Data is the New Guano

0:00:07.9 Paul Matzko: Have you ever heard someone say that…

0:00:10.3 Speaker: Data… Data… Data is the new oil.

0:00:13.8 Paul Matzko: Data is the new oil.

0:00:15.5 Speaker: Our data is now the most valuable asset in the world, and we are none the wiser, we’re not seeing a dime of it.

0:00:22.6 Paul Matzko: That’s not totally accurate. What they’re trying to say is just that data is valuable, like oil was valuable, and that corporations are interested in profiting from it, but that applies to just about everything in our modern economy. Data is not the new oil, it’s sh*t.

0:00:42.8 Speakers: Huh?

0:00:44.5 Paul Matzko: Guano, that is. Now, they don’t usually cover this in the standard high school or college US history class, but there was a moment in the 19th century when guano suddenly became one of the most valuable substances by weight in the world. It set off a global gold rush. Well, not really a gold rush, a sh*t rush, a race to get as much of that rich avian effluvient as we could. We needed that bird sh*t to save our agricultural economy, but our desperation for that sweet, sweet guano meant we didn’t think twice about creating the first overseas outposts of what would become a globe-​spanning American empire. Does that sound familiar yet? Think of how similar the story of bird guano is to the rise of digital data today. You see, the internet has turned something with previously little value, the aggregated information of the masses, into an invaluable resource that propels economic growth, fuels further globalization and is transforming modern society.

0:01:57.4 Paul Matzko: Data is the new bird sh*t. But data is even better than guano, or than oil, for that matter. That’s because it is functionally inexhaustible, and rather than just affecting agriculture or transportation, data is creating new efficiencies and propelling economic expansion in every aspect of the global economy.

0:02:23.2 Paul Matzko: I’m Paul Matzko. This is Building Tomorrow, and I want more data.

0:02:29.2 Speaker: I want some more.

0:02:29.3 Speaker: More.

[foreign language]

0:02:39.4 Speaker: More.

0:02:54.8 Paul Matzko: Nobody was more obsessed with poop than the 19th century American farmer.

0:03:00.1 Paul Matzko: Bird crap was the key, the mitigating, what historians call the agricultural fertilizer crisis. You see, when you’re a farmer, soil exhaustion is no bueno. Plants don’t just need sunlight and water to grow, they need nitrogen from the soil, nitrates to be specific. Let’s say you’re a farmer by the name of oh, I don’t know, McDonald, you’re Jerry McDonald, you stayed home on the family potato farm when your more famous brother Ronald went off to clown school, and then clogged the arteries of American children. How do you put the nitrogen your crops suck out of the soil back into the soil?

0:03:45.0 Daniel Immerwahr: Here’s the way it goes. If you are doing the sort of traditional old MacDonald had a farm type of agriculture, you plant things and you eat them, you consume them largely locally or on your farm, and then you poop and then the animals eat things and they poop.

0:04:05.1 Paul Matzko: That’s Daniel Immerwahr, a historian who has written about the guano craze in his award-​winning book, How To Hide An Empire. Okay, pooping, that’s great. The old McDonalds are great poopers. Really just superlative sh*tters, but how exactly does that help?

0:04:24.3 Daniel Immerwahr: And basically, there’s a kind of recycling in the soil, and why that recycling matters is that one thing that plants need to grow is they need nutrients, and most notoriously they need nitrates to grow, and nitrates… Like there’s kind of… Up until the 20th century, there’s sort of like a fixed number of nitrates, it’s hard to make new nitrates, but despite the weird fact that the air is like 80% nitrogen by volume, it’s actually really hard to break that triple-​bonded nitrogen away from itself and form the compounds that are plant edible, which are nitrates. So basically all of farming is like an extended experiment in managing nitrogen flows.

0:05:05.0 Paul Matzko: In pre-​modern agriculture, that nitrogen flow was pretty manageable. Subsistence farmers essentially cycle their own poop and that of their livestock through the fields to keep them stocked with nitrates, but when agriculture scaled up, as we started to shift from a society that was 90 plus percent farmers to the current 1.3% of Americans who live on farms, that meant the old nitrogen management system was suddenly insufficient.

