I want to talk about four things. They're all to do with counting. One of them is money, one of them is health, one of them is power.
And one of them is how we can do all of those things better. But throughout, I'm going to talk about this underrated thing that we do, which is to count stuff. And to begin this journey through the past, the present and the future of counting, I want to start by talking about a sad French woman. Not because she saw the football result last night, but she was sad for a different reason. So her name is Denise Schmand-Besarat.
She was born Denise Besarat in France in 1933. Obviously... The 1930s is an awkward time to be born in mainland Europe, but actually Denise had a pretty decent time during the war. She had a rich uncle and she was taken off during the war to some castle where he lived.
And she recalled, I have wonderful memories of that time, being a child with a castle. They had a donkey, they had ducks. What more could you wish for? So that's a good start.
And then after the war she got interested in languages and in the 1950s, in 1954, she met a German man, a young dashing German philosopher called Jürgen Schmand-Besserat. And she said it was not very fashionable in 1954 for a French woman to be engaged to a German. Nevertheless she persisted with that.
plan and eventually her parents came around to the idea and she became Denise Schmand Besarat. She married the love of her life. And by 1960 Jürgen had a good job. He was traveling the world, he was doing interesting things. Denise was living with him with their two healthy sons in a suburb of Paris, a beautiful garden, they had access to a swimming pool, they even had a maid.
And Denise said, instead of counting my blessings, I felt frustrated. What was wrong with me? What was wrong with her?
Why didn't she feel satisfied? She seemed to have this idyllic life. She seemed almost a charmed life. And yet she was unhappy. And one evening over a dinner party...
A friend of hers said, well, you know, whenever, when I felt really at a loss, when I was unsatisfied with my life and I really wanted to make a difficult decision, what I did was I went to the tomb at the heart of the Great Pyramid of Cheops and I sat there all by myself and thought it through. This is the kind of circles Denise mixed in. I sat by myself in this burial chamber and I thought it through until I had solved my dilemma.
Well Denise thought that sounds like a pretty good idea. However, the Great Pyramid was not available. So instead, one day after the children had been dropped off at nursery, she was all by herself, she drove to a nearby Japanese garden on the outskirts of Paris and She saw this beautiful bench overlooking a bridge and all around her the raked gravel, the morning sun was just beginning to warm this very special place and she said to herself I'm going to sit on this bench and I'm just going to think and I'm not going to move until I figure out what's wrong with me.
And she sat on the bench and it didn't take long. She realised... She had basically surrendered to this role of being a housewife.
She had this husband who had this incredible career, going all over the world, meeting all these interesting people. And she said, well, I want those things too. And you know what that means?
It means going back to school. And so she went back to school. She enrolled at the Louvre and studied archaeology, which didn't go very well first year because she was pregnant again, so she kept falling asleep in the lectures. Don't get any ideas, any of you, OK? Stay alert.
So she went back again the next year, not pregnant, and this time she really discovered this love for archaeology. And by the late 1960s she had a... position as a researcher, she was affiliated with Harvard University, she was actually making a go of being an archaeologist.
And she went, I interviewed her not long ago and she said it was a very brave thing for me. I went all by myself, this is 1971, I went all by myself all around the Middle East. I went to Syria, I went to Iraq, I went to Iran, all of these different places. It was pretty scary but as a woman all alone in these places but I just...
needed to explore all of these digs, all of these sites, where all these archaeologists were finding these incredible things, these amazing discoveries. I wanted to do that too. And at the time archaeologists were fascinated by all the cool things that the Mesopotamians were doing. These great cities, their structures of governance, the taxation, agriculture, politics, that sort of thing.
You know, exciting, exciting stuff that archaeologists could study. But just to give you a sense, the Epic of Gilgamesh, one of the first great epics ever written, he built the town wall of Uruk, the city of sheepfolds. Look at its wall with its frieze like bronze. Gaze at its bastions, which none can equal. So, you know, gaze at its bastions.
