I’m confused about what sin Amazon is supposed to have committed — October 10, 2014

I’m confused about what sin Amazon is supposed to have committed

I don’t have time to write about it right now, but Matt Yglesias’s post today on why calls to fight Amazon’s ‘monopoly’ are misguided did hit the mark. I wanted to write something the other day when John Gruber predictably snarked in favor of the Justice Department fighting Amazon’s ‘monopoly’.

There’s no there there, seriously. I’ve been waiting patiently for someone to make a good case that Amazon has done anything wrong. Seems to me that their worst sin is … negotiating very hard against publishers? And using their market power to demand lower prices? This is good for readers, isn’t it? It makes books cheaper. Maybe you could argue that something which is good for readers is bad for authors, but *that requires argument*; it can’t just be asserted. I had this same problem with George Packer’s argument against Amazon a few months back.

To put it in perhaps a few words: whatever Amazon is guilty of, Wal-Mart is guilty of too. And I don’t see anyone pushing to break up Wal-Mart. They’re both just large retailers pursuing high volume and low profit margins, perhaps at the expense of their suppliers. That’s all. What am I missing?

My paranoid backup scheme — June 11, 2014

My paranoid backup scheme

Inspired, I think, by Marco Arment, I trebly back up my computers:

1. To a Time Capsule at home.
2. Via SuperDuper to an external hard drive that sits at work. (I couldn’t really tell you how this differs from just using dd(1), other than that it has a nice UI, only copies the diffs, and seemingly makes the external disk bootable. In any case, it’s great.)
3. In the cloud to Backblaze.

I have my laptop set to automatically back up to the cloud at all times, and my girlfriend’s laptop set to do the same. Then I use the Backblaze iPhone app to periodically ensure that all my backups are up to date. It’s awesome. The best backup is the one you never have to think about, and I definitely don’t have to think about this one.

…and if you decide to use Backblaze, too, I can get a cut. It’s great. I would never recommend a product I didn’t use enthusiastically, and I wholeheartedly recommend Backblaze.

(As it happens, I also wholeheartedly recommend my Time Capsule and SuperDuper, but they offer me no way to get filthy rich, like Backblaze does.)

Of all the stupid modern tribes to belong to, the tribe of the corporation is probably the silliest — April 23, 2014

Of all the stupid modern tribes to belong to, the tribe of the corporation is probably the silliest

On the occasion of Apple’s releasing their revenue numbers, it’s fair to point out that lashing yourself to a particular company is really stupid. Daring Fireball, for instance, exists to defend Apple and lash its detractors. Which is fine as far as it goes: I read DF every day, and I like his style very much. John Gruber is very much a part of the corporate-tribalism nonsense, and he makes a good living from it: people invite him to give talks to defend and explain Apple, and there are rumors (unclear how accurate) that he makes half a million dollars a year from it. And good for him.

Of course there’s a tribe on the other side, namely the tribe of Android. Inasmuch as I use Apple products, I guess I’m not a member of the Android tribe. I like Apple products.

But here’s the thing: this has nothing to do with me as a person. Yet the weird stupid modern tribalism requires that your choice of technology have something to do with you as a person. If you use Android, you probably have a neck beard, for instance. If you use Apple, you’re probably effete and eat kale. Or whatever. (Turns out I eat a lot of kale, you guys.)

Starting from this base of letting the technology determine your personality, the next step is to care very much about the companies that make them. I am supposed to be personally invested in the success or failure of Apple Inc. Turns out I’m not, though. I like their products. I will keep buying their products because I like them. If they go out of business, I will be sad, because then I will have to use products that I wouldn’t otherwise have chosen. Only, it seems really hard to imagine Apple’s going out of business, so … I guess I have no reason to be sad. Problem solved!

Apple doesn’t need your support. Neither does Google. Apple and Google will do just fine even without bands of true believers furtively tossing grenades at the other side. Use their products if you like them; don’t use them if you don’t like them; lobby the company to change things (in its dealings with Chinese manufacturers, for instance) if that’s what you want. But defining yourself as an “Apple person” or an “Android person” is just pathetically demeaning to your stature as a human being.

