Congress and engineering

In the wake of SOPA/PIPA, there’s been some renewed talk about the necessity of getting more technically sophisticated people into office.  I’d like to see this happen myself, though I’ve generally assumed it would come about on its own as digitally-illiterate generations of legislators naturally age and are replaced.  But a passage in what I’m reading today is, uh, not encouraging on this score.

Keep in mind that much of this book is about exactly when and why it isn’t safe for engineers to assume that any property is infinite.

technology and humility

With the internetty part of SXSW well behind me, it seems like a good time to talk a bit more about the state of our community (if it can be called such a thing) and its origin myths.

It seems as though disillusionment and cynicism are growing within our ranks. Alex Payne does a nice job rounding up some examples. As noted here recently, it’s difficult to separate this sense from the process of growing older (Alex is not much younger than me). Still, I perceive it to be a real shift. I think there are specific reasons for it, and that its arrival was inevitable.

It’s an axiom of internet true-belief that things will get faster. Mostly, this is true. Moore’s Law is habitually misstated, but even when the details are mangled the speaker’s point often remains defensible. Digital technology really does get cheaper in a hurry, and this has real consequences for its reach and impact on our lives. Faith in this ceaseless acceleration reaches its apotheosis in the idea of the Singularity, an long-nascent idea popularized by Ray Kurzweil, which posits that technological change will reach a velocity beyond which meaningful predictions about human society are impossible to make. Many people guess that this will involve mind-uploading, but the point is really just that we won’t be able to anticipate the true nature of the silicon godhead. Kurzweil was at SXSW, but I can assure you that he wasn’t the only person in Austin hoping/expecting that his mind will live forever in a machine or otherwise achieve technological transcendence.

Many find this faith in technological acceleration to be cheering, but it’s got some problems. Perhaps most importantly, it ignores the reality of diminishing marginal utility. Put simply: useful interventions tend to get less useful as you apply more of them to a problem. Digital technology can do a lot of things, and its declining cost means it will do more and more of them. But it’s important to realize that many of the things it could do but didn’t back when it was more expensive weren’t done because they weren’t worth the expense.

My toothbrush has a microprocessor in it; it’s great, and it only cost me $50 or so. It times my brushing and it ramped up the strenuousness of its sonic ministrations when I first began using it. Two decades ago, no such toothbrush was available. Why? There are a number of possible explanations. Perhaps the technology didn’t exist? No, that’s not right, it certainly did. Perhaps the toothbrush makers of the past were less clever than the folks gathered in Austin? No — that seems improbable, or at least egotistical. I think a better explanation is simply that, at the time, the advantages of a microprocessor-enabled toothbrush didn’t justify the expense of building it. The expense diminished, so now we have them. But it’s the costs that fell; the utility has remained static, and meager. This is strictly long-tail stuff.

Put another way: the most urgent uses for a technology will be addressed first. That’s the beauty of markets. But a consequence is that, during a given technological milieu, engineers will generally find themselves working on tasks that increasingly seem trivial.

I don’t think developers should be discouraged by this, but I suspect that many of them will be. That’s too bad. There’s no shame in doing honest work for honest pay. I’ve argued before that software development is a trade comparable to carpentry. I still think that’s about right. Building someone a home may not be “innovative”, but it’s a necessary and important thing. There will always be meaningful work that needs to be done, and, if you pursue work that is especially difficult or neglected by the marketplace, even work that is novel and exciting.

For what it’s worth, I don’t think that ours is the first engineering discipline that’s had to go through this process. You only have to look at the automobile-triumphalist designs of Corbusier to understand how easily the early returns to an invention can be incorrectly projected into the future (in many such cases we should be happy to have gotten away with disappointment rather than outright disaster). Similarly crazed optimism affected observers looking forward to a future full of steel, supersonic flight, radiography and who knows how many other unhelpfully applied technologies.

Now, I shouldn’t overstate all of this; I don’t want to claim that Progress Is Over just because people at SXSW are self-congratulatory dopes. So two caveats. First, it would be a mistake to think that diminishing returns to digital technology means social progress in this realm must be stagnating. Marginal utility declines, but computational power ascends. Individuals are able to perceive the slope of either of these functions and use it to fuel their sense of utopianism or malaise. But to know whether society as a whole is slowing down would require combining the two, which would require more fully characterizing them, and only particularly learned econometricians are wise or foolish enough to attempt that sort of thing.

Second, I dropped the phrase “for a given technological milieu” a couple of paragraphs back, but I’m asking it to do a lot of work. It seems likely that there are thresholds to technological progress at which discontinuities occur — emergence, they call it. Adoption of smartphones, for instance, seems to have mostly been a question of a better kind of telephone/PIM getting cheaper and cheaper, but once everyone has a network node with a flexible UI on them at all times, it seems likely that huge social benefits will emerge that weren’t a consideration in the initial consumer-adoption calculus. But more on this in another post (one that will probably have the word “innovation” in it an embarrassing number of times).

To retreat to my point: I think that in IT, especially, we have a tendency to get confused about whether we’re building a locomotive or inventing the steam engine. It’s a willful confusion, it’s driven by vanity, and it’s now leading to disappointment — and will do so all the more as we find we have enough locomotives on hand.

let’s pause and promote something else (Open States!!!)

Since I might have your renewed attention, briefly (“but I thought this RSS feed was broken”), and since I’m already frantically promoting it on Twitter, let me draw your attention to this morning’s announcement from Sunlight’s Open States team.  In short: we’ve got a new mobile app, and our services now support all fifty states (plus DC and Puerto Rico).

