IFTF hosted the second Quantified Self Show&Tell, and since then they’ve been curious about and supportive of this extended investigation into the meaning of what more academic observers call “personal informatics.” Recently, Bradley Kreit interviewed me about the implications of The Quantified Self for the IFTF Health Horizons Report. His interview and excellent editing helped me express what I think is happening in a fairly concise way. I’ve republished it with Bradley’s permission below.

—

IFTF Health Horizons

Gary Wolf is a contributing editor at Wired magazine and the co-host of The Quantified Self, a blog dedicated to self-knowledge through numbers (www.quantifiedself.org). At Wired, he has been the author of a number of the magazine’s most frequently cited articles, including “The Curse of Xanadu,” about Ted Holmes Nelson and the invention of hypertext; “The World According to Woz,” about Apple co-founder Steve Wozniak; and “The Wisdom of St. Marshall, Holy Fool,” about Marshall McLuhan. He has also written about Piotr Wozniak, creator of the memory program SuperMemo, and recently about Craigslist and its founder, Craig Newmark.

IFTF: The phenomenon of the quantified self is an early form of personal health forecasting. What is the idea behind it?

GW: Numbers play a key role in analyzing all kinds of phenomena, from the largest phenomena of the cosmos using radio telescopes to the smallest phenomena in the universe—the analysis, say, of subatomic particles. We have statistical tools of great sophistication for gathering data and finding meaning in it. It seems only natural that we would want to use some of these techniques to gain knowledge about ourselves.

This is so obvious that it might almost seem trivial, except when you realize that we usually associate self-knowledge not with numbers but with words—a kind of inner voice of consciousness and conscience. I think that supplementing that with quantitative tools is one of the most interesting trends emerging in our culture today. This interest is based on the highly practical results of experiments that people are doing in collaborative diagnosis and collaborative evaluation of treatments for chronic conditions, as well as experiments that involve the analysis and acceleration of learning.

IFTF: In some of your writing about the quantified self, you’ve talked about a concept called a macroscope. What do you mean by that, particularly as it relates to health?

GW: The word macroscope has been used quite a few times in quite a few contexts. It’s an interesting word; its meaning is trying to emerge and everyone’s taking a crack at it, but it’s finally settling down into a useful concept.

My meaning is taken from Jesse Ausubel, a climate scientist who is also a professor at The Rockefeller University. It simply refers to gathering data in nature through distributed methods, often through sensor networks, and then analyzing it on a computer. The particular pieces of technology for gathering this data are familiar; it is how they are now being combined that is interesting. We are beginning to see them being used in the context of a social process that produces data that would be inaccessible to an individual researcher trying to build this network from scratch.

The macroscope concept can be applied to the many individuals keeping track of some aspect or aspects of their lives. You have people tracking sleep, diet, exercise, productivity, symptoms, and so on. With all this tracking, a tremendous amount of health-related data is being produced. When that data is analyzed, you learn things that would be much harder to learn using the traditional methods of a clinical trial or a population study.

IFTF: Do you expect self-tracking will become widespread over the next ten years?

GW: I think it will become a mainstream, almost ubiquitous practice and at the same time will become invisible because it will be blend in with daily life. I think a good comparison is with the fate of computing. At one time, the people who used computers tended to be the kind of people who liked it. Over time, the process of computing has been incorporated into so many technologies and devices that many of the things we do that involve computing don’t seem like computing at all. Think of using a pedometer or step counter, or standing on a digital scale. The computing component is disappearing, and the self-tracking aspect will, too.

Self-tracking will disappear because it will be taken for granted. The quantitative tools in our lives will produce data that will be incorporated into some feedback mechanism; we will look at those mechanisms and they will influence us in some way. For instance, we will get biometric data in the form of feedback about how well we’re eating and sleeping, but we won’t have to peel back that information and do the analysis ourselves. Of course, the people who will be making these products and services will be highly aware of their tracking components, but if they’re successful, users won’t think about those aspects.

IFTF: Do you foresee any difficulties with privacy or concerns over control of information? Will individuals not want to share the detailed and intimate information that will be collected about them?