That's the kind of thing that archaeologists were doing. But... Denise, she said to me, and maybe it's a woman thing, but I was just kind of more interested in in the everyday stuff, in the humble stuff, and she became fascinated by clay.
How did people use clay? What were the earliest uses of clay? And you think about what are the earliest uses of clay?
You might think to yourself bricks maybe or pottery, tiling, that sort of thing. But no, the very earliest uses of clay in Mesopotamia were these things. They look a bit like game pieces. And archaeologists discovered them decades ago. And Denise verified for herself, both by looking through the records and by going to these different sites, they were absolutely everywhere, all over the Middle East.
And they were the oldest layers of these archaeological digs. These were thousands of years old, 7, 8,000 years old, 9,000 years old, and everywhere. And the curious thing was, all the archaeologists she dealt with had absolutely no interest in them. Now what were they? You know, they're game pieces or something, you know, toys maybe?
And one archaeologist said, well they look a bit like suppositories to me. What they're actually used for is anyone's guess. Well that wasn't quite enough for Denise.
They're obviously important. They're everywhere. They're part of some system, they're in every site, they're very old. As she said, it's very surprising that the first use of clay is these tiny, tiny geometric shapes that are absolutely ubiquitous.
So what are they used for? Well, Denise started to look into counting. It turns out, she told me there are so many ways to count. She's giddy with enthusiasm. There's so many ways to count.
And by so many ways to count, I don't just mean, you know, one, two, three, or, you know, eins, zwei, drei, or ich, ni, san. I mean just different ways of approaching counting altogether. So, for example, in some languages... Generally not modern languages, modern cultures, but in some more traditional, older cultures, counting specifies the particular kind of thing you're counting.
We have a little bit of that in English. So, for example, triplets means three of something, i.e. babies born on the same day. A trio refers to three musicians.
So we have these hints, but generally... You just go, well, it's three of something, and then I'll tell you whatever it is that the something is. But there are languages where three of one kind of thing, you use a different word for three of another kind of thing.
Or you can count without numbers altogether. So the Veda in Sri Lanka used to, I mean, this has all died out now, but they used to count by just saying, OK, let's say we want to count coconuts. I've got a pile of coconuts. How many coconuts are there? OK, well, what you do is you take your coconut.
Put it there, and then you take a pebble. You take your second coconut, another pebble. You take your third coconut, third pebble. You just keep going.
There are no numbers. How many coconuts have you got? That many.
And it seems a bit pointless, but it's not... Actually, clearly it's in many ways more convenient to say, well, there are eight coconuts. But the pebbles will do for a society that doesn't have a lot of need for counting.
And one thing you can check is, has somebody taken any of my coconuts? Have I still got the same number of pebbles as there are coconuts? Because if there's more pebbles than coconuts, then somebody has stolen my coconuts. Always trolling with me by adding pebbles. I mean, one or other of those things.
So Dhanush Shman Besarat says, look, these things, they're like the pebbles that the Vedas are using to count coconuts. But it's much more sophisticated than that. So there are different incisions, different marks, different patterns on these baubles, which are not baubles, they are tokens, signifying different numbers. So you can have something that signifies a single object or something that signifies five or ten or sixty. And the shape of the...
Object indicates what you're counting. So you might be counting jars of oil, or you might be counting jars of wheat, or you might be counting sheep. So this is a very early system of accountancy.
And it's been completely overlooked for decades by archaeologists who were interested in the great sheepfolds of Uruk and its friezes of bronze and all that kind of thing, interested in all this manly stuff. And along comes Denise Schman-Besarat and she says, look, these little things that you think are basically suppositories or whatever, you don't care what they are. They are fundamental to the Mesopotamian system of mathematics and of economics because these are economic commodities.
So the whole system is somehow running on these. She went further. It turns out there's this big puzzle going on in ancient history, trying to understand what's going on in Mesopotamia. And the big puzzle is these things. So these are these are tablets written in cuneiform and they're very very old.
I'll tell you how old but I've forgotten. It's thousands of years old. Much older than other forms of writing that have been discovered.