0:05:34.8 Daniel Immerwahr: And you just want to grow tobacco again and again and again, and just squeeze as much tobacco out of the soil as you can, and you’re not consuming it locally, you’re sending it to a distant market. That is an absolute disaster, because basically what you’re doing is you’re sucking the nitrates out of the soil and then you’re delivering them in plant form somewhere else. And if you’re growing grain and you’re just doing fields of grain over and over and over again, you’re basically sucking them out of the soil, you’re giving them someone else, someone else gets the grain, the nitrate is in their poop and they poop it out, and if they’re in New York, they flush it into the ocean.

0:06:08.2 Paul Matzko: And it’s gone.

0:06:09.6 Daniel Immerwahr: And those nitrates are not going back to the soil. So people in the 19th century as industrial agriculture starts to come online, this is like its Achilles heel, and people in the 19th century start freaking out about it and they recognize the problem, they… Basically, they just try to come up with ways of fecal repatriation. Could you take the poop from the cities and bring it back and then put it back on the fields, like that would solve the problem? It gets wild, and the number of quite serious intellectuals who have opinions about this that are gross, gross opinions about this, is large, and we solve this problem ’cause we have chemical fertilizer. That’s the thing they don’t have.

0:06:48.1 Paul Matzko: But the problem is that chemical fertilizers derived from fossil fuel by-​products aren’t available at sufficient scale until the 20th century, but right now it’s the 1850s and crop yields on the McDonald potato farm are starting to fall as the nitrogen in the soil is depleted, and no matter how many “Ee-​I-​Ee-​I-​Ohs” you give, you can’t rely on the old methods to replenish your farm. What do you do?

[music]

0:07:24.5 Paul Matzko: Your salvation comes from an unlikely source. Off the coast of South America, there are a number of small islands, just dots really in the vast Pacific Ocean, but if you were one of a group of oceanic fishing birds, say a cormorant or a Peruvian booby, these islands were avian paradise. They were located right in the middle of rich anchovy and sardine shoals, a perfect source of nutrition for young hatchlings, which could be raised on these dry, isolated islands that provided natural shelter from predators and the elements, and these birds pooped. My goodness, did they poop. They fished, ate and sh*t. It was sheer bird heaven. Multiply millions of birds sh*tting day in and day out for hundreds of years, and by the mid-​19th century, these islands had grown from just small rocks peeking above the waves into literal mountains of sh*t, and that sh*t might as well have been gold.

0:08:29.1 Paul Matzko: Because guano is a fantastically rich source of nitrates. Fish are nitrate-​rich. The bird’s digestive system acts like a refinery, distilling the fish down to a purer form, which meant that there were millions of pounds of nitrogen-​rich fertilizer sitting there unclaimed in the ocean. Americans weren’t even close to the first to realize that this guano had potential. Centuries before, the Incan empire had relied on guano mining on the Peruvian coast to sustain its agriculture in terraced fields throughout the Andes Mountains, like those you see if you visit Machu Picchu, but the Peruvians didn’t have the boat technology or demand to mine the Guano Islands off the coast.

0:09:15.7 Paul Matzko: And so when US farmers started to feel the soil exhaustion crunch in the 1850s, the reports from whaling boats and explorers of these entire islands covered in natural fertilizer kicked off a guano rush, a quite literally sh*tty scramble for the resource. Entire fortunes were made and lost in the guano rush. When Congress was debating some kind of government action to secure the Guano Islands, a congressman claimed, perhaps with a touch of hyperbole, that a single Peruvian island was worth more than the Gadsden Purchase, Cuba and the entire Caribbean combined. So Congress passes the Guano Islands Act in 1856, which leads to more than 100 islands all across the Pacific becoming US possessions. Indeed, the Guano Islands Act is the first overseas expansion of American empire, which to this point had been limited to westward expansion across North America. American empire is quite literally founded on bird sh*t.

0:10:23.8 Paul Matzko: Isn’t that astonishing? Certainly, nobody involved in this story, not the ship captains claiming the Guano Islands, not the farm owners desperate for the guano, not even the poor souls mining the guano, none of them could grasp the full implications of the guano rush and how it reshaped their country and the entire world. But the guano mines were exhausted a few decades later, and eventually chemical fertilizers were developed that surpassed the old guano fertilizer, both in cost of acquisition and in nitrogen-​replacing efficiency. Today, the US has dropped its claims to most of those islands, leaving them once more the sole possession of the cormorants, boobies and their offspring. Who knows? Perhaps in a few hundred years, the islands will become mountains once again.

