Sully’s blog.

The fallacy that Sully commits is called denying the antecedent. As Wikipedia, everyone’s go-to logician, says:

One way to demonstrate the invalidity of this argument form is with a counterexample with true premises but an obviously false conclusion. For example:

  • If Queen Elizabeth is an American citizen, then she is a human being.
  • Queen Elizabeth is not an American citizen.
  • Therefore, Queen Elizabeth is not a human being.

That argument is obviously bad, but arguments of the same form can sometimes seem superficially convincing.

As the Wason selection task shows, we human beings are not so great at conditional reasoning. We are prone to making logical mistakes like denying the antecedent. It’s perhaps not too surprising that we make this mistake when we talk about business models for blogging or whatever. But it’s time to stop, for logic’s sake.

Feel free to talk to me on Twitter: @RagingTBolt.

Satisficers > Maximizers

People who write about tech seem to be obsessed with maximizing — attempting to find and get the best, always.

There is The Wirecutter, a website that is dedicated to find the best in everything tech (and some more).

Ben Brooks wrote about his obsession with the best a while back:

[...] it’s something that is an ongoing pursuit in my life: to get the best X that I can get. [...] It really is the little things that count, because if you improve enough of the little things (and the big things don’t suck) then pretty soon you are going to have a lot of great things going for you and thus you will be happier. This is the reason I often write about recurring topics on this site — it’s a documentation of my pursuit to find the perfect thing for me.

More recently, Dustin Curtis’s eponymous declaration of his obsession with the best has been making the rounds in the Appleverse:

[...] trust me: the time it takes to find the best of something is completely worth it. It’s better to have a few fantastic things designed for you than to have many untrustworthy things poorly designed to please everyone.

Marco Arment echoes the sentiment:

This is why I research and review everyday objects like light bulbs: I have no patience for poorly working, poorly designed, or low-quality products.

Patrick Rhone agrees too:

This is something I believe in strongly. The reason is simple, choosing the best is a final choice. A final choice means I never have to spend the mental energy on that choice again.

These quotes all converge on two themes.

  1. Best or Bust. If you are not getting the best, then you are getting something untrustworthy, poor, or low-quality.
  2. Bestness = Happiness. If you do try to find and get the best, even at the cost of spending a great deal of time and money, you’ll be happier in the end.

Both are false.

Best or Bust is simply a false dichotomy. In most instances, there is middle ground between what is the best and what is untrustworthy, poor, or low-quality. There is such a thing as being good enough — being above some acceptability threshold relevant to one’s needs and wants. Often, the second best in many categories are good enough. iPad 4 is the best, but iPad 3 is good enough.

Unlike maximizers, satisficers do not always attempt to find and get the best. Satisficing is about attempting to find and get what is good enough.

As the Wikipedia entry on satisficing recounts, Herbert Simon argued in 1956 (!!!) that, given our cognitive limitations, rationally we should be satisficers rather than maximizers:

He pointed out that human beings lack the cognitive resources to [maximize]: we usually do not know the relevant probabilities of outcomes, we can rarely evaluate all outcomes with sufficient precision, and our memories are weak and unreliable. A more realistic approach to rationality [that is, satisficing] takes into account these limitations.

In other words, even if you attempt to find and get the best, in an uncertain world you often won’t succeed. To say that the world of tech is rather uncertain is a serious understatement. Even if you researched endlessly and settled on the iPad 3 as the best tablet your money can buy, six months later that choice probably wouldn’t look so optimal after all.

Given human cognitive limitations, Bestness = Happiness turns out to be false too. More from Wikipedia:

Maximizers tend to use a more exhaustive approach to their decision-making process: they seek and evaluate more options than satisficers do to achieve greater satisfaction. However, whereas satisficers tend to be relatively pleased with their decisions, maximizers tend to be less happy with their decision outcomes. This is thought to be due to limited cognitive resources people have when their options are vast, forcing maximizers to not make an optimal choice. Because maximization is unrealistic and usually impossible in everyday life, maximizers often feel regretful in their post-choice evaluation.

In other words, you might aim for more happiness with maximization, but you won’t actually have more happiness. If anything, the attempts to find and get the best make you less happy in the end. Perhaps somewhat paradoxically, satisficing is the way to maximize happiness.

Once we recognize our own cognitive limitations, it’s obviously more rational to settle for good enough than to keep searching for and trying to buy the best. Be a satisficer, not a maximizer.

Feel free to talk to me on Twitter: @RagingTBolt.

Price Points and the Paradox of Choice

On the day of the 4-inch iPhone announcement, I was deeply tempted to rewrite and rerun some old stuff about silly tech claims (“3.5-inch screen is the perfect size!”) based on armchair sciences. The point is not just that these claims are wrong (though they certainly are) but that they are the results of terrible, yet typical, post hoc reasoning patterns. The purported reasons are just decorations for a predetermined conclusion.

Instead of a rerun, I thought I would show a new example of this kind of bad reasoning. Jim Dalrymple Shawn King writes (on the basis of lists made by Stephen Hackett):

[...] take a guess at how many price points Amazon has for the Kindle and then guess how many Apple has for the iPad. It’s hard not to imagine that Amazon is creating their own “Paralysis of Choice”.

Sounds plausible… until you actually look at the research on the paradox of choice.

A meta-analysis of the last 10 years of research concludes that there is zero overall effect size on the link between the number of choices and adverse effects on decision-making. (But see some criticisms here.) There is an overall effect size of zero because there are quite specific conditions that need to be met for the paradox of choice to arise: lack of clear categories, the presence of difficult trade-offs, and an induced time pressure (p. 419).

None of these conditions are met with the Kindles. There are clear categories: if you just want a reader, you can get the classic Kindle or the Paperwhite, and if you want something more, get one of the Fires. There is no presence of a difficult trade-off: again, your choice is fairly clear depending on your needs. Finally, there is no induced time pressure: the Kindles will be available for a long while and you can order whenever you want. So, if you look at the scientific literature more carefully, the paradox of choice is unlikely to arise for Kindles, despite the oh-so-many price points.

Oh, by the way, guess how many price points there are for all Kindles? 19. Guess how many there are for all iOS devices? 17. Not a very dramatic difference. Even just counting the iPads, there are still 8 price points. The difference between the price points of Kindles and iPads is nowhere near the magnitude of the contrasts used in the various experiments. For example, in the most cited demonstration of the paradox of choice, the difference is between a choice from 6 kinds of jam and a choice from 24 kinds of jam.

The point, again, is not just that Dalrymple King is wrong (though he certainly is). The point is to highlight the kind of irresponsible and lazy reasoning pattern that shows up again and again in the Appleverse. At least we’ll always have the “yep”.

Feel free to talk to me on Twitter: @RagingTBolt.