And that's a bit of a problem because there was always a theory. So where did writing come from? Well, the theory was basically where you start by drawing a little picture.
So, you know, sheep, fine, you draw some sheep. And then writing becomes kind of more stylized and you move from pictograms to some symbol that looks a bit like a sheep. And then in the end, it doesn't really look like a sheep anymore.
And that's writing. That's abstract writing. But it's evolved from pictures. You can imagine hieroglyphs like the Egyptians used to use.
The trouble is. These cuneiform tablets, there's, I mean, they don't look like pictures and they've got no precursors. They're incredibly old and there's no, there's no pictures on tablets that are older than these. Well, maybe they didn't write them down on tablets.
Maybe they wrote them down on something that rotted away. I mean, maybe. It's not terribly compelling as a theory. Dhanush Besarat says, well, actually... If you look, you can find evidence of cuneiform writing that is the tokens themselves.
These things being pressed into clay tablets. And you start with maybe a clay ball. And you press each token onto the outside of the clay ball and then you put it in the ball.
And so you have a written record on the outside of the ball of what's inside the ball and you seal the ball up. To actually discover whether this is true or not, one of Schman-Besarat's teachers, Pierre Amier, basically gouged into a... A 3000 BC, 5000 year old clay ball.
He was just super curious about what was in there. And it turns out, no, it's the tokens. And you can see each token matches what's going on on the outside. But pretty quickly, you realize that's a big old fuss, having these clay balls with the tokens inside and the kind of impressions on the outside. You don't need to do that.
You just need to, you can just press the tokens into the clay and then you can put the tokens away. You can throw them away or you can store them for another use. And then after a while, why bother with the tokens?
You don't actually need the tokens, you just need a stylus, and the stylus can make these impressions. So this hasn't evolved from pictograms, this has evolved from impressions of tokens. So it turns out we're not just looking at the earliest mathematics, we're also looking at the earliest writing.
And we're also looking at one of the earliest instances of trying to organize a complex society. Remember the great city of Uruk with its sheepfolds and its freezers of bronze. These are among the most sophisticated societies anywhere on the planet at the time. And they're sophisticated because they have complex economies, they have hierarchies, they have some of the first cities, they have writing, the first writing, they have the first accountancy. This is a spreadsheet, and it all comes back to counting, to these little things that the archaeologists thought, ah, they're toys, they're baubles, they're suppositories, they're whatever, we don't care.
And along comes Denise Shmand-Besarat, and identifies how fundamental counting is to this entire project of civilisation. That's the first point I want to make this afternoon, is that we underrate counting. We overlook it. We think it doesn't matter.
We don't really care. We think it's, you know, for kids and for women. but actually it is fundamental to everything.
When I spoke to Denise, this is a picture of her about 10 years ago, she's still going, she's 90. Actually she's probably 91 now, she was 90 when I spoke to her. She was saying this is so important, this is, what did she say, this is a turning point in communication and data storage. Now we don't think of this as data storage, we don't think of this as data storage.
But it is, far below, before the computer age. This is civilization being formed, being built on foundations made of counting tokens. But when I spoke to Denise, she said, you know, the one thing, people talk to me about my work. The one thing no one's ever written, Tim.
No one has ever written, and she has worked hard. So, okay, I'm putting it on the record. I'll put it in my next book. She's worked hard. Okay?
This is important. This gives us a sense of just how important counting is. So here's to Denise Shaman Besarat. That's the first point I want to make. Let's change the story.
Let's talk about another ancient civilization, another ancient piece of writing. This is the oldest medical advice known to any civilization on the planet. It was discovered in some tomb in ancient Egypt. It was brought to the West by a collector called Edwin Smith.
So it's known as the Edwin Smith papyrus, which is a little bit problematic, but there you go. And Edwin Smith spent decades trying to translate it and failed. He lived in the 1800s. It was eventually translated in about 1930. So we now know...
What this medical advice consists of, we know what it says. Basically what it says is a whole bunch of things that will not make you better. It is incredibly bad medical advice.