GW: Although gathering personal data will become mainstream, I don’t think most people will want to share their data. We can identify some people as sharer types with respect to their health and biometric data; they are closely linked to the pioneer type because they have a vision of what sharing may bring. But for the most part I think the benefits of the macroscope will be very hard to achieve under a system in which people can be punished harshly on the basis of their numbers. And we live in a world where if you have bad numbers, you will be punished.

IFTF: Isn’t one of the core challenges that the data is most useful in large-scale aggregations, but to get that you have to be able to get people to share their data?

GW: Let’s back up a bit: useful to whom? The data is very useful to you, whether or not it’s aggregated. You can see the macroscope as having multiple guises: there’s the social macroscope, which aggregates data across individuals, and that’s where the privacy issues come in, but you can also interpret the macroscope on an individual level. I can have multiple sensors at multiple times, all aggregating the data for me; I can do experiments of one, and the data never has to leave my computer.

IFTF: So how do you bridge that gap to make the social macroscope feasible?

GW: We need to articulate as clearly as possible that there must be a transformation in terms of how we look at what health and health care mean. As long as health care is considered from the perspective of the individual, there are many benefits that we’ll be missing.

Do you know anything about Qoogle? I am not linking to the site directly, because I don’t know what they are up to, and though I suspect it is just a link-farming scheme, it could be something more nefarious. But here are some screenshots.

I searched this morning to check the online commentary about Ryan Sorba, who made some news with his anti-gay rant at the recent CPAC conference.(Andrew Sullivan has an account here.) Here is a screenshot of the first page of results on Google. The text on this image is hard to read, but the third link (pretty good, Mr. Black Hat!) is a web page called Ryan Sorba, and below it is a snippet that contains some “word salad”: Feb 19, 2010 … This brueghel is an ryan sorba of the gipsywort disentangled by jawless oxidization sardinian in the solon, cryogen, and dreaming of …

I’ve been getting this sort of thing quite often when searching on names. If you click on the link, you get to a fake Google page at Qooglesearch.com, showing results for Ryan Sorba that appear to be scrapped from Google, along with two “sponsored links” from “flashbuzz.net” at the top. I assumed that these sponsored links were the payload, but on subsequent clicks they have disappeared. Another link associated with Qoogle appears to be imasion-corp.com. There is an intermediate link that appears also: tdss… – but it is hard to catch and I didn’t get it.

Below is an image of the Qooglesearch page.

Sort of convincing as a Google page, if you aren’t paying too much attention.

But the best image comes from clicking on the “cached” link under the Qoogle “Ryan Sorba” entry on the original Google search page. That shows you what Qoogle is showing the Google Bot as it surfs the web. Colorful!

Well, it may be just another case of “pissing in the pool” as Craigslist CEO Jim Buckmaster calls it. But I’d like to know more about Qoogle, and if you have any tips, please send them along.

Two years ago I traveled to Costa Rica and met two amazing scientists: Daniel Janzen and Willie Hallwachs. Janzen and Hallwachs and their many colleagues – professional scientists, students, and an important group of “parataxonomist” collectors – work the Area de Conservación de Guanacaste, where they catalog and study the organisms that live there, while trying to protect and expand the park.

Automeris zugana?

A few weeks later I went to Ontario to meet Paul Hebert, a geneticist at the University of Guelph and the inventor of a novel method for identifying species using very short snippets of DNA. Hebert was using material – moths and wasps, mainly – collected in Guanacaste to prove that his idea was valid.

When I wrote about it, this method was highly controversial. Could you really take part of the job of the expert taxonomist – the identification of species – and routinize it, reduce the cost, and provide identifications of even very hard to tell apart animals as a kind of standard technical service? Or, as Dan Janzen liked to ask: could we build a species tricorder?

A couple of months ago I found out that my story, A Simple Plan to ID Every Creature on Earth, is going to be awarded the 2009 AAAS Kavli Science Journalism Award for the best science story published in a magazine during the preceding year. I’m very proud of this award, and grateful to Dan, Winnie, Paul, and their colleagues for having been willing to answer endless questions about their work. Critics of barcoding – also eminent scientists, including the Director of the Jepson Herbarium Brent Mishler - were equally open and helpful.

I took many photos on my trip, and I thought I would post a few of them along with some updated barcoding material before I head down to San Diego for the AAAS meeting.