Does Heartbleed mean that C should die? — April 12, 2014

Does Heartbleed mean that C should die?

The “Does that pretty much wrap it up for C?” piece (via my man Jamie Forrest) is interesting, but I think he needs to talk it out a bit more. I mean, at *some* level, *someone* is going to have to do memory allocation on bare metal. And what do we do then? And there are always going to be functions that need high performance, because they’re in the middle of some tight inner loop. Or in the SSL case, *someone* is going to need to do very specific things with memory, like making sure it’s not holding any sensitive data.

My understanding of modern malloc implementations is that they include all kinds of sophisticated ways to prevent buffer-overflow attacks. When you request a block of memory, they set it up such that requests past the end of your block cause a segfault. Or they randomize the blocks they give you, so that you can’t just grab the next few bytes and expect there to be anything there.

I’m not a C programmer (I really need to know it, I think, to be a complete programmer), but all of this says a couple things to me:

1. If you use the right libraries, you should be protected against a lot of stupid behavior. Makes you wonder, for instance, why the OpenSSL team wasn’t using tcmalloc or ptmalloc. I’m sure there’s a reason; I just don’t know the problem space well enough to say.
2. Any serious software system, whether down at the bare metal like C or higher up like Python, is going to require lots of testing, regardless of whether it’s got compile-time type safety. There should be lots of unit tests. Ideally, the unit tests would also be able to simulate other components, using mock objects and whatnot. And then you need integration tests to see how well your component integrates with others. And then, in the case of a secure system, you probably need to bombard it with very focused buffer-overflow attacks, written by dudes who know the code inside and out. (Sort of like penetration testing within a company, on the assumption that you’re most vulnerable to your own employees.) And for performance reasons, you should also test it by bombarding it with millions of requests per second and seeing where it breaks. Testing is hard. QA is hard, and is very often not respected as a peer of engineering. Engineering is sexier. If you’re really good at QA, you’re spending your time writing systems to test many thousands of cases rather than just grinding out the same manual test over and over, and you’d probably rather be off building something new. Engineers also feel this way: they’d rather be writing new versions of the code than maintaining the old stuff.
3. An ideal team will learn from its mistakes and build systems that prevent the same bug — or similar bugs — from reappearing.
4. Building good software requires a good organization and good management (whether by “management” we mean someone who’s controlling the work product of his direct reports, or something broader like “group structure”). This is a variant of Conway’s Law: “Organizations which design systems are constrained to produce systems which are copies of the communications structures of these organizations.”

Let me be clear that I say all of this with absolutely no understanding of the OpenSSL code base, much less an understanding of the OpenSSL team’s structure. But it just strikes me that blaming an OpenSSL bug on the C language doesn’t really get at the problem. A successful software system will fix this mistake and ensure that it never happens again. A successful *open-source* software system will take community direction to build such a resilient system, and will do it all with a fully open process. That goes beyond narrow issues of language choice.

Merging and de-duping contacts under iOS and OS X — April 5, 2014

Merging and de-duping contacts under iOS and OS X

I guess I’m some sort of completist when it comes to having accurate contact information for my friends. OS X has made it really worth your while to keep that information up to date, for a few reasons.

First, if you have a friend with an oddly spelled name, and you don’t have that person’s name in your contacts, then OS X will flag the spelling of your friend’s name with the dreaded red wavy underline; it flags no spelling problems if you have them in your contacts.

Second, if you keep all the various ways of addressing a given person under the same contact, OS X lets you search for that person using every available identifier. For instance, I have a colleague whose real name is ‘Junrui’, but whose nickname is ‘Hench’; if I’m diligent about including the ‘nickname’ field and the real first name in his contact entry, then I can search Mail.app (inter alia, presumably) using either the nickname or the first name. The same goes for an old friend, for whom I have eleven email addresses and two family names, one hyphenated. If I don’t keep these contacts merged, then I need to remember whether the email message I’m searching for came from him when he was just “John Smith” or had become “John Smith-Jones” (note: not his actual name). With merged contacts, you can just search for ‘Smith’ and be done with it.