There's something mentioning willow bark, which does chair a chemical compound with aspirin, although it's difficult to get enough to be therapeutic from willow bark. And apparently the advice on spinal injuries is okay. Apart from that, useless.
This is the first piece of medical advice we've got. It's three and a half thousand years old. Absolutely no help whatsoever.
But you... okay fine, so doctors have had three and a half thousand years to get better. So what have they been doing?
So let's see what Galen was writing two thousand years ago. Galen, one of the fathers of modern medicine, or of medicine anyway, he says, I've got this amazing new treatment. He says, in a short time, except though it's incredibly helpful, in a short time, except those whom it does not help, who all die.
However, in case you're thinking that doesn't sound like a good treatment, it says no, no, no, it is a foolproof treatment. It fails only in incurable cases. Okay, that's not very good.
That's not very good. So what are you going to do about that? Well...
We've learned that in order to evaluate medical treatment, what you do is you run a trial. So you give the medicine to some people and you don't give the medicine to other people and you see if the people you gave the medicine to get better compared to the people you didn't give the medicine to. It's a pretty simple idea.
It's astonishing how long this idea took to catch on. So in the 11th century, the Chinese had a go. They ran a trial of ginseng. They got two runners. each ran a mile.
One of them got ginseng, one of them didn't. The one without the ginseng developed severe shortness of breath. Well the one who took the ginseng breathed evenly and smoothly. Which is evidence in favour of ginseng?
It's not very good evidence though. But here's the thing, so this is the 11th century. They're so close. If you gave the ginseng to a hundred people and no ginseng to a hundred people and got all 200 people to run a mile You got it. Okay, you actually have nailed it But no, they somehow couldn't quite get around this idea.
It turns out to be very elusive. There was an Arabic scholar Called Abu Bakr al-Razi in the 10th century. He ran a proper clinical trial Unfortunately, he found that bloodletting cured meningitis and bloodletting doesn't cure meningitis. So something went wrong with Al-Raz's trial.
We don't have enough detail to know. Probably what went wrong is it wasn't randomized. So probably what went wrong is he looked at a bunch of people and said, OK, you know, you look like the kind of people who benefit from my bloodletting and you look like the kind of people who won't. And.
That's probably why the bloodletting appeared to cure the meningitis, even though bloodletting does not cure meningitis, but we don't know. Anyway, it wasn't properly conducted. And then, about 1650, if I remember rightly, Jan Baptist von Helmont had a real beef with the other physicians. He's like, you guys are rubbish, I'm great, and I can prove it. Let's draw lots.
We'll take 500 people from the lepers colonies and from the hospitals and half of them are going to get your treatment and half of them are going to get my treatment and we shall see how many funerals each shall have. So that is a proper randomized trial. The only trouble is he never actually did it.
So he was so close, he never did it. Then this is the man, James Lind. So, James Lind, famous for running the first ever clinical trial.
I mean, with the exception of all these other clinical trials I just told you about, which perhaps don't count. So James Lind, he's interested in the problem of scurvy. I'll just give you an account of scurvy from the Middle Ages. With violent pains in the feet and ankles, their gums became swollen, their teeth loose and useless, while their hips and shin bones first turned black and putrefied.
Finally, an easy and peaceful death, like a gentle sleep, put an end to their suffering. So this is the problem with Scurvy. And Lind is serving on HMS Salisbury in 1747. And he says, and Scurvy's a terrible problem for the British Navy and for anybody else who's trying to go on these long sea voyages.
And so Lind says, look, I'm going to sort this out. I'm going to figure this out. So here's what he has.
He writes in his diary, 20th of May, 1747. He finds 12 sailors who have scurvy. Their cases were as similar as I could have them. They all in general had putrid gums, the spots, and latitude, and weakness of the knees. They lay together in one place and had one diet common to all.
So to two of them, he gave elixir of vitriol, 25 guts three times a day. For those of you who don't speak. 18th century English that is 75 drops a day of sulfuric acid on an empty stomach Two of them two of them got Two spoons of vinegar three times a day two of them got half pint of seawater Two of them got a quart of cider which I mean it's probably gonna make you feel better.