0:11:23.9 Paul Matzko: But the century after the guano rush, America faced another kind of productivity crisis.

[music]

0:11:32.9 Paul Matzko: In the 1970s, our economy ran out of steam once again. If you’re a certain age, you’ll remember stagflation, gasoline shortages and, worst of all, ugh, disco.

[music]

0:11:50.9 Paul Matzko: Economists call this period in American economic history the Great Stagnation, a time of much slower productivity growth, stagnant wages, and spiraling wealth and inequality. In part, that’s because the new energy tech of the mid-​20th century, nuclear power, didn’t pan out for a variety of reasons. The atomic age never fully got off the ground, leaving America dependent on the same old fossil fuel energy inputs, despite the growing awareness of their environmental costs. Yet the economic doldrums of the 1970s didn’t last for half a century. Even without new energy inputs, growth continued, albeit at a slower pace. Why? Well, it’s because we found new ways to eke out productivity gains without new energy inputs.

[music]

0:12:42.4 Paul Matzko: The rise of the internet boosted productivity, not via the old mechanism, unlocking new denser energy inputs, but by allowing us to do more with less, to use the same energy inputs, but get a greater return from them. You see, the internet is a kind of sustainability tech. By weaving together people from all over the globe into a worldwide web, it makes it possible to identify and exploit inefficiencies in a million unforeseen ways. That allows our economy to be more efficient, to do more with the same material goods.

0:13:20.7 Paul Matzko: For instance, it’s transformed the workplace. You might look with dread at your burgeoning email inbox, and yes, we all hate you inbox zero people, you know who you are, but imagine the bad old days when you had to do all of your business communication in the pre-​digital era, writing or typing out a letter, mailing it, waiting for it to arrive and then waiting for the response. Highly inefficient. But today, we think nothing of shooting an email or a Slack message or a text to a co-​worker somewhere else around the globe, and getting a response in minutes. Talk about efficiency.

0:14:00.3 Paul Matzko: Now, if guano as fertilizer was the key to maintaining America’s agriculture-​based economy in the 19th century, then data is the new life blood of the digital economy in the 21st century. It’s data that extracts those efficiencies and propels our economy forward, but the similarities between guano and data go deeper than the fact that both are really important to their respective economies. First, both of them are previously valueless things that quite suddenly became incredibly valuable. That rapid re-​evaluation has made fortunes, toppled governments and transformed global society.

0:14:44.3 Paul Matzko: Can you think of anything that’s more valueless than bird sh*t? Guano was a thing that had almost no value on its own, sitting there on that island, but it suddenly became valuable when extracted from the island and placed in a different context, on the farm in a farmer’s field. The same is true of data. Your data is financially worthless on its own. “What? I mean, Facebook, Google, they’re worth billions of dollars. How can my data be so worthless?” Well, it’s because data only gains value when it is collected and moved to a different context, when it’s aggregated by a digital platform.

0:15:26.0 Paul Matzko: Here’s a simple illustration. My favorite ice cream is Breyer’s mint chocolate chip. It’s so much better than every other kind of ice cream that I sometimes wonder what’s wrong with folks who order Rocky Road or, ugh, vanilla. So vanilla of them. Now, what’s the material value of the fact that I like Breyer’s mint chocolate chip ice cream? In a pre-​digital world, that fact is too individualized and locked away for advertisers to target ads to me. Instead, Breyer’s might hire someone from a Mad Man-​esque ad agency to create a campaign promoting the flavor on billboards, then newspapers, to a broad audience, but that’s a lot of wasted advertising on all those poor benighted folks who prefer other flavors.

0:16:17.6 Paul Matzko: So this data point, this factoid, that I like Breyer’s is nearly worthless in the pre-​digital era, but in a digital world, my revealed preferences, from what I search for, the groups I join, what I post and so on, can be more narrowly identified, combined with others like me and then used to target ads. This is a win-​win for everyone involved. I get more ads and coupons and the like for ice cream flavors I prefer, Breyer’s probably ends up selling more ice cream with less food waste as a result, and the ad revenue goes to more efficient online markets and advertising. Thus in the digital era, my data, generated by a combination of me and my preferences, plus an online platform, now has monetary value, but there’s an important distinction to be made before we go any further.