Third, the Calendar app can display a feed of contacts’ birthdays, which you can selectively choose to display or hide. I like that much better than having to keep a separate annually recurring calendar entry for every one of my friends. Instead I just stick their birthdays (and anniversaries) in the appropriate contact entries, which is where they belong.

Social networks make things a little more complicated. iOS lets you pull in contacts from your LinkedIn, Facebook, and Twitter accounts. To some extent this is all to the good: now your contact entries will have your friends’ Facebook profile pages, their Twitter handles, and so forth. But now you also have multiple contacts for each person: one contact for John Smith via LinkedIn, one for John Smith via Facebook, and so forth. Annoying.

Enter iOS and Mac contact merging. This is useful, but frustrating. You can merge contacts on OS X, and merge them on iOS, but I have thus far not been able to make them sync as smoothly as one would like. Syncing between iOS and OS X is otherwise quite smooth, using iCloud: you change a field in the contact on your Mac, and by the time you come to your iPhone, the contact is updated there. Great! Except that merging contacts under OS X doesn’t seem to merge the corresponding contacts on iOS.

So then, if you’re as (apparently) anal about this as I am, you go over to your iPhone and merge there, too. (You do this by editing one of the contact entries you want to merge, going to the bottom of the entry, selecting ‘Link Contacts’, and then selecting — one by one — all the other contacts to merge.) This is twice the effort, because anal.

Now you have your contacts merged on both devices. New problem: how does the Apple universe de-duplicate entries? Suppose you list your home address on Facebook as “Cambridge, MA 02139”, whereas you list it on LinkedIn as “Cambridge, MA 02139-3986”; how do you de-dup that? At least two plausible alternatives come to mind: either include *both* addresses in the merged contact, or merge them into a single address with more information in it — in this case, a single address with the ZIP+4. iOS seems to go for the first option: include multiple nearly identical addresses. Which is a little annoying because, if you’re OCD about this sort of thing (as I apparently am), you are then very tempted to then go through and de-dup everything properly.

I don’t fully understand the internal logic of merged iOS contacts. I think they’re designed with the assumption that you can un-merge them later; depending upon how they implemented that, it may mean that you can’t delete a superfluous address entry: you may un-link the cards later, and will at that point need to recover the original address. So it’s possible that when you delete a duplicate field in a merged contact entry, iOS keeps the deleted field around, hidden, and available for un-hiding if you decide to un-link. It’s confusing to me.

While writing all of this up, I discovered some oddities about the merging/syncing logic:

* OS X has a “maiden name” field in contact entries. That’s handy: again, I may want to search for “Jane Smith” before she became “Jane Jones”. It doesn’t seem like maiden names propagate to iOS contacts.
* Duplicate home addresses seem to show up in iOS but not under OS X.
* If I mark Jane Jones’s maiden name as Jane Smith under OS X, I would expect that I could search Mail.app for Jane Smith, and get all of Jane Jones’s emails in the same result set. It seems to not be so. I was really excited that OS X had seemingly mastered the notion of “identity”, whereby Jane Smith and Jane Jones are really different labels for the same identity. I think maybe OS X has a different notion of identity than I do, though I could probably defend their notion of identity if I were called upon to do so. Maybe, for instance, it’s sensible to treat Smith and Jones as different identities, because maybe sometimes you want to search for things from a particular period in Smith/Jones’s life; in that sort of case, you’d want to make a distinction between Smith and Jones. That said, see the next bullet.
* If I type “Smith” into Mail.app’s Spotlight search field, it doesn’t return me any results from the contact whose maiden name is listed as Smith. That seems odd, and cuts against the interpretation of OS X identity that I posited in the previous bullet: it can’t be that OS X is trying to keep your options open by giving Smith one identity and Jones another. Smith’s identity seems to have disappeared altogether.
* Actually, now that I focus on it more, I notice that OS X doesn’t unify identities at all in the way that I’d hoped. Again supposing I have several email addresses for Smith/Jones, one of which is jane@smith.com and the other of which is jane@jones.com, I would expect to be able to search Mail.app for “emails from Jane Jones” or “emails from Jane Smith” and return email messages from any of them. It seems to not be so.

