This post is cross-posted from my joint newsletter with Ranjan Roy, The Margins. Please check it out, and consider subscribing.
Two weeks ago my co-host wrote about the digital exhaust, and mentioned how a surreptitious Nest thermostat can keep tabs on the new owners of a house. I’ve experienced something similar myself. My previous partner had a Google Home smart speaker in our living room. After I moved out, it took me a few days to realize that I was still logged in to the Google Assistant on my phone and could literally see what she was saying into the speaker. There wasn’t anything particularly scandalous, yet idly observing the activities of your former partner from your phone, albeit in extremely low fidelity, has a tinge of voyeurism to it.
Ranjan’s post was more about the “exhaust”, the data that gets inadvertently generated and forgotten. Yet, there’s even a more fundamental issue that I think that deserves attention here; that is identity management.
It’s hard to pinpoint a number, but most people seem to have around 100 or so accounts online. My own highly biased Twitter survey of people who use password managers puts that average to over few hundred. That is an obscene number of identities for a single person to handle.
I shouldn’t have to spill more ink on why you should use a password manager, and how the initial minor pain of setting all that stuff up on your devices pays off huge benefits later. But this is my soapbox for now: You should use a password manager. I use 1Password on all my devices, enable 2-Factor authentication where possible. I have in my memory 3 passwords only, and they are actually all passphrases.
There’s a part of me that enjoys watching this rather complicated (if not convoluted) setup work like butter with FaceID and TouchID and all the other Apple’s biometric wizardry. As much as it creeps me out that my phone is taking a biometric photo of me every time I open up WhatsApp, I enjoy being able to pay for the Tube in London with a combination of mathematical models of my face and some radio waves. I’ve paid for this iPad I am writing on by buying it on Apple’s website, which used TouchID on my laptop. The entire flow feels both cool, and secure.
But there’s also another part of me that finds this setup insanely complicated and brittle. For every website that 1Password’s browser extensions work with, there are a few more where I have to copy a password from one app, and paste it into another. The mere digital UI trickery involved in generating correct identities in 1Password with the “Website” field set in is barely within my reach, and I’ve built such UIs myself for years. The way 1Password app matches the passwords it has on file to the accounts I have on different services is smart, but it does require you to understand it fully (so maybe not so smart).
Moreover, I live in constant fear of somehow my database of passwords across my devices getting out of sync or losing all my devices at the same time. Every time I enter a new password in one device, I make a mental note to open up 1Password in the other devices to make sure it gets picked up.
This stuff is just bonkers.
And this is just the tip of the iceberg, that I have some a modicum of control over, and tiny bit of visibility. Behind each of those accounts lie separate databases, which are connected to other databases, that hold dossiers of information on me. Some of that data is stale the minute it is entered in, some of it is utterly incorrect. Yet, they lie there dormant, until someone does something (maybe good, maybe bad) with it. These databases, as I’ve mentioned before, tend to make their way into the public sphere often, exposing their inaccuracies for the whole world to exploit. Let’s not even get into what happens when the companies that own these databases change owners, and the new management has different ideas on what to do with the data
This is admittedly a pessimistic view of the world. For most people, the small amounts of data they enter into an app is quite irrelevant, and the damages are quite minuscule even in the worst of all outcomes. Modern economies have ways to hedge these possible downsides like insurance. We are probably not pricing the risks correctly yet, but it’s definitely possible. Nevertheless, you simply can’t deny things are slowly getting out of hand, with more and more of our lives take place in the bits territory, instead of atoms.
I’ve written before that another way minimize these types of risks is to move to a more ephemeral model of data storage. The point I’ve made before wasn’t that we should never be holding on to any data but that we should be thinking of the entire lifecycle, including its disposal:
If every product manager in Silicon Valley thought about how their teams would eventually have to delete the data, we wouldn’t be in this mess in the first place. If right to erasure was part of the technical calculus, alongside maintenance and performance requirements done by tech leads, deletion would also work. If every engineer thought about the data she’s sending over the wire when they log an error message or send it through a PubSub system, she would be writing better code in the first place. The data wouldn’t seep into the machinery, like a viral infection that you can’t even diagnose, incubating for years and years, only to have a outbreak that almost destroys Western democracy.
Writing pieces toiling the long-term benefits of such a vision is fun, but I also try to practice what I preach. I, somewhat performatively, frequently delete all my tweets, in order to keep more of a fleeting presence on the platform.
It’s not particularly a novel idea, but it’s one becoming more common and even attracting investment capital. Just recently, the makers of the famed Sunrise calendar app came up with a new company called Jumbo. Their app is essentially a productized version of what I do with a mish-mash of Ruby scripts to delete my tweets and likes.
Platforms such as Facebook and Twitter both provide tools on paper, but in reality they are barely usable. Zuckerberg’s promised “Clear History” functionality is still nowhere to be seen. Twitter only allows deleting your last 3200 tweets programmatically. The aforementioned deletion wizard Jumbo seems to rely on a liberal read of the platforms Terms of Service agreements, and brittle hacks to impersonate user behavior.
The larger insight behind apps like Jumbo is that users only own their data only to the extent they can manipulate it as they wish, including deleting it altogether. This notion of ownership that’s predicated on operability is much more comprehensive and reflective of how people think of owning a good, then the narrow legal sense tech companies espouse.
This is where identity management and data ownership tie back together. One way to think of your identity online is as a combination of all the data that’s spread around behind hundreds of different accounts. Ephemeral data makes each of those individual accounts both less risky, and also more reflective of things work in the real world, with timeliness as a natural part. This is the part Jumbo attacks.
And identity management approaches the other variable, all the different logins and accounts on all the services. This is where companies like 1Password and LastPass operate.
I see these two approaches as attacking the problem from two different angles. The enterprise side of identity and access management has already made huge stride. Until very recently, the demand on the consumer side hasn’t been high, but clearly things are different now.
It remains to be seen how the future trends, along with aggressive regulatory moves like Europe’s GDPR or California’s best imitation of it will change the landscape. However, to me, it feels like we are on the cusps of big changes of how we are thinking of identities online, and how technology will let us manage presence online better.
What I’m Reading
Definite Optimism as Human Capital: Dan’a blog is one of my new favorites and he just became a Bloomberg writer too. This piece about how optimism is a hard to renew resource made me question my own skepticism and cynicism often, and is one that I keep myself going back to often.
It’s straightforward to measure a recession’s effects on employment and output. But what if the psychological impact of a recession is much more severe than we thought, to the extent that it could make a dent in long-term productivity growth? If we accept the idea that recessions linger in the form of psychological scars, lower expectations, and greater risk aversion, then it makes more sense to do a lot to avoid them. And it weakens the Austrian case for recessions as healthy corrections that improve capital allocation, because they cause a great deal of unseen harm as well. If we treated definite optimism as a function in human capital and productivity growth, then we could be slightly more rigorous in considering the broader effects of recessions.
How Game Theory Helped Improve New York’s High School Application Process: I have an odd fascination with the admission-industrial complex, especially in highly selective sectors. This 2014 piece is more about the former, with a mathematical tinge, and is fascinating.
Before the redesign, the application process was a mess. Or, as an economist might say, it was an example of a congested market. Each student submitted a wish list of five schools. Some of them would be matched with one of their choices, and thousands — usually the higher-performing ones — would be matched with more than one school, giving them the luxury of choosing. Nearly half of the city’s eighth graders — many of them lower-performing students from poor families — got no match at all. That some received surplus offers while others got none illustrated the market’s fundamental inefficiency.