Here is a photo of Dan at his desk. Below it is the description of him working that opened my piece for Wired.

The utopian lepidopterist holds a pin in each hand. His style is ambidextrous and probably unique. He catches two forewings of a dead moth simultaneously and pins them to his drying board, and then, in a continuous sweep, he does the same with the hind wings. He repeats these motions again and again, like a conductor with tiny batons. Outside, it is hot and bright. Inside, it is hot and dark. The lepidopterist, whose name is Dan Janzen, has been working here in this Costa Rican forest for more than 40 years. He is married to his research partner, Winnie Hallwachs, and the two of them occupy a small house with a roof of corrugated metal whose eaves cast deep shade. During the day they work under artificial light. At night bats flit through the gaps at the top of the wall, do hairpin turns in the air, and exit again without slowing. The utopian lepidopterist’s aim is to put names on all the moths and butterflies in the forest. He wants to know more than just the names, of course; he wants to know who lives where and who eats whom and to unravel the mysteries of the ecosystem. But his first question is always the most basic one. This moth, here on the drying board: What is it called?

Here is one of the rearing stations Janzen and his colleagues operate. Each bag contains a caterpillar and some of the leaves of the plant it was found on.

[The area is...] strung with ropes. Beneath the ropes hang hundreds of plastic bags full of leaves, and inside every bag there is a caterpillar, a pupa, a moth, or some flies or wasps that have managed to parasitize the caterpillar, eat the pupa, and emerge into the middle of this scientific experiment. Like the insects in the neighboring bags, the destiny of these parasites is to be frozen, dried, identified, barcoded, and shipped to a museum for reference. Here, and in 10 other caterpillar-rearing stations in the forest, Janzen, Hallwachs, and their many local collaborators have solved taxonomic mysteries that go back hundreds of years. “Some of these moths have had names forever, and their caterpillars, too, and they’ve never been recognized as the same species,” Janzen says.

Janzen is 71. I enjoyed grabbing this picture of him on the road the rearing station.

There was electricity available at the rearing station, so I got to sit outside and catch up on my notes….

What has happened after the story appeared? Was the controversy resolved? Over the last two years, the advocates of barcoding appear to have won the day. I keep up with ongoing news via Mark Stoeckle’s detailed posts on The Barcoding Life Blog. Recently, reported on a project to use barcoding to determine the species of fly larvae that are important in forensics; although the larvae which feed on corpses can be telling evidence of the time of death, figuring out what species you have can be hard when all you can see is a tiny wriggling worm, with virtually no distinguishing features.

Barcoding also made news recently in Science, with a report by Elizabeth Pennisi on the recent completion of a barcoding analysis of a collection of more than 2500 parasitoid wasps. (The image below is from Dan and Winnie’s collection of material, it shows the actual wasp larvae (Euplectrus walteri) emerging from the caterpillar – yum!)

Pennisi, in her story, The Little Wasp That Could, (full text behind a subscription wall, unfortunately) gives an excellent description of why figuring out what wasp you have on your hands is interesting, important, and – before barcoding – very hard:

The British geneticist J. B. S. Haldane once famously quipped that God seems to have had an inordinate fondness for beetles, given their numbers and diversity. If so, then he must have been besotted by parasitoid wasps. Tinkerbells of the animal kingdom, many of these insects are no bigger than fleas, yet they may well outnumber beetles.

Unlike beetles, however, parasitoid wasps aren’t exactly charismatic. “You get one in your eye and pull it out with your finger and think it’s a piece of dust,” says Daniel Janzen, an ecologist at the University of Pennsylvania. “There’s millions of individuals out there, and you don’t even know they exist.” Yet these inconspicuous insects play a crucial role in natural ecosystems and in agriculture. They destroy the eggs, larvae, or cocoons of countless species of insects and arthropods, sometimes with hugely beneficial effects: The U.S. Department of Agriculture (USDA) estimates that parasitic wasps save the United States at least $20 billion annually by controlling invasive species. “I think very few people realize what a force they are in the biology of our planet,” says Michael Strand, an entomologist at the University of Georgia, Athens.