Some friends and I experienced no end of trouble, along similar lines, when we tried to carry on a long-running group chat via iMessage. For reasons that still make no sense to me, a sequence like the following would happen often, whenever there were more than two people on the thread:

1. Person A would start chatting with people B and C on message thread 1.
2. Person B would reply to A and C.
3. Person B’s reply would show up on A and C’s devices on message thread 2. (I say “devices” rather than “iPhones” because at least two of us were trying to conduct this conversation from *both* an iPhone *and* a Mac. I don’t remember whether that fact caused any of the particular problems that we ran into. At this point, even the memory of our difficulties is frustrating enough that I don’t want to revisit it.)

They wouldn’t be threaded together. This is incredibly annoying. It seems like ultimately this is related to Apple’s confused notion of “identity”. Again, it seems to me that your device ought to treat anything that comes from me — from any of my email addresses, phone numbers, Twitter handles, etc. — as an emission from “Steve Laniel”, the metaphysical entity. The fact that one of the senders happens to be steve@stevereads.com, while another happens to be steve@laniels.org, should be immaterial. I think this distinction confused iMessage, even though I think all of us were diligent about telling iMessage that any of our phone numbers, email addresses, etc., should all be valid senders for iMessage purposes.

Again, there may well be good design reasons not to unify people into “identities” of this sort. Perhaps there are certain cases where it makes sense to unify, and others where it makes sense to separate. I would just like to see Apple defend its separate/unified identity design decision. But Apple doesn’t, generally speaking, explain its design decisions to anyone, so I’m not holding my breath. All I know is that the results are often confusing, as it stands.

I’m tempted to say here that Apple is very good when they control the entire experience, start to finish: if the universe of your usage lies within Apple products, then they’ve got the user experience handled. As soon as you need to incorporate the outside world in some way, though — by syncing across devices using Google’s sync services; by using multiple email addresses from non-iCloud providers; by loading iCal calendar feeds from other providers; by pulling in contact information from LinkedIn or Facebook — the experience gets much more muddled, and I don’t know that Apple does terribly well. They’re getting better over time, and I have hope that they’ll eventually get it working well enough that a lot of the muddle I described above will go away. They need to, because over time our mobile devices and computers are incorporating more and more sources of information in more and more complicated ways. Apple has historically done very well at forming a coherent whole from its own universe of products; I now really hope it’ll form a coherent whole from people’s online identities.

I wonder if Five-Whys-style sessions ever get this far — March 7, 2014

I wonder if Five-Whys-style sessions ever get this far

We join our conversation in progress:

“This bug slipped through because the product was poorly QAed.”

“Why was the product poorly QAed?”

“Because our QA is poor.”

“Why is our QA poor?”

“Because it’s a low-caste job.”

“Why is QA a low-caste job?”

“Several reasons, all of which generally fall under ‘QA isn’t respected as much as engineering.'”

“Why isn’t it respected as much as engineering?”

“Well, it’s not paid as well as engineering is, for one thing.”

“Why isn’t it paid as well as engineering?”

“Because what they’re doing is fairly rote testing, rather than actually building test harnesses and so forth. If QA were actually engineers, they’d be paid like engineers.”

“Why isn’t our QA actually engineers?”

And *that* is where the real discussion should start.

(Hat tip to my friend Dan Milstein, who’s all about the 5 Whyses.)

George Packer’s piece about Amazon is terrible — February 14, 2014

George Packer’s piece about Amazon is terrible

I have a deep love for George Packer’s work, going back many books. [book: Blood of the Liberals] is one of my all-time favorite books, though viewed in the context of the other books and articles he’s written, it’s condescending: *his* deep desire to find a liberalism that can appeal to the common man is somehow at odds with everyone else’s. Likewise, when George Packer came to oppose the War in Iraq, *his* ultimate decision to oppose it was thoughtful and well-meaning, whereas everyone else’s was reflexive and irrational.

