My grandfather was no slouch: he was a Rhodes Scholar and had successful careers in the foreign service and pharmaceutical industry.  But that’s not what I remember him for.  Instead I think about his absent-minded professorial air; his tendency to try to repair everything with epoxy; his fondness for hollering out Aura Lee after too many sherries; his kindness, which was so profound that it — and it alone — could bridge the bitterness of his kids’ divorces.  Part of the reason these traits took the foreground was that he was by nature a humble guy.  But a big part was that he happily shared a stage with my grandmother.

She was an incredibly impressive woman — possessed of a forceful mind, a New Englander’s unshakable belief in liberal values, and a sharp sense of humor.  Recently my aunt sent me an interview that she’d just discovered, in which my grandmother spoke frankly about her life as a foreign service officer’s wife (very much a career in its own right); her experiences being stationed in Senegal, New Zealand, Australia, Jamaica, Iceland and Belgium; her sense of Europeans’ comparative drinking habits; and our family’s run-in with some of the worst of the McCarthy era.  I think it captures a lot of what made her so great.  Maybe it won’t be of interest to anyone else, but believe me: if you’d known her, you’d be rapt.

More »

IBM’s victory at Jeopardy has led to the usual woe-is-humans handwringing, including (charmingly) from Jennings himself.

But look, while I agree that the logic of machines crowding out off-the-shelf humans is compelling, I think there’s some reason for hope.  I, and probably most of you, already carry a substantial neural prosthesis around with me all the time, which confers superhuman trivia capabilities at a level even more impressive than the Watson system.  It’s just that a couple of the system buses are a bit slow.  The neural bus will remain relatively sluggish, but for now it’s the interface that’s the real problem.

I think there’s a lot of low-hanging fruit that can be grabbed on this front: voluntary cochlear implants, haptic interfaces, mobile or ubiquitous ambient displays — but at some point we’ll be talking about neural interfaces, and the problem’s going to require technology that’s currently only speculative.  That’s going to be a difficult problem to solve.  More difficult than creating artificially intelligent beings?  Honestly, I don’t think we know enough to guess.  But I suspect the answer could go a long way toward determining whether or not our descendants spend their days fighting Terminators.

Let me add: it occurs to me that one interesting/tragic side-effect of the steadily improving quality of these interfaces is likely to be the disenfranchisement of older workers who are biologically incapable of utilizing these self-enhancing tools.  I don’t just mean in an “old people can’t work the VCR” way (though I think that may be related); I mean that neuroplasticity declines with age, and it’s conceivable that some of these technologies will simply be impossible to install in people with less malleable nervous systems.  My understanding is that there’s already some reason to think this phenomenon affects users of cochlear implants (though I’d want to do a bit more reading about it to be sure).

How are you going to compete in the workforce against someone who can Google without a keyboard?

Over at Grist Tom Philpott has written a few posts about the dangers of diet soda, and I’ve been acting cranky about them on Twitter. To be honest, it’s not that the posts are that bad, or even that I disagree with their gist. I tend to think that soda isn’t very good for me in general, and I avoid artificial sweeteners because (in order of decreasing importance to me) I don’t like the taste, have mostly weaned myself off sweeteners, and put at least some stock in the health case against them. So I haven’t really got a dog in this fight, other than general reactionary impulses. I am also not anxious to get into a big huge debate with Tom, because frankly he knows much more about food science than I do. Besides, he seems like a nice guy.

Still, these posts got my goat. I think it’s because they’re emblematic of a number of rhetorical tics that people use when trying to write scary stories about chemicals. And I hate that shit. The audience for that kind of thing is useless at picking or prioritizing their battles; I tend to think it leads to more anxiety, fewer vaccinated kids, and absolutely no effect on policy or health outcomes (other than those related to communicable disease, of course).

Also, I’m confident that Tom’s a smart guy, so the posts struck me as deliberately engineered linkbait, which I also dislike.

Anyway, the Twitter can of worms has been opened, so let me explain here what bugged me about these posts and made me feel that they were representative of bad online coverage of environmental toxins in general.

