Monday, September 16, 2019

Leaving Google

I probably should write some definitive words about why I left Google. I guess I'll do that here, on my lonely blog.

Let's start with the short version. There are lots of things you could say about what a computer really is, about the meaning and purpose of computers. My favorite one, the one I find most inspiring, comes from Steve Jobs back in the '80s: “The computer is a bicycle for the mind.” I used to believe Google made more bicycles for more minds than anyone else. I don't believe that anymore.

That's it, this is the tl;dr line, you can all go home now. What follows is more words. More words from the guy that customized his Blogger stylesheet to get wider text so his paragraphs wouldn't look so long. More words from a guy that's said a few regrettable ones already.

I put Jobs' remark in quotation marks above but it's more of a paraphrase. He said things like that at least a few different times and often phrased it in different ways. I imagine at the time he was saying it he biked places sometimes, or at least had done so in recent memory. A couple decades later he was a rich old crank that bought a new car every week or so, so that he could always be in the grace period where he didn't have to install permanent license plates.

One of the other things you could say about the meaning and purpose of computers is that computers, like a lot of other technology, make people more efficient. The people that can make the most of this efficiency reap the most gains. Often existing powerful organizations are well positioned to make the most of technological efficiency; sometimes entrenched powers are the only ones that can do things. Sometimes upstarts do very well for some reason. Whatever that reason is, they then become powerful organizations that attribute their further successes to the same myths as the early ones. Google is like that.

And that's fine. Google is full of the myth that it succeeds in 2019 because it's full of smart people that make great software. I don't know, I think most stuff Google does succeeds mostly because Google is big and has an established user base? I'm not an expert in this stuff, but... come on. A lot of people can look really smart when they have access to all the stuff Google has already built up. So it's not like they aren't full of smart people that make great software, and if they didn't they'd fall off pretty fast! That's just... whatever, it's fine, it makes people feel good about their work. Sometimes the exceptionalism is grating; that's not why I left.

One of the funniest things in Google culture is the idea that Google is “data-driven”. Ridiculous! Google has made a lot of cool things throughout its history essentially as halo projects. They could do them because they were printing money selling search ads, and any eventual benefit to the bottom line was ultra-speculative. Some of them even worked!

And that's fine. GMail was in one sense a flex: look how much storage we can just give away! We won't even put banner ads in your inbox! All you have to do is log in — and now people are logging in to Google. I don't know whether people tried to use “data” to estimate the ultimate value of that to Google but... how would they even? How could you estimate the value of something whose payoff depends on so many other unknown factors precisely enough to prioritize the work against other things you could be doing? It has to be an intuitive belief. Meanwhile Google+, with its explicit goals around driving logged-in usage, flopped famously. Another way to look at GMail is as the result of laziness, impatience, and hubris. “Webmail sucks! Configuring SMTP sucks about as much! Spam sucks so much! I could do better!” The laziness and impatience were in tune with the zeitgeist; the execution matched the hubris.

You know, it's better than fine. Being data-driven is for robots. Be vision-driven! Use data to spread your vision! Use data to hone your vision! Use data to question your vision! But never let data grab the handlebars — the computer is the bicycle, the mind is the rider.

I got to work on Google Maps for four years. I'm sort of proud of that. I think Google Maps is one of the coolest apps in software history. I'll never forget the first time I played around with it. It changed the way I saw the world! I'd already been the sort of person that was easily fascinated by studying maps because of how much you can discern about the logic of a place by studying maps. Here was one big, map of the whole world!