The piece on Amazon is a long string of [foreign: ad hominem]s, including the requisite slam on engineers: “Everyone there is so engineering-oriented. They dont know how to talk to novelists.” That one example, it seems to be, contains the key to what’s wrong with the whole piece: throughout the piece, you ought to be asking, “Compared to what?” People with an engineering focus can’t talk to novelists, sure. So I assume Random House is filled with artsy types who are willing to forego a profit to take a flyer on some unknown, promising author? I have no experience in the publishing industry, but I am willing to wager huge quantities of money against that premise. Take a nice anonymous survey of authors — including aspiring or failed authors — who’ve worked with large publishers and let’s see what they think of the publishers’ author-friendliness.

Amazon is terrible for local bookstores, sure. But compared to what? How about you Google for [bookstore market share 1998]? Up comes a [newspaper: New York Times] article from that year titled “Independent Bookstores Struggle Against the Tide”. Quoth that article:

> In Tarrytown, the American Booksellers Association, a trade association, reported that while the independents held a market share of 31 percent in 1991, that number had dropped precipitously to under 19 percent five years later.

So let’s not romanticize the world that Amazon inherited. It was dominated by Borders and Barnes & Noble. At one point there was Waldenbooks, too.

It’s hard to find a sentence in Packer’s piece that doesn’t contain a tendentious interpretation of data that we all already experience. To pick just one:

> The digital market is awash with millions of barely edited titles, most of it dreck, while readers are being conditioned to think that books are worth as little as a sandwich. Amazon has successfully fostered the idea that a book is a thing of minimal value, Johnson said. Its a widget.

Did Amazon create this attitude? Who you gonna believe: George Packer, or your own lying eyes? Here’s what my impression tells me: back in the day, I used to buy music CDs, each of which I treasured and obsessed over. I’d buy an album or two, then spend the next few weeks digesting it lovingly. I’d read all the liner notes; I’d listen to it until I’d memorized every lyric and every last bridge. Then MP3s happened. Now I don’t see any liner notes; I don’t see cover art. For a time I used Napster, which allowed me to get unlimited access to free music. Each individual track, then, was valueless — literally costless. Nowadays I use Amazon MP3s, where most tracks cost $0.99. I also use Rdio, which allows me to stream most any song for free. I’ve used Songza and Pandora for similar purposes. In fact the default now seems to be that music is free (ad-supported). I don’t know, but I assume none of these services pays artists particularly well.

So music, in any case, has long since moved from a model where each work was an individual perfect snowflake to a model wherein it’s all basically wallpaper: you can get all the music you want at any time of day or night, and each individual track is a fungible commodity.

As for print media: you could argue that Amazon turned any individual bit of writing into a commodity, but that’s obvious nonsense. The presence of blogs had much more to do with that than did Amazon. The fact that I can get my hands on any newspaper from anywhere in the world had much more to do with that than did Amazon. My Instapaper queue is enormous, meaning that I have mountains of fungible text awaiting me. Amazon did not invent the commoditization of everything electronic; the Internet did.

Listen, I’m not 100% happy about this. I wish this blog post you’re reading were the most brilliant thing you’ll read all week. I wish you carried it rolled up under your arm; that you highlighted interesting passages with a pen; that you photocopied it at work and shared it with all your officemates. But that’s not how the world works anymore. Someone writes an article in a newspaper or magazine, and within hours thousands of blogs have digested that article for you. Now you can choose among thousands of blogs, each of which approaches that newspaper article from thousands of perspectives. I’m sure I could find a libertarian gun nut’s take on George Packer’s piece if I looked long enough in benighted corners of the Internet. But the point is that it’s all a commodity now. George Packer’s apparent dream, wherein each book is treated as a perfect, crystalline work of art, is many years out of date.