It doesn't cure scurvy Two of them got a purgative mixture Not good. Two of them got oranges and lemons. And within about a week, the ones who got oranges and lemons had completely recovered. And the ship had run out of oranges and lemons, which is a shame. But this was James Lind finally using a clinical trial to demonstrate if you want to prevent scurvy, oranges and lemons will do it.
There is just one problem. So, in 1912, I read from a diary the description of a certain Mr. Teddy Evans. Mr. Evans is turning black and blue and several other colours as well.
Mr. Evans is gradually worse. It's no use closing our eyes to the fact. This morning we were forced to put Mr. Evans on his ski and strap him on as he could not lift his legs.
Mr. Evans is in a very bad state. If this is scurvy... I'm sorry for anyone it attacks. Uh, Teddy Evans. was one of the people accompanying Robert Falcon Scott on his expedition to the South Pole.
And part of the way on that expedition to reach the South Pole, some of the men were sent back. They'd been there in a supporting role. They were sent back. One of them was Teddy Evans.
He arrived back at base camp in the condition I've just described. In fact, he didn't actually make it to base camp. He collapsed about 30 miles away from base camp.
And one of his comrades... walked alone through Antarctica 30 miles to get help, bring back and get Teddy Evans who then survived. It is highly likely that all of Robert Falcon Scott's men, including Scott himself, suffered from scurvy. And when you read Scott's diaries, it's weird because they're kind of, they're just getting weaker and weaker and weaker and he doesn't really know why everybody is suddenly getting weaker.
Almost certainly scurvy because of a lack of vitamin C. But you're currently thinking to yourself, Tim, this is 1912 and Robert Falcon Scott was a Navy officer and wasn't James Lind in 1747 working for the Navy? Didn't he just prove how to prevent scurvy? And the basic answer is, yes, we thought we knew how to cure scurvy and then we kind of forgot. And in a nutshell, this is a long and complicated, fascinating story, long and complicated story that I'll cut to the chase.
Basically, a lot of the oranges and lemons that they were giving people didn't have a lot of vitamin C. And in particular, for imperial reasons, the British Navy switched to limes from they had been growing Sicilian lemons. They switched to limes from the West Indies.
Limes don't have as much vitamin C in them. then you'd store the lime juice for a long, long time, which gradually lets the vitamin C deteriorate. And then you store it in copper containers, and the copper attacks the vitamin C.
And so a lot of it was basically didn't have any vitamin C, and it wasn't actually very effective. But nobody really noticed because the steam age had come in, and so ships were moving around much, much more quickly, and so people weren't at sea as long. And so the whole kind of relationship between the citrus fruit and the scurvy broke down. And then on these... Polar expeditions people started noticing that the lime juice they were being given, the Royal Navy was so famous people called them limeys, like because they always had lime juice.
They're going to the North Pole drinking lime juice, it's not working, they're getting scurvy, and so people start to think oh Lind was wrong all along. And given that germs had just been discovered, this is the the late 1800s, people started to develop this alternative theory that scurvy was in fact some kind of had some kind of bacteriological cause, and the whole thing about the lemons was wrong all along. And you might think, poor old James Lind, he solved the problem and no one believed him. But he didn't actually solve the problem. He seems to have no idea what it actually proved.
And he advocated the use of a boiled lemon syrup, which is not what he tested in his randomized trial, in his clinical trial. In his clinical trial it was fresh oranges and lemons. But boiled syrup, you boil the syrup, you destroy the vitamin C. So even the thing that he was advocating using would not have in fact worked. What am I talking about?
Why am I saying any of this? The point is, if you carefully count cases of diseases being cured and diseases not being cured in a properly structured randomized trial, that is absolutely essential to making progress with medicine. But you have to do more than that. You have to keep a really close track of exactly what it is you have actually measured. It's not limes that have been stored in copper containers for months.