0:17:12.3 Dirk Auer: So I think it’s useful to draw a distinction between data and information.

0:17:21.6 Paul Matzko: That’s Dirk Auer, an economist at the International Center for Law and Economics, who specializes in tech policy.

0:17:29.9 Dirk Auer: So data would be specific tiny pieces of information that when put together can enable an economic agent to draw valuable insights. So in practice, what that means depends very much on the sector. In the tech sphere, data would typically be, I interact with a digital platform, it could be Facebook, it could be Google, and in doing so, I generate large amounts of data. This could be the search queries I looked for, metadata, where I was searching, at what time, and tech platforms are going to be able to put those pieces of data together and usually to draw insights about who I am. So for example, they may be able to figure out that Dirk Auer is a 30-​something-​year-​old male who’s interested in barbecuing, and so for them, they’ll be able to target ads to me based on that profile.

0:18:29.3 Paul Matzko: To put that another way, all information is made up of data, but not all data is information. Raw data is just a collection of factoids. Let’s go back to my ice cream example. Here’s a collection of facts. I like the flavor of mint, I like ice cream, my favorite ice cream brand is Breyer’s. That’s data, but information, on the other hand, is how you contextually make sense of that information. If I gave you those three data points, but didn’t tell you specifically that my favorite flavor was Breyer’s mint chocolate chip ice cream, you could probably still figure that out. In other words, using the power of your mind, you turned isolated bits of data into useful information.

0:19:17.2 Dirk Auer: Information requires more than just data. Information requires data associated with human intellect, so the ability to make sense of those pieces of data. So for example, in the tech platform example, it’s one thing to have all pieces of information saying this IP address searched for that query, and maybe made a search from that location at that time. It takes something more to then infer the general characteristics of that person, and then make that sort of the marketable information that you can then sell to advertisers.

0:20:00.0 Paul Matzko: Now, realize this isn’t all that different from guano. Bird sh*t has no inherent value in its existing context on an island in the middle of the ocean, it just sits there. But if you change its context, if through human ingenuity and effort, you move it from point A, the island, to point B, a farmer’s field, it now has immense value, and not to be too obvious, that required, shall we say, effort, both by the birds who sh*tted it and the company that extracted and shipped it.

0:20:36.5 Paul Matzko: In our digital world, we generate data as automatically as a bird generates sh*t on a guano island, but it’s not until that data or guano is efficiently gathered, refined and distributed that it has value. Our data has no value in and of itself, no real value apart from being turned into information, and that process is a collaborative effort. A platform like Facebook or Google or whomever takes worthless bits of my data, algorithmically analyzes it, and turns it into information with value. You have to understand this point to get just how idiotic it sounds when various activists call for laws or rules that would force companies to “give you back your valuable data,” as if you originally possessed something with inherent financial value that was then taken from you by a social media giant and evilly sold for profits.

[music]

[laughter]

0:21:36.7 Paul Matzko: But that attitude betrays a very basic ignorance about both the difference between data and information and how online data markets work, and the danger is that these sometimes well-​intentioned souls could pass laws or rules that destroy or hinder those data markets, turning valuable information back into valueless data and making all of our lives worse off as a result. Why is that? Well, here’s the second point you need to grasp. Our data, and I’ll just use that term instead of information for sake of convenience, it has value because we found a way to drastically lower the transaction costs to create it. That’s a bunch of fancy $5 words. Transaction costs is economists’ speak for the cost of doing business. Let’s say you want to buy a Christmas present for your kid, who demands an air rifle, and not just any air rifle.

0:22:32.8 S?: I want a Fisher Red Ryder carbine action, two-​hundred shot range model air rifle.