I think it’s even out of date for beautiful books published by reputable publishers. Years ago I read Yochai Benkler‘s [book: The Wealth of Networks], which is really a lovely book published by one of the best university presses. A university press! Shouldn’t they be the last bastions of hope for authors who aren’t chasing a profit? University presses are associated with non-profit institutions whose goal is to contribute to the sum of human knowledge. Yet even Benkler’s book, which is sort of a landmark in the field, was horribly edited. When I asked around about this, the word I got back is that authors are now expected to do their own copyediting, and are expected to submit “camera-ready” manuscripts. So much for publisher support.

At the risk of waving my hands too broadly at “society”, I am willing to blame capitalism just as much as I’m willing to blame Amazon or the Internet. Eventually everything becomes a commodity. Eventually everything gets driven down to marginal cost. If Amazon didn’t do it, someone else would.

I don’t mean to absolve Amazon. In fact I mean to praise them, which Packer is somehow unwilling to do. Everyone loves Amazon, right? They’re good for customers. Here’s Packer:

> Those were sweet words for a company that declares itself to be Earths most customer-centric company. Even its bitterest critics reluctantly admit to using Amazon, unable to resist its unparalleled selection, price, and convenience. When Bezos talks about serving the customer, its as if he were articulating his purpose in life. The customer is almost theological, James Marcus said. Any sacrifice is suitable for the customer.

That’s basically the extent of Packer’s praise for Amazon, which is incredibly odd. This is a company whose success is changing one industry after another, and Packer’s piece somehow attributes all of that success to malign influence — such as the infamous engineering attitude. How about this: it’s succeeding because it does right by its customers? Packer is just unwilling to admit that, because it would undermine the entire rest of the article — an article whose premise is that Amazon’s band of barbarians is toppling a once-civilized industry. This premise gets no support from Packer’s article. And somehow Packer never really addresses a question that ought to be fundamental: how could something be good for readers and not good for books, or good for authors, or good for publishers? I’m open to the possibility that there’s a conflict there, but really Packer ought to be asking: does Amazon make people read more, or less? Offhand, I assume that it makes them read more, because books are now cheaper. When people read more, that’s good for authors and publishers. That seems to be the fundamental calculus here, and Packer never once addresses it squarely.

I hate to say it, but Packer’s piece is garbage. I hope you read it in costless form on the Internet. When you do so, I expect that Packer will shed a single proud tear for the commoditization of his heretofore priceless work.

__P.S.__: The costlessness of the Internet means that my own blog post here is just a disposable commodity. Either I learn to deal with that, or I don’t. The dialectic doesn’t care especially much what I think about it.

I submit the radical hypothesis that communications media ought to be used insofar as they’re useful — February 11, 2014

I submit the radical hypothesis that communications media ought to be used insofar as they’re useful

Austin Frakt says conference calls should be more like Twitter. I agree! I would go further: I think that people should use each medium exactly so far as it’s useful, and should use each to augment the others. For instance, I believe that people should use email up to the point that everyone is talking past everyone else and the conversation is muddled, but no further. If there’s information to be conveyed that can’t be conveyed via telephone, I believe that it shouldn’t be conveyed via telephone. Likewise, if a rapid back-and-forth would be useful, I believe telephones should be used for that. If the conversation is more asynchronous, particularly if it’s happening across many time zones, I believe email is valuable for that. And I believe conducting part of a conversation over email, part over voice, is the right idea.

My “some methods are appropriate in some contexts, some in others” approach has made a big splash approximately nowhere.

Three humble requests for the creators of online stores — December 19, 2013

Three humble requests for the creators of online stores

1. You don’t need to ask me for *both* my city/state *and* my ZIP. The latter implies the former. We have the technology.
2. If I enter a ZIP+4, and your system isn’t equipped to handle 10-character ZIP codes, just silently drop the final four digits. Here, I’ll give you the Perl code you need: $zip =~ s#(d{5})-(d{4})#$1#; . You can thank me later. (I gather that it’s similar in PHP.)
3. Don’t tell me to leave spaces out of my credit-card number. Again, we have the technology. ($credit_card =~ s#s+##g;)
4. Just don’t create a retail website. The only company that would be allowed to create a retail website in Stevetopia would be Amazon. Maybe Google.