Cherry-picking research

This one’s easy. Confirmation bias is real (look around and you’ll see it everywhere, the joke goes). I particularly winced when I read the phrase “Italian researchers”. Admittedly, I don’t know the state of Italian biomedical research. But I do follow some alternative energy stuff, and I can tell you with confidence that that country produces more cold fusion “breakthroughs” per capita than anywhere else on earth. Plus, people smarter than me seem to think the Italian academy is a joke. But hey, if the abstract says what we want it to say…

Succumbing to this temptation is understandable, and I’ve certainly been guilty of it myself.  You need to be shut up in a doctoral program for a very long time to begin to possess any kind of immunity.  Frankly, I doubt that any of us are equipped to truly resist it.  But it’s still irritating to be on the other side of it.

Nomenclature as indictment

Particularly infuriating. Lots of chemicals sound scary, and health writers never resist the opportunity to deploy a 5,4-methylpropylbenzepene or whatever into a laboriously recited ingredient list. But it’s meaningless. The nomenclature got a lot scarier when chemists changed how they name compounds. There were good reasons for doing this (reasons that open data evangelists would find particularly appealing!), but the change does seem to have some unfortunately profound side effects on the human mind.

Similarly, pointing to constituent elements or the chemical class to which a compound belongs is rarely illuminating. At one point in the third post Tom discusses how caramel coloring is made by denaturing starches with heat and “ammonia-based” compounds. A shout-out to something you keep under your sink next to the drain cleaner can’t help but cause alarm. But of course fertilizer is also “ammonia-based”; and nobody seems to worry much when it gets turned into protein and eaten.  If you are Derek Lowe or someone like him, you can start to develop rules of thumb about the biological effects of compounds based on their structure.  If you aren’t, you can’t and shouldn’t.

Tom would probably object to this characterization, as his account goes on to discuss a byproduct of the denaturing process that’s carcinogenic. I’d say that that falls under “cherry-picking research”. In truth, lots of cooking processes produce carcinogenic compounds (the process he describes is not dissimilar to how pretzels are made). Remember the scare over acrylamides in french fries and other oil-cooked carbohydrates from a few years back (or the one over the char on your steak)? Suddenly it became apparent that we were all consuming boatloads of cancer-causing chemicals!  But it turned out that the situation’s more complex than that, and a study suggesting that one component of a food has shown carcinogenic effects (often only in an animal model) isn’t nearly enough evidence to justify a freakout over the food in general.

As another example, consider coffee, a beverage that Tom points to approvingly at the end of the second post. The roasting process produces a huge array of chemicals, not nearly all of which have been characterized and studied, and many of which have demonstrated carcinogenic effects (here’s a Wikipedia link, though like me you’ve probably already heard this elsewhere). But that doesn’t matter, because coffee is categorized under our semi-arbitrary “good/natural” heuristic, albeit perhaps with various qualifying words attached to it in our laughable mental medical tag cloud (parkinsons/osteoporosis/miscarriage/anemia/whatever).

Ignoring scale

Less relevant to this series of posts, admittedly, but a habitual mistake of this kind of writing. Humans are awful at understanding differences of even several orders of magnitude, and authors are even worse at bothering to convey them.  For our purposes, this is a problem on two fronts: both chemical concentrations and the scale of measured health effects. All you have to do is say “there’s uranium in your drinking water!” and people will start going nuts. You know what? There almost certainly is uranium in my drinking water, and yours too. There’s nasty stuff everywhere; you can’t avoid it. Things diffuse! Hell, cosmic rays transmute atoms! But until you quantify the scale and demonstrate the effect, these facts are meaningless. Bad things will get into your body, but usually in very small amounts, and your liver and kidneys will probably have no problem doing their usual kickass job of dealing with them (if the bad thing bioaccumulates, you should admittedly be less sanguine about this; but that’s not applicable to the compounds under discussion).

But this is not how the human mind works. You only have to invoke a substance and a sort of animistic power takes over. It’s homeopathy at an angle. It’s why we have airport security theater. Numbers mean so very little to our idiot monkey brains.