This is probably the right place to say: sometimes thinking of the world through the lens of one big map misleads as much as it enlightens. That isn't why I left Google, or Maps, it's just a thing that's true: Google Maps can't be the only map. No single map can be the only map. Some people think about places differently than others, even in ways that contradict others. Some places, some institutions, work differently for some people than others. Like I said, Google Maps changed the way I saw the world. I could see all the places I might go running, crossing freely between towns, on and off of campus, through parks. But I didn't always acknowledge the perspective I had. Just randomly:

  • I remember once saying that essentially every place on the continent was connected by a network of pavement. My friend Heather challenged this as depending on a particular idea of what a “place” was, and I didn't really get it, but this was a pretty important point!
  • Mercator projection! It's... a pretty bad way to draw maps of big parts of the world! It's a pretty practical way to draw city-scale maps in parts of Earth that have cities (for people that want north-up orientations)! It's definitely the worse-is-better solution (“It is slightly better to be simple than correct”). Anyway Google Maps has probably made more people look at Mercator projections of big parts of the world than any other discernible thing in the history of maps? And now there's a 3-D mode... it still enforces north-up like a globe with an invisible stand you can't break free from. That's probably a reasonable choice! I just think we shouldn't forget it is a choice.
  • In the first season of Serial the host explained the local term “strip” as a particular named place where you might go to buy or sell drugs. I bet some such strips are in more common parlance than many neighborhood designations that appear on Google Maps! I actually used this as a conversation-starter with colleagues sometimes but had a hard time getting the idea across.
  • Of course an expansive view of the world showing near-total freedom of movement across political boundaries is not a reflection of reality for many people.

Again, these aren't reasons Google Maps is bad, they're reasons it can never be enough. We'll always need more ways to think about geography. We'll always need ways that a corporation will never give us. We'll always need ways that the software world is unlikely to give us (if you have an idea for something the software world is unlikely to give us, but that yet could use some software work, I may be able to help, with the caveat that I'm pretty useless generally; hit up my DMs or whatever).

Google Maps has some “halo project” vibes, like GMail. At times it's been used to flex: way before I worked on Maps they, for a time, connected the continents in directions by instructing users to, “Swim across the ocean.” As with GMail, who could possibly compete with what Maps offered? A fast site with a cool UI and, with no need to immediately make money, no annoying ads?

I guess that's one thing that feels properly sinister. Google at its best shows the world how to build really cool and useful stuff... that you couldn't actually build without the Google-sized money printing machine. And when you graft on the Google-sized big-data machine?

I have some particular frustrations around the way Google Maps chases users and gathers data these days but it seems to be tied to a general idea that runs through the biggest companies in computing. We aren't about empowering users. Users aren't our partners. They're sources of data (as long as it's the data we want when we want it). Their attention is the battleground where we fight against the other big companies to stay at the front the longest.

One of my particular frustrations has to do with driver distraction. I've always been uncomfortable with GPS screens in cars. I sometimes drive cars with GPS screens active on mounts around relay races and I find it distressing. There's a feeling when you're merging in traffic and monitoring a situation over your shoulder, then suddenly realize you haven't been keeping track of what's in front of you; for me, driving with a GPS screen often feels like that. Apps like Waze take this to another level, making driving something between a game and a social network.

I can't stomach Google Maps following Waze down that road: actively prompting users for input on road and traffic conditions while they should be watching traffic ahead. People are, in general, deluded about their ability to multitask. That is, to divide their focus and switch between tasks without losing accuracy, and to manage distractions and prioritize the most important things. One of the reasons I did leave Google is that I was very disappointed by Google's direction on this, and by responses to my concerns about them, which relied on pure fantasies about people's ability to prioritize in the face of distraction. We supposedly had principles about distraction; they ended up as little more than footnotes when it was time to break them. That made me really angry.

That loomed large for me but of course it was just one part of an overall change in how I saw Google and other big tech companies. The change in the overall picture is why it was a moment where I didn't want to be a part of Google, rather than a moment where I wanted to fix Google. Maybe fairly little of that change is really a change in what Google is or what it does. Certainly a lot of it is position: things that seemed cool as “betas” from small companies can look more oppressive from behemoths that are already heavily entangled in our lives. Some of it is experience: I didn't anticipate some of the ways software companies would change the world and I haven't liked some of the changes I thought I'd like. And some of it is certainly a change in how I see the world. I understand the significance of politics and power more, for instance.

For now I just want to do things that are more straightforward. Maybe make some bicycles for some minds if I can.