A 2008 DNA bar-coding analysis of 2597 parasitoid wasps from his collection turned up 313 species, not the 171 researchers had previously thought. M. Alex Smith of the University of Guelph in Canada and Josephine Rodriguez of UC Santa Barbara discovered that what was believed to be a single species—a 2-millimeter-long wasp called Apanteles leucostigmus with a black body and a white rhomboid patch on its wing—proved to be 36. And there were many more examples of previously unrecognized species, Janzen, Rodriguez, Smith, and their colleagues reported in the 26 August 2008 issue of the Proceedings of the National Academy of Sciences.

Alex Smith, mentioned in the above passage, was on one of the trips I made to Costa Rica. I enjoyed his company tremendously. Here is a picture of him with a beetle on his head. He did not put it there. We were collecting moths at night, and just saw a nice spot and decided to land.

Finally, here a couple of good videos that capture some of the interesting stuff going on in this line of work. The first is a short video about parasitoid wasps; it is very well explained and there are a few great images.

But this one, a Google talk by Dan Janzen and Paul Hebert, is more informative and inspirational. Here Janzen and Hebert lay out the whole picture of what it would mean to be able to identify species on the planet, and how we could get there.

There is an excellent collection of links here, at the Ontario Genomics Institute, with more information about the relationship between DNA barcoding, species identification, and environmental conservation.

Stumbled upon the chanterelle motherload yesterday – one day late. They were waterlogged and hopeless after a solid night of rain. These are California chanterelles, which got proper acknowledgment as a distinct species only in 2008,when David Arora and Susie Dunham published an article in Economic Botany giving it a name of its own: Cantherellus californicus.

I have always wanted to write a long profile of Arora, one of the most independent thinkers I have ever met. His analysis of the relationship between ecology, conservation, and commercial use of natural resources changed my way of thinking about nature.

In a story I wrote about the commercial mushroom harvest in the Yukon territory, I had a chance to spend a week with him in the far north. I got a few paragraphs about him into the story, but most of what we talked about is still sitting in notebooks, waiting for the opportunity to become part of a story. His house in Northern California is very inspiring and strange, with a pond on the inside where he grows algae, and a forest outside that he manages to encourage the appearance of hundreds of fat Porcini when the time is right.

Arora’s giant book, Mushrooms Demystified, is used by everybody on the west coast with an interest in mushrooms. My copy is falling apart and indispensable. He has never updated it. The reasons are complicated, and related to his shifting views of the relationship between wild nature and the humans who are part of it.

But that’s a story for another day…

The story in this month’s Wired started when the magazine’s editors asked me a pointed question: how can a site that’s so good be so bad? Serving a vast community at an irresistible price (mostly free), craigslist nonetheless seemed the antithesis of what a modern web business should be. Oblivious to innovation and stuck in a 1997 mindset, craigslist was hogging the sector and holding things back. When the editors invited me in to propose my writing the story, they wanted an exposé.

Meanwhile, I had been wanting write about craigslist for years, ever since I saw founder Craig Newmark and CEO Jim Buckmaster give their now famous talk at the 2004  San Francisco Web2.0 conference.  (The full recording can be heard here.) Newmark stood on a milk crate, which put him nearly eye-to-eye with Buckmaster. They described their philosophy of branding (against it), of graphic design (against it), and, most intriguingly, of money. Money they were not against. But they were not exactly in favor of it, either. They seemed to think of money as a danger. Their extreme, almost theatrical caution in relation to cash was familiar to me. Many writers, musicians, and artists have it. But these were not writers, musicians or artists, they were two men running a classified advertising site. As craigslist grew dominant, its managers’ profession of disinterest in profit came to seem more and more anomalous. I did not assume they were hypocritical. I assumed they were interesting.

When we first discussed the story, the vehemence of the editors’ point of view caught me off guard. I use craigslist. The couch in my living room is from craigslist. I got rid of my moving boxes on craigslist. On the other hand, somebody tried to rob me once when I met them to purchase a DVD player they had listed on craigslist. I find searching the site to be absurdly laborious. I could see the points the editors were making against craigslist. Nonetheless I didn’t hate it. I didn’t know anybody who hated it. The most negative response I’d ever seen to craigslist  was a shoulder shrug and an eye roll that meant: “Hey, it’s free; what do you expect?”