Pretending epidemiology hasn’t been invented

Perhaps I will be accused of a kind of libertarian placidity for this one. That would probably be fair. Still, I think that writers often lead with “we are all unwitting participants in a huge and unethical natural experiment!” and then decline to ask whether any data’s come in from this bit of mad science. And the fact is that it has. “It’s a mixed bag” is the answer that then invariably follows, but that’s sort of telling in itself, right? People have been consuming aspartame for a while now — some of them even doing so while participating in vast longitudinal studies — and it’s just not the case that it makes you start sprouting tumors or hemorrhaging spinal fluid or growing hermaphroditic genitalia. I’m all for getting BPA out of the lining of tins of baby food, but it’s pretty hard to make the case that any of the worrying maybes in Tom’s posts are actually manifesting themselves as health crises. Yes, someone needs to be manning the watchtower; yes, industry has incentives for bad behavior. But in this particular case I think the decades’ worth of rhetorical and scientific stalemate over sweeteners is enough to merit some shrugging and eye-rolling.

Diet sodas are not a serious health problem. They’re just not. It may be that avoiding them is a good idea at the margin — and like I said, I do exactly that, because hell, I don’t enjoy ‘em anyway. But it’s hard for me to take myself very seriously as I do so. Sure, I don’t like the idea of methanol in my body, but I dislike it in almost exactly the same way that Scientologists don’t like the idea of thetans in theirs. It’s just a neurotic tic; an echo of a sort of Jungian archetype about impurity; really, a hobby at best. I mean, I still drink alcohol, and exercise less than I should, and go out in the sun a lot when I’m on vacation. Despite this disregard for my health I am still probably among the healthiest people in history, thanks to having enough money and living in America in the era that I do.

But look, perhaps you are healthier still. That is definitely possible! If so, by all means, freak out about aspartame if you want. It just strikes me as unlikely to deliver a very good return on the worry you intend to invest in it.

Like Ezra, I find this explanation for Dropbox’s success intellectually attractive.  Minimalism!  Of course!  It’s so simple!  It’s exactly the kind of thing that software developers get off on. For one thing, it legitimizes our lust for Apple products — despite the fact that Apple UI betrays the idea more often than we admit — because most of us can’t or don’t want to distinguish between good aesthetics and simplicity.  It’s sort of zen.  It makes us feel wise.  But in this case I don’t buy it.

Here’s what I think happened.  Amazon launched its S3 service.  Suddenly you could buy Big Storage services at essentially marginal cost, immediately.  No physical capital expenses, and human capital expenses were dramatically reduced (now you need a programmer who can work with AWS, but not someone who can source, install and manage huge RAID arrays in a datacenter somewhere).

That happened, and then a bunch of people tried to resell this service with some paper-thin rebranding/UI work.  Kevin Rose and Leah Culver with Pownce, and some Flickr clone with too many O’s in its name, and whatever, a bunch of people.  Dropbox was the first — as far as I know — to come up with a funded business model that could provide a useful amount of synced storage for free.

Also — and this is something that the Quora answer completely underplays —  Dropbox is quite technically sophisticated.  It’s not just rsync on a minute cron, you know.  It’s hooking into filesystem interrupts to notice when stuff changes in the synced folder, and doing it natively on every major OS.  It’s got quiet but powerful ways of dealing with versioning conflicts.  It’s also doing all of this with a high degree of polish (I mean: Growl notifications, c’mon).  Plus it’s smart enough to do things like notice when it needs to sync within a LAN instead of over the net, avoiding complexities you might not have considered like NAT traversal.  It’s not that it’s so simple; it’s actually a very sophisticated execution.  It’s just that those parts aren’t necessarily visible (and no, many of its competitors were not as clever).

Now, this is minimalism, in a sense.  But it’s not the sort of minimalism pointed to in the Quora answer, which amounts to “Let’s offer fewer features than those other jerks and we’ll all get rich!”  It’s more about doing things that are sophisticated and difficult, and not wasting time on UI afterthoughts.

It was 2006 when S3 launched, but a few years isn’t THAT long for a specialized market to shake out.  Besides, S3 prices have been falling since its launch, so it could’ve either been a lack of investment in the synced-storage space or just the need to wait for a cheaper equilibrium point that delayed the rise of a winner in the space.  At some point a critical mass was reached and brand recognition took over.

Anyway, Amazon’s contribution to web infrastructure is the key here.  Its transformative field-leveling effect on the industry (and the web’s increasing reliance on it) is a story that ought to be explored more in the popular press.  AWS deserves a bit more of the concern that Google commonly attracts for its market power.