0:22:38.4 Paul Matzko: You know the one. Buying that rifle doesn’t just mean paying for the raw materials that went into the rifle or the cost of manufacturing it. You also have to pay for the expense of the maker shipping it to thousands of stores around the country, the cost added on by those retailer middlemen, the cost of the credit card fees for the purchase, the cost of the gas you spent driving to the store, and so on. These are transaction costs above and beyond production costs, the cost of exchanging that item between owners. Of course, you can lower those transaction costs by, say, cutting out the retail middleman, selling directly to the consumer, or by developing drone delivery technology that decreases pick-​up costs, or in a hundred different ways you can make the transaction more frictionless, which makes the good cheaper and thus increases the value generated by that good as more people are able to afford it.

0:23:33.4 Paul Matzko: This too happened with guano once upon a time. Although it was primarily a demand-​side shock rather than a supply-​side shock, the development of better boat technology, the excess cargo capacity of old whaling vessels as the whaling industry declined in the 1850s, all of that reduced transaction costs for the delivery of guano to American farms. So to today, global digital interconnectedness has drastically reduced the transaction costs for data collection and use. In the pre-​digital era, how would one collect people’s data and organize and aggregate it into useful information? I should note, it wasn’t entirely impossible, it was just really, really inefficient.

0:24:18.8 Paul Matzko: Here’s one common use of personal information captured from pre-​digital data, the Rolodex. If you’re of a certain age, you remember a time when any business person worth their salt had a big circular device on their desk containing the cards and information for all their business contacts. This was hugely important. They quite literally taught classes in business schools about building out your contacts and getting a nice big Rolodex. Fast forward to today, and digitization has lowered all kinds of information-​sharing transaction costs.

0:25:01.8 Paul Matzko: Now, we all have a kind of Rolodex via our social media accounts, no business school classes required. Furthermore, because the transaction costs for data are so low, there are a thousand novel ways that have evolved to use data in cool, salutary ways. It’s not just advertising. That’s the lowest hanging fruit. Consider one use of our data that has transformed our daily lives for the better that’s become so routine, we now take it for granted: Route mapping.

0:25:34.1 Paul Matzko: I’m old enough that I remember what it was like to prepare for a long drive in the pre-​online era. You would pull out your Rand McNally atlas, not a sponsor, and look up the place you were going and figure out which interstate ran closest to it, but if you got stuck in a traffic jam, you never had any idea how long you’d be stuck there. Could be 15 minutes, could be four hours. So you hauled out the atlas and then agonized over whether you should try to get to the next exit and take the state highways, or whether that would get you lost, and mom and dad would be arguing in the front about what they should do, and inevitably you made the wrong choice.

0:26:13.0 Paul Matzko: It was such a commonplace experience that we based entire movies and TV show episodes around having to rely on directions from random unreliable gas station personnel, you know, the type who probably just wanted to route you to the nearest haunted house or mountain cannibal tribe. You heard it here, folks. Route mapping saves lives. But today, you just open up your Google or Apple Maps app and let it plot a course for you. I can’t remember the last time I got legitimately lost on a trip. You’re stuck in traffic, the app will tell you how far ahead the accident is and re-​route you.

0:26:52.3 Paul Matzko: It can do that because a company like Google not only digitized all the maps, but because it calculates from the data created by millions of fellow mappers just like you how long it normally takes to get from point A to point B. It even actively encourages drivers to share information about police speed traps and other obstacles. And as a result, we all spend less time getting to where we’re going, and with less stress in doing so. It also means knock-​on benefits, like less gas consumption, fewer carbon emissions released into the atmosphere, fewer encounters with mountain cannibal tribes and so on.

0:27:31.5 Paul Matzko: So something that was previously locked away and nearly valueless, my individual trip planning and route data, has now been aggregated with the trips of millions of others, algorithmically analyzed and turned into an incredibly valuable application that invisibly improves our daily lives. And for every example like that, there are ways that digitally lowered transaction costs will continue to unlock our worthless data and create new forms of valuable information that we can’t even fathom at the moment.

0:28:06.8 Paul Matzko: So, we’ve lowered transaction costs and unlocked the incredible potential value of data. But a question remains: Who should own that data? Should I own it? Should online platforms own it? How do we decide?