Getting to fifty is a big deal, and I’m extremely proud of Sunlight and the Open States team for reaching this milestone.  No one’s done this before; it took us years to do it ourselves. This data existed, but in a completely piecemeal manner — there was no consistent location, API or data model. Getting it collected, cleaned and accessible will have real impacts for journalists, watchdogs and citizens.  If you saw Sunlight’s recent work on mandatory ultrasound legislation — well, suffice it to say that there’s much more coming on that front, and Open States is what’s going to make it possible.

This was a huge undertaking — honestly, I don’t think anyone knew if creating and maintaining a legislative data collection architecture pulling from so many disparate sources would even be possible when we started. It’s a tremendous testament to the contributors and developers who’ve made this possible, and the leadership of my colleague James Turk in particular.  Nice job, you guys.

SXSW: my panel is over / where are the margaritas

It went fine! The panel was located inconveniently in spacetime (far from downtown/first thing in the morning) and the room was huuuge (Texas, etc.) but we still managed to draw forty people, which I’ll count as a win.  Also, I’m pretty sure I didn’t say anything that’s going to get me fired, which is the only real criterion I use to judge these things.

Apropos of nothing, during preparation for the panel I discovered that the Census TIGER/LINE GIS dataset has a logo, and it’s totally awesome:

I feel like an embroidered Gosling/Drive jacket with this logo would get you a lot of action at the geoDC meetup.

SXSW: lessons from the distant past

Anil Dash is on stage right now talking with Nick Denton about a variety of different things, including the failed 2011 Gawker-wide redesign which, if you remember, was going to Change Webdesign Forever and then didn’t. (Don’t worry, I’m sure Pinterest’s look will echo through the ages.)

More to the nominal point of the session, this is from Denton:

“The idea of capturing the intelligence of the readership [now] sounds like a joke.”

This strikes me as right, and completely at odds with the optimistic, inclusive, democratic vision of our networked future that dominated SXSW in 2007.  But I think it’s true, and basically compatible with the (admittedly ugly) view I expressed back here, in which part of new sites’/network’s success hangs on their ability to attract early adopting elites who are genuinely better at the internet than the general online population (which arrives later, bringing advertiser dollars and ruin).

The net really is more democratic than other mediums. But that doesn’t mean it’s egalitarian, it just means the gating/sorting mechanism is (arguably) based on less arbitrary/unfair personal attributes than other parts of our society.  This is about as Randian an opinion as you’re likely to get out of me, but as far as I can tell it’s true.

Denton went on to opine that democratic moderation of comments doesn’t produce interesting content because it suppresses minority views; and expressed the view that anonymity is essential to both the production and consumption of Gawker’s content.  Which again, all seems right to me, though it’s a little bit sad.

MORE PROSAICALLY: Denton mentions that Gawker doesn’t have some super-secret or cutting-edge analytics platform. Editors use different things — all commercial — based on their own preferences and priorities.  Which makes me feel better about my general lack of interest in these tools (beyond the basics).

SXSW: honesty in open gov boosterism

@urbanfran and @fkh did a really nice job on their panel. For one thing, I love transit data. For another, it’s good to acknowledge that projects that dream of large-scale user adoption/participation need to grapple with what people actually want to do. They don’t want to come home, make dinner, and start researching potential wastefraudandabuse on, for instance.

SXSW: I ate some tacos

Also, went back to the airport and rented a car. As predicted, everything seems a bit better now. My substantive thoughts about this whole thing remain unchanged, but I’m at least once again able to blithely ignore the parts I don’t like.

SXSW: gloom

I’m at SXSW, and rather than tweeting about it incessantly, I think maybe I’ll just post some things here.

I suspect that my mood will improve with the weather and after eating some tacos, but right now SXSW is a very alienating experience.  I was last at this conference in 2007, with a bunch of people from EchoDitto.  Emily and I had just started dating. Twitter launched here that year, and although I never thought it’d take off the way it has, it did immediately feel like something special — maybe it would only ever be used by the kinds of people there in Austin, but for us it would be a new, enriching layer on top of reality. I was thrilled and amazed at what technology could mean and by this new, fascinating class of people who seemed to believe — quite earnestly and intensely — hunches about the future that I had secretly harbored but been half-embarrassed to express. I’d never known these people were real, but now I was really one of them and everything was going to come true.

That feels very far away right now.  The weather’s worse, and there are way too many people — they’re everywhere, all the panels and shuttles and bars are bursting.  They’re the same kinds of people, but by now everybody’s settled on the same dream — the mad-lib details vary, but the basic story is about startups and a kind of “innovation” that seems indistinguishable from a dreary march toward the combinatorial exhaustion of the current technological landscape’s possibilities.

I’m older, and it’s easier to see the contours of feasibility and to recognize how many of those dreams are unworkable or ill-advised or just mundane.  Websockets are a great technology, but they’re going to change interfaces, not lives, you know?  Real-time data and self-quantification are interesting and will be useful in a few domains, but reality is not as procedural and deterministic as programmers’ minds; most of these speculative possibilities will not pan out, even after we’ve built the dehumanizing foundations necessary to really put automated Taylorism through its paces and see what it can do.  There is a terrifying amount of talk about cognitive psychology results, and I’m starting to think that the actually-quite-limited predictive power of these models at an individual level is the only thing that can possibly save us from people like my twentysomething year-old self.

Well.  Like I said: tacos are indicated.  In a moment, off to the airport to pick up a rental car and escape the shuttles. Later: panels from activists, journalists and political folks, which I think will be both less inspiring and less frightening, both of which strike me as healthy things.