But as I began to talk to people inside and outside the company, I began to understand that while the intensity of feeling against craigslist was a minority view, it was helpful to think about, because it came from a rare appreciation of the company’s power. To a user craigslist appears to be a local listings site. But behind each page of blue links is a publishing system that serves more than 45 million people each month, and produces 100 million dollars in revenue annually. More importantly, craigslist’s success has helped take down an entire sector of the publishing industry. The anger against Craigslist, an anger that I thought deserved to be expressed with some irony, was nonetheless  an honest tribute both to its significance and to the hopes and expectations it provokes.

In the end, after a couple of months of research, I told the editors I thought I was prepared to answer the question they originally asked: “Why can’t craigslist be better?” The answer exposed some normally unchallenged assumptions about what better means. In the series of posts that should go up over the next couple of weeks, I’m going to go into some of the details that didn’t make it into the magazine version of the story, details that may interest people who find themselves fascinated, as I was, with one of the most unusual businesses in the world.

[Update: I posted two follow-up pieces on Wired's Epicenter blog: Craigslist vs. eBay, and Bad Advice for Newspapers]

Watch this video published from a story in New Scientist by Lars Hall and Petter Johansson.

[I]n an early study we showed our volunteers pairs of pictures of faces and asked them to choose the most attractive. In some trials, immediately after they made their choice, we asked people to explain the reasons behind their choices.

Unknown to them, we sometimes used a double-card magic trick to covertly exchange one face for the other so they ended up with the face they did not choose. Common sense dictates that all of us would notice such a big change in the outcome of a choice. But the result showed that in 75 per cent of the trials our participants were blind to the mismatch, even offering “reasons” for their “choice”.

This is troubling enough, but there’s more. When people are fooled into thinking they made a different choice than the one they actually made, and then articulate their “reasons” for this supposed choice, they then may actually change their future preferences to conform to their confabulated preference.

Importantly, the effects of choice blindness go beyond snap judgments. Depending on what our volunteers say in response to the mismatched outcomes of choices (whether they give short or long explanations, give numerical rating or labeling, and so on) we found this interaction could change their future preferences to the extent that they come to prefer the previously rejected alternative. This gives us a rare glimpse into the complicated dynamics of self-feedback (”I chose this, I publicly said so, therefore I must like it”), which we suspect lies behind the formation of many everyday preferences.

Lars Hall and Petter Johansson lead the Choice Blindness Laboratory at Lund University, Sweden. At the end of their New Scientist piece, they suggest that learning about this experiment should make people better at understanding their own choices.

In everyday decision-making we do see ourselves as connoisseurs of our selves, but like the wine buff or art critic, we often overstate what we know. The good news is that this form of decision snobbery should not be too difficult to treat. Indeed, after reading this article you might already be cured.

Unfortunately, this is not convincing. It is common for biases persist even when we are warned about them. I suspect we are in no position to stand guard over our judgments without the help of machines to keep us steady. Assuming, that is, that deliberative consistency is a value we care to protect.

The number of Web applications to collect data and information about yourself continues to grow; if you want to track something, most likely there’s an online tool to do it. This is great – especially since a lot of the applications seem to have a lot of users, which means an interest in data… However, as users, developers, and designers, we shouldn’t be satisfied too quickly with what we have. Want more. Demand more. It’s interesting and oftentimes fun to log data about your life – whether it be when you go the bathroom, your sugar levels, or your mood. You get some nice graphs and charts, it looks cool, and maybe you learn something about yourself.

But all the self-surveillance tools so far are mostly about a single dataset or two at most. You track your weight and what you eat, but it’s more complex than that. Life is complicated and data is an abstraction of life after all. Do you eat when you’re depressed or are you depressed when you eat? Do you feel better if you exercise? What about sleep? How much sleep and exercise is best for you? What days should you exericse and how many days in a row and for how long? What truly makes you happy? I want my self-surveillance application to not only give me the ability to find these answers but to give them to me with very little effort on my part.

Nathan argues that any good solution for part of the problem ought to at least aspire to solve all of it. He wants the tools to include some data processing, and to be ubiquitous, so that you can post from anywhere.