Facebook Graph Search is pointless, says Farhad Manjoo without meaning to — January 16, 2013

Facebook Graph Search is pointless, says Farhad Manjoo without meaning to

Farhad Manjoo’s piece about Facebook Graph Search is the best possible case for why we should care about Facebook Graph Search, which leads ineluctably to the conclusion that we shouldn’t care about Facebook Graph Search. Here’s the sort of thing that Facebook Graph Search lets you do, according to Manjoo:

> The most interesting searches werent necessarily the most complicated, but those that asked Facebook to combine its knowledge in ways that other sites cant. In an effort to suss out authentic cuisine, I tried, Mexican restaurants in New York liked by people from Mexico.

Let’s pick apart what’s wrong about this. First, the problem for any search engine is that it has to be not only *better* than Google, but better than Google by enough of a margin to make switching worthwhile. So you have to ask yourself whether searching like this is going to get you better results than just Googling for [authentic Mexican New York]; and if it does get you better results, will it require a lot more effort to do so? How about if you, like many people, eventually find websites that you trust for food questions? (Me, I use Chowhound.) How easily can you answer this Mexican-restaurant question by narrowing your search to Yelp?

Much more fundamentally, though, the problem with Facebook Graph Search is that it’s for people who don’t want to interact with people. Look at Manjoo’s Mexican-food example: I’m supposed to ask Facebook to search among those people who are from Mexico. Why wouldn’t I, instead, just *ask people from Mexico?* It’s supposed to be a *social* network, right? Like, with people? So shouldn’t I ask people things? Manjoo wishes he could ask FGS about “running shoes liked by people who have run marathons”; I know what I’d do in that case: I’d post on my friend Laura’s wall (or better yet, email Laura), “Hey Laura: which shoes should I get?”

In its defense, maybe FGS will not just query what my friends like, but rather query what all Facebook users like. But if that’s true, FGS is even *less* valuable: I *don’t care* what all Facebook users like. I hardly care what’s outside my own network, *contra* Manjoo’s example of searching for “photos of friends of friends who like Girls who live in NYC who are single women between 20 and 34 and like Arcade Fire.” That is a question that no one has ever had or will ever have. Either you’re querying the tastes of people you know, or you’re querying the tastes of strangers. And if you’re querying the tastes of strangers, why not use a site that makes no pretense about being “social” — like Google or, in that Arcade Fire dating example, like OkCupid?

The useless use cases come thick and fast in Manjoo’s article. I’m thankful for that, because he’s trying as hard as he can to show why we should care about this thing; given that no one could possibly care about these examples, there must be no reason to care about FGS. Like this:

> What captivated me was Facebooks search interface. Its unlike any search box youve ever used. Googles search is based on keywords. If you type restaurants chicago into Google, it guesses that youre looking for restaurants in the city even though you typed just two nouns. Facebook, by comparison, wants you to connect nouns with verbs. You can ask, restaurants in Chicago, or restaurants liked by people who live in Chicago or restaurants liked by my friends who are from Chicago. (The search box offers drop-down suggestions as you type, so you dont usually have to finish writing these full queries.)
> This method of searching is instantly intuitive. After just a few queries, I started asking the engine for more and more complicated things, just to see if it could keep up. I tried: My friends of friends who work in Palo Alto, California and are from California and are male and who like Indian restaurants.

Is natural-language searching something that anyone needs? Was it a problem that Google required you to search for [restaurants chicago] rather than “I would like to know where to eat a meal in Chicago”? This is a natural-language solution in search of a keyword problem.

If you want to know “restaurants liked by my friends who are from Chicago”, shouldn’t you *ask your Chicago friends which restaurants they like?*

The fact that Facebook and Manjoo think there’s a problem a computer can solve here is bottomlessly sad to me. It exposes a deep poverty of interpersonal relationships.