0:28:21.6 Paul Matzko: Well, the first step is remembering that this data doesn’t exist apart from a collaboration between each of us and the online platforms we interact with. Pulling out of those systems doesn’t give you back your data, it just destroys the ability to create useful information from that data. This is fundamentally different from other forms of possession. If you let someone borrow your car, when you decide to take it back, the car still exists, it still has value. If you rent out a room in your house and you decide to stop doing so, the room still exists, it still has value. But with digital data, its value is almost completely collaborative. Thus the goal ought to be to ease the creation of data that adds value to our lives, without making people feel compelled or short-​changed by that process. The creation of data value is good for us, good for online platforms, good for the economy.

0:29:21.0 Paul Matzko: We want to do more of this, not less. Some people have proposed that the best way to ensure fairness in the digital marketplace is to formalize property rights, to pass legislation that would give consumers some kind of absolute right over their data. And that sounds reasonable, even libertarian, given the importance of property rights to economic growth. But formalizing property rights could actually hinder the creation of this new form of property. You see, formalizing the property right often means getting rid of a prior informal system of property that actually works even better, and replacing that informal property right with a formal property right can destroy the value of that data itself.

0:30:08.0 Paul Matzko: Let me give you a concrete example of those thickets that Dirk is talking about. I once lived on a street called Bentonville Way. It was a fairly busy road in a residential neighborhood, which was great, because it meant that lots of folks drove by, including several junk collectors. I call them junk collectors, because you could put out your junk, and not trash, junk, like broken dishwashers and old furniture. And it would be gone within hours. The junk men would take it, they’d fix some of it, re-​sell it or scrap it for raw materials. Think about what’s going on there. I had an object that was valueless to me, which otherwise I would have had to take to the dump, but it had value to the junk men, but only so long as the transaction costs were negligible, which they were.

0:31:00.8 Paul Matzko: The junk was free, right there on the side of the road. And the exchange of property between us was informal. Both sides knew that stuff on the public right-​of-​way was fair game, something that technically still belonged to me, but which I was willing to give to them. And this system worked just fine. This is what data is like, it’s trash that becomes junk in combination with another party. One man’s trash, everyone else’s trash, and a little algorithmic pixie dust can become junk or even treasure.

0:31:35.2 Dirk Auer: I think of data, not as much as the new oil, but sort of the new trash, so it’s something that you… It’s a by-​product of your online activity that you never cared about or thought about, but that someone else can actually make money from.

0:31:50.3 Paul Matzko: But now imagine if my busybody neighbor, let’s call her Karen, as a concerned, progressive, upstanding good citizen decided that this system was unfair, that this junk clearly had some value to me that I wasn’t getting back, that was accruing instead to the benefit of the junk men. So my neighbor, oh, so graciously lobbies the Town Council to pass a law formalizing my property right over that junk. Now, in order for the junk men to take possession, they need to give me some percentage of their proceeds from any future use, they need to sign a contract with me in which I agree to the transaction, they need to file that contract away so that regulators and courts can make sure that the exchange has been fair, and the junk men have to do that for every piece of junk they pick up at every house they visit. What would that actually accomplish?

0:32:46.7 Paul Matzko: Well, it would boost the transaction costs of those simple exchanges exponentially. Given that the value of my little informal Bentonville Way junk recycling program was not large, it would effectively kill it. That would hurt the junk collectors, of course, and it would not only fail to return any value to me, it would make my life demonstrably harder. Now I have to find a way to get the stuff to the dump myself. In other words, it would turn junk into trash. Sometimes informal evolved property rights are superior to formal imposed property rights, and that is what formalizing property rights in data could do to the digital economy. Rather than returning real value to consumers, it could turn valuable information back into valueless data.

0:33:44.7 Paul Matzko: It would destroy the value, not transfer it back. To put this in concrete terms, imagine that if every time an online platform aggregated our data, it had to get my express permission. Let’s say they were combining my preference for Breyer’s mint chocolate chip ice cream with the thousands of others who share my predilection for the sweet minty frozen ambrosia of the gods and selling access to that anonymized market segment to advertisers. Now, imagine if every single time they wanted to do that under our hypothetical situation, they would need to ask my permission. It’s my data. That platform would have to get my affirmative consent for that transaction, via, I don’t know, a pop-​up or an email or a text, or the company couldn’t share it.

0:34:37.5 Paul Matzko: Now, imagine them having to do that for every individual user, dozens or even hundreds of times per day; imagine being that user getting deluged with those requests, it’s insanity. Obviously, the transaction costs would skyrocket and that wouldn’t result in me getting my data back, no, it would just kill the data sharing market. Are you happy now, Karen?