It's not boiled lemon juice. You gave people oranges and lemons. If you're going to rely...
Until you understand what vitamin C is, you actually have to do what the original trial did. There's a whole structure of knowledge, of keeping track of what you've counted and what you've measured. that people seem to be very bad at doing. And I think it's astonishing that the first properly conducted, double-blind, randomized controlled trial published its results after the atomic bomb had been detonated.
It's that late. So it's such, it's this thing that Denise Schman-Besarat discovered of people overlooking counting again. It's such a simple idea. You get a bunch of people. You give some of them the treatment, you give some of them nothing, just in its most basic idea, and see...
Whether the people who got the treatment get better relative to the people who got nothing, it's such a simple idea. And the fact that you never saw it properly executed until 1946? Astonishing.
In fact, as late as the 1970s, the great Scottish doctor Archie Cochrane was still trying to convince the medical profession that, you know, you actually had to run clinical trials. You couldn't just rely on expertise. And in one particular case, this is in 1971, so he's running a randomised controlled trial of coronary care units. So the question is, should you treat people who've had heart attacks in a specialised coronary care unit, or should you basically just send them home and provide care at home?
And all the doctors who ran these coronary care units said it was incredibly important that these patients be treated in hospitals. You can't allow them to go home, it's dangerous. So Archie said, well...
Have we ever run a randomized trial? Turns out not. So he's running a randomized trial.
He tried to run it in Bristol, but wouldn't let him in Bristol, so in the end he ran it in Cardiff, because he had a joint professorship in Bristol and Cardiff. And so halfway through he writes, the results at that stage showed a slight numerical advantage for those who'd been treated at home. I rather wickedly compiled two reports, one reversing the number of deaths on the two sides of the trial. As we were going into the committee in the ante room, I showed some cardiologists the results.
They were vociferous in their abuse. Archie, they said, we always thought you were unethical. You must stop this trial at once. I let them have their say for some time and then apologized and gave them the true results, challenging them to say as vehemently that coronary care units should be stopped immediately.
There was dead silence and I felt rather sick because they were... after all, my medical colleagues. So these are the length, rather impish length, that Archie Cochranee had to go to in order to persuade doctors of the value of just systematically counting stuff, rather than going on the vibes.
We're experts, we think the coronary care units are good, why don't you actually measure it? Why don't you measure it properly and measure it fairly, which means a double blind clinical trial. But then Archie realised something else. You see, you can And this kind of goes back to the James Lind thing and the fact that all that knowledge was kind of just disappeared.
You can't just run clinical trials. You actually have to have the clinical trials available to the medical profession when they need them. And you need all the clinical trials available. Not just the one that happens to come across your desk, or not just the one that the drug companies have decided they'd like to publish. You need all of them available.
And so he wrote in 1979, It is surely a great criticism of our profession that we have not organised a critical summary by speciality or subspeciality, adapted periodically, of all relevant, randomised, controlled trials. Such a simple idea. Not only are we going to run these trials, we're going to organise them in a place where everybody can access them.
After Cochranee died, the Cochranee Library was set up, where those of you who are geeks like me may well know you can get rigorous summaries of all the best evidence available for any medical condition you can imagine. It's such a simple idea. The fact that we didn't really do it until the 1990s just goes to show again how much we undervalue the power of simple counting. So I talked about money, I talked about health, let's talk a little bit about power.
So in the 1880s, the American Census Bureau had a bit of a problem, which is that they ran a census in 1880, they're planning to run a census in 1890, and it became apparent that they might not actually finish analysing the results of the 1880 census before 1890. You see the problem was the census questions were getting more and more complicated. Not only you've got more and more people, but you're asking more and more detailed questions. The old census used to just say, well, you know, how many people are living in the house? And what are their genders?
Is there anybody blind living in the house? So you might write down, there's a blind person living in the house, but you don't know. Are they male or female?
Is it the head of the household? Or is it the adult? Is it the child?