In the end, I want all of my data in one place with some machine learning in the background and the ability to analyze and visualize easily and thoroughly. We’re not quite there yet, but I’m looking forward to when we do. Information overload? No. Better-educated decisions and a completely different view of ourselves and our surroundings? Definitely.

Nathan is building his own multi-tracker at your.flowingdata.com. Right now it is by invitation only, but you can follow yfd on Twitter to connect to the next wave of invites.

At the second QS Show&Tell, Joe Betts-Lacroix gave a short talk about his dream system: a website that could receive data and put it into a database, with data would be gathered by little devices that could beam it to the web using simple protocols. (The picture below is of Dan Brown, not Joe Betts-Lacroix; Dan happened to be in the first frame of this segment of video, which is automatically used for reference. Joe shows up a few seconds in.)

In this version of the dream the ideal Web site would have some
simple graphic tools, the ability to export data, good security, and some sharing and privacy options. Since then, there have been quite a few demonstrations of various ideas at the QS Show&Tell meetings, as well as a steady stream of products and, naturally, announcements of products (cf. Fitbit) that aim to achieve some parts of what Nathan, Joe, and other self-trackers have called for.

My own vision is slightly different. I think we are inevitably going to see a bunch of competing solutions, most of which will seem pretty good to some people and deeply flawed to others. People come at self-tracking with different goals and values. Daytum, which is mainly about self-expression, will be nifty for the person who uses data mainly as a feature of personal identity. Daytum’s origin is in the Feltron Annual Report by Nicholas Felton; an annual report serves many purposes, but data analysis is not one of them. Meanwhile, Zume Life is designed for people tracking serious health conditions, who are often trying to manage complex prescription drug regimens. Zume Life has made the data entry process almost as easy as imaginable until the day arrives when we can beam our data from monitoring devices without intervening at all. Along with an iPhone app, there is a voice transcription service. Just press a button, say your number (or food eaten, or exercise accomplished) and a person will transcribe it into a database. This is going forward by going backward, and there is a kind of genius to it. The type of person who likes Daytum is not going to bother with Zume Life, and for the user of Zume Life, Daytum is pointless.

We are headed into a messy, confusing, and interesting period in self-tracking, when lots of new solutions emerge, each claiming a piece of territory and pushing up against neighbors. There will be no ideal system, but a bunch of different sytesms, and then a bunch of different solutions for gluing different parts of different systems together. Some of the people who manage to aggregate lots of users will find excuses for holding on to them. (This may not always be a bad thing: see Dead ends and walled gardens.) But others will see that making the data collected on their system available in standard form will speed adoption.

But what is the standard? You can insert your own favorite standards horror story here. But after you’ve given yourself the shivers, you can recover with the realization that this conflict over standards occurs because everybody can finally see the prize they are wrestling for, which means that a substantial amount of agreement has been won. Once important people start making highly emotional arguments for how quotidian personal data (QPD – now it’s official) should be represented, you can start celebrating.

Note that I said “quotidian personal data” and not “healthcare information.” QPD is the type of thing you are willing transmit in a text message, and SMS is an easy bet for the ubiquitous medium for QPD. But if you call it healthcare information, I’m out. They fight dirty.

Kurzweil’s notion of a singularity is taken from cosmology, in which it signifies a border in spacetime beyond which normal rules of measurement do not apply (the edge of a black hole, for example). The word was first used to describe a crucial moment in the evolution of humanity by the great mathematician John von Neumann. One day in the 1950s, while talking with his colleague Stanislaw Ulam, von Neumann began discussing the ever-accelerating pace of technological change, which, he said, “gives the appearance of approaching some essential singularity in the history of the race beyond which human affairs as we know them could not continue.”

Many years later, this idea was picked up by another mathematician, the professor and science fiction writer Vernor Vinge, who added an additional twist. Vinge linked the singularity directly with improvements in computer hardware. This put the future on a schedule. He could look at how quickly computers were improving and make an educated guess about when the singularity would arrive. “Within 30 years, we will have the technological means to create superhuman intelligence,” Vinge wrote at the beginning of his 1993 essay The Coming Technological Singularity: How to Survive in the Post-Human Era. “Shortly after, the human era will be