0:35:04.2 Paul Matzko: Now, if that sounds too fantastical, consider what hath HIPAA wrought. HIPAA stands for the Health Insurance Portability and Accountability Act. You might have heard recently Congresswoman Marjorie Taylor Greene citing HIPAA as a reason for why she didn’t have to answer a journalist’s question about her vaccination status. But don’t listen to Marje, HIPAA has nothing to do with that. It’s a law passed in 1996 meant to make it easier for doctors and insurance companies to share health data while still protecting patient privacy, but it also formalized a kind of property right over your health data.

0:35:43.4 Paul Matzko: What did HIPAA actually do? Well, it required affirmative consent from patients for every data transfer between your doctor and any third party. It’s why when you send medical records, you have to sign a piece of paper with a bunch of legalese on it. So it formalized a property right in a particular kind of data, data about your health, but rather than encouraging sharing, it really just jacked up transaction costs. Any violation of the law can result in significant legal liability for providers, but even more onerous are regulatory compliance costs. Remember, your data is supposed to go point to point with no unauthorized third parties in the middle.

0:36:26.5 Paul Matzko: Now, think about that for a second. What’s the problem with that conception of data as a thing that simply goes from point A to point B? The internet doesn’t work like that. It’s a worldwide web. When you send an email, it bounces between a host of servers, mail servers, DNS servers, all run by multiple companies in dozens of countries around the globe. Email is almost invariably in violation of HIPAA, which is why your doctor’s office probably sends you messages via a proprietary message system with an interface that looks like something from the ’90s. Gee, thanks, HIPAA. And that has harmed the collective health of Americans.

0:37:09.2 Paul Matzko: To give you just one concrete example, I was once prescribed a heartbeat monitor. I paid a very large amount of money for an FDA-​approved, doctor-​prescribed, HIPAA-​compliant heartbeat monitor. After wearing it for 48 hours, the results came back negative for heart issues and I went on my merry way, but the doctors only got a tiny slice of my heartbeat history. What if that was a particularly good two days for my heart? What if I had a murmur that surfaces less frequently than that, once a week or once a month? Well, I have a smartwatch device that can track my heartbeat day in, day out, and do so about as well as the device that my doctor gave me. That’s data that a doctor could turn into useful medical information, but I can’t easily share that data with him, because it would pose HIPAA compliance issues.

0:38:02.6 Paul Matzko: Every year, there are Americans with heart problems that will drop dead simply because HIPAA made it too hard to share heartbeat information. It prevents us from turning useless data into useful medical information, and this problem remains true for a host of potential health tracking applications. Our medical data system is a quarter of a century behind our social data system. Why? Precisely because we formalized a medical data property right. So maybe we should think twice before doing the same thing to our other forms of data.

0:38:48.6 Paul Matzko: Thus far, I’ve been discussing the ways that data is like guano, but I do want to touch on some of the ways it’s not like bird sh*t at all, and why that’s a great thing. First, and this is fairly obvious, extracting the bird sh*t was really sh*tty. The odor was incredible, there was no shade, mining companies basically kidnapped people, then forced them to work on the islands for months at a time and paid them almost nothing, as Daniel Immerwahr puts it.

0:39:18.1 Daniel Immerwahr: I asked 19th century historians, what is the worst job that you can have in the 19th century? And there are a lot of answers, but guano miner is absolutely up there. Okay, first of all, you’re a miner, so that is bad, right, I mean, like mining, it’s a really hard job. You’re breathing in all kinds of particulates, and that’s certainly the case in guano mining, this stuff stunk, it had this sort of ammonia smell. If you were around a ship hold full of guano, the accounts of this are… They had sailors down there with… In a guano ship hold, would just come out within minutes, with nose bleeds and temporary blindness, it is not good for you.

0:39:55.6 Daniel Immerwahr: The other part of it, though, is that to be a guano miner, you’re also on an island, which is a unique kind of hell, so you’re shipwrecked at the worst job.