You don't know. You certainly can't answer questions like, how many blind children are there in my city? Because I want to know, as the city mayor, I want to know whether I need to open a school for the blind.
These questions are impossible. So you start trying to answer more questions and be more systematic about all the questions that you ask. And you just lose track. And in the end, you get to a situation where there's no way.
we're going to be able to evaluate all the census returns by the time we conduct the next census. Don't even know what one journalist said. We don't even know how many married or single people there are in the United States and it's been eight years since the last time we run a census. So, um The Census Bureau decided it needed to solve the problem and it ran a competition for anybody who could think of a better way to count the census and there are a couple of teams of like uber nerds who approached this they had this very very organized counters and they were gonna have one of them had like different colored pens and one of them had stickers and it's like absolutely it's like it's like me in you know when I was 12 years old like pens all the stickers we're gonna do this we've got we've got the system for evaluating the census and basically they were given census returns from a city from the 1880 census and said how quickly can you tabulate everything get it all sorted.
So those were two teams There was a third team in this competition that the Census Bureau ran and the third team was a bit weird Was using this thing. That's this thing that looks a bit like a piano or like what are those they're sort of clocks? 40 clocks This is a punch card tabulation machine And it was invented by a man called German Hollerith, who asked his family for financial backing to help him make the machine.
They said, you're absolutely crazy and cut him off. But his idea was you have these cards. Previously, they'd be used in weaving. You have these cards with holes in them. And rather than a weaving pattern being expressed by the holes, you could express different answers to census questions.
with the holes in the punch cards. And you put the punch card into the machine, and these spring-loaded pins drop down, and the card blocks the pins, unless there's a hole, in which case the pins go through, and the pin makes contact with a little cup of open mercury. And that completes the electrical circuit and so you can start counting but you can also cross tabulate So you can answer questions like how many blind children are there in this city? so you could you cross reference the city and The age and the disability and you you come out with these answers and the Hollerith Machine was operated the guy who operated the machine. He was basically like I was amazing.
It's like communing with the Angels He was completely high on mercury But Was incredibly effective it was not even close So the human teams are taking weeks to tabulate this one city and the the holo earth machine was taking days And they weren't even asked to do cross tabulation the holo earth machine would have been even further ahead So this is a proto Proto computer. It's an incredible invention. This is Mr. Hollerith.
I'm saying nothing. So it's absolutely amazing. It saved the census, US census, millions of dollars.
It paved the way to the computer age and revolutionized the way we counted. So this is the industrialization of counting, which is great. When Hollerith retired, his company was, there was a merger and it became IBM, International Business Machines.
That was all brilliant until the 1930s when the German government said we really like some of these machines, we want to do a census 1930. The census had been postponed in Germany but when Adolf Hollerith came to power in 1933, suddenly very very keen on a census, very very keen on a census and get all the categories right. Let me read It's a wonderful history of the census by Andrew Whitby. The chair of IBM's German subsidiary in 1934 gave a speech. We are recording the individual characteristics of every single member of the nation onto a little card so that our physician, Hollerith, can determine whether the results calculated in this manner stand in a harmonious, healthy relation to one another.
Or whether unhealthy conditions must be cured by corrective interventions. I think I'll say no more about that. Let's talk about doing better.
And fundamentally we've got this incredibly powerful thing. Counting. Systematic counting. And it's a thing that we underrate, it's a thing that we undervalue, it's a thing that we underuse, just ask the last three and a half thousand years of doctors.
And it's also a thing that can be dramatically abused if we do it wrong. So if you just take health records, think about NHS records. NHS records are millions of people going back decades, and it's an incredible resource that we could use to unlock insights about different medical conditions and how best to help people.
Amazing resource. I mean, obviously, a little bit sensitive, what's in your NHS records, you know, when was the last time you were treated for sexually transmitted disease, for example, or did you wet the bed when you were 12? sort of thing.
So you wouldn't want your personal records getting out. But that's okay because we can release records, we do research on records, we can unleash the incredible power of counting and we just we take the names off them right. Take your name and address off and be fine. So Massachusetts did this. Hello everyone in Massachusetts.
Massachusetts, you see, topical. Massachusetts did this in the 1990s and William Weld, the governor of Massachusetts at the time, they did this for government employees. It was the government insurance company released all these anonymized records of everyone who was in the government health insurance scheme.
And William Weld, governor of Massachusetts, says this is fine because all of this data is anonymized. So... Latanya Sweeney, who's now a Harvard professor of computer science but at the time was a PhD student, said I'm not actually sure that's right.
So you think about it, think about my medical records. Let's say you want to find out the last time I was treated for a sexually transmitted disease, whether it was before or after I met my wife. Say you wanted to figure that out.
And let's say you had a whole bunch of medical records but they had all the, she's over there, she's absolutely looking, I'm not making eye contact with her. You can just like feel the left hand side of my face melting at this point. Let's say you wanted to know that.
So let's say you have a whole bunch of anonymized medical records that don't have my address and don't have my name, but do have my date of birth. So I don't know, how many dates of birth are there out there? So 365 dates of birth a year, go back 100 years, that's 36,500.
Let's not go back that far. Let's say there's 25,000 unique dates of birth. So.
Population of Oxford, I don't know, 250,000. So there's only 10 people sharing each particular date of birth. So you've just had the date of birth.
You've already narrowed it down to 10. So you've got the sex, you've narrowed it down to five. And then, you know, if you know, for example, that I live in Jericho, just because they couldn't persuade me to come very far to give this lecture. You know, you've got me and you can, you know, you can find out that, you know, the last time I had any treatment for any such thing.
You've nailed down. who I am from a completely anonymous database, because it turns out anonymization is not as effective as you might think. Latanya Sweeney reckons that if you've got the birth date, sex, and the five-digit zip code, which is quite a broad zip code, you can identify 87% of the population of the US just from those three pieces of information. And just to demonstrate her point, she downloaded an incredibly sensitive individual health record of an individual, all the diagnoses, all the treatments that were being prescribed and she sent it to the individual in question William Weld, Governor of Massachusetts.
Can we do better? Well yeah, I mean there's a group of people here in Oxford who are trying to do better. Is anyone from the Bennett Institute here?
No, excellent, so no one can say I'm doing, I'm describing their work incorrectly but here's the basic idea. You take these anonymized records and instead of just mailing them out and hoping the anonymization holds, you keep them in a secure data warehouse. Any researcher who wants access to the records, instead of just saying, oh I'm a researcher, I'm doing good work, trust me, send me the data.
Instead, the researcher has to give code querying the data to a nice little robot. The robot goes into the data warehouse, runs the code, comes back with the results. So you never get to touch the data, you never get to see the data, you only get to send your query into the data warehouse and bring it back. There's a lot of kind of clever stuff here, clever details that I don't have time to discuss, but one of the nice things about this is while you get absolute privacy for the patients, the researchers are doing everything in the open. So every line of code you write and submit, Every other researcher gets to see, every other researcher gets to use, every other researcher gets to share, every other researcher gets to debug if you've made a mistake.
So you've got complete transparency for the researchers but you've got complete privacy for the patients. If anybody is more interested in this kind of thing, check out OpenSafely who run basically a couple hundred yards away from here. The point is we have this power and I think we're not using it very well. We're systematically undervaluing how useful counting can be.
Do you know the... everyone talks about food poverty, everyone talks about the cost of living crisis. The Department for Work and Pensions published their first ever estimate of the use of food banks last year. Until last year we had absolutely no idea how many families in the UK were using food banks.
It just gives you an example of the way that we just don't bother to collect relevant data. So we overlook data, we underuse data and misuse data, and we abuse data. I think we can do vastly better. I think the Open Safely team are an example of how we can do vastly better. I think there's an enormous amount of potential still to be explored.
I've talked for far too long. I've left only five minutes for questions. I apologise to Daryl, who told me to leave 15. But there you go. Questions, we have a magic microphone for questions. Please ask your questions.
And thank you all very much for listening.