0:40:04.7 Paul Matzko: Eeuch. But by contrast, while guano mining was terrible, the creation of data just means you playing Six Degrees of Kevin Bacon on Wikipedia, watching cat videos on TikTok or writing a tweet complaining about whatever the latest viral meme that boomers are circulating on Facebook. A little bit different experience, wouldn’t you say? Indeed, data’s revaluation has had the greatest impact in precisely the places where even small boosts in value have outsized impacts. For example, user data is being used to bank the unbanked in Turkey, to help farmers in India get a better read on seasonal crop demand and so on.

0:40:51.0 Paul Matzko: In fact, there is a project to create a cell phone app for Kenyan goat herders that collates location data, satellite images of vegetation and information about local grazing conditions to help the herders maintain a sustainable herd size and prevent overgrazing. Given that 60% of Kenya’s economy is livestock-​based, this is potentially a big freaking deal, and it’s brought to you by data. We’ve come full circle to agriculture again, guano boosted productivity by renewing the soil, data is boosting productivity by encouraging more efficient and sustainable use of what’s already there in the soil.

0:41:33.3 Paul Matzko: So compared to guano, data is a much easier resource to extract and has a far wider set of beneficiaries, but perhaps the most significant difference is that guano is an exhaustible resource, while data is the ultimate renewable resource. Eventually, the guano was all used up, and the guano rush was over, but data is a functionally infinite resource. Every one of the 7.8 billion people in the world produce it as a matter of fact, and as more and more people go online, more and more of their data is turned into information and made more and more valuable. It will never be exhausted.

0:42:16.5 Paul Matzko: A data-​based economy is here to stay. I for one am happy about that fact. If you’re not, well, be sure to pick up your smartphone and register your disapproval on the social media platform of your choice, so that the algorithms can serve you up some ads for techno-​skeptical YouTube videos. But what’s next? Is there another resource on the horizon that could do what data and guano did, suddenly go from valueless to valuable? You’ve probably heard science fiction author Arthur C. Clarke’s quote that “any sufficiently advanced technology is indistinguishable from magic.”

0:42:56.0 Paul Matzko: Well, Noah Smith, columnist, economist and Twitterer extraordinaire, says that we should get ready for some hocus pocus.

0:43:07.1 Noah Smith: I would say that if there’s a technology that’s similar to the classic definition of magic, which is that we can do things, but we don’t understand how we just did them, the principles behind what we just did, when we can systematically do them, then that’s AI. I would say AI is, depending on whether… How far we get with explainability of AI, AI might actually be true magic, which is to say, spells that we can cast, algorithms that we can sic on problems that have insanely good predictive power, but whose sort of… But we never understand the difference between the algorithms that do well at prediction, the algorithms that do poorly at a kind of structural sense, like why is this getting it right? If we never understand it, then AI actually is magic. Literally, we’re casting spells.

0:43:55.0 Paul Matzko: You see, when a technological innovation comes along, we briefly marvel, then we make it a routine part of our lives, and then we take it for granted. The magical becomes mundane. But with all of these technologies, it’s a bit like we’re alchemists who finally unlocked the secret of the philosopher’s stone. We turned bird sh*t into crops, we propel our vehicles at speeds that would have boggled the minds of our ancestors by exploding dead dinosaur residue in metal boxes beneath our feet. And now we’re starting to turn useless data into valuable information that will create artificial intelligence. Heck, we’ve already used simple machine learning algorithms to help us predict the weather with ever-​increasing accuracy. That might have once got you burned for witchcraft, and more complex artificial intelligence is going to transform all of our lives in ways that we can now at best just glimpse. I’m Paul Matzko, this is Building Tomorrow, and I want the next bird sh*t.

[music]

0:45:28.2 Paul Matzko: In a fitting end to an episode about valueless data becoming useful information, let me give credit to those without whom this episode wouldn’t have been worth much more than a pile of bird crap. Thanks to our guests, historian Daniel Immerwahr. Go check out his book, How to Hide an Empire. Economist Noah Smith, who has an excellent newsletter at Noahpinion@​substack.​com, and the economist Dirk Auer, who is on Twitter @AuerDirk. Our producer is Landry Ayres with the support of Greta Langhenry, Natalie Dowzicky and Aaron Powell. If you want more podcasts and content like this, visit lib​er​tar​i​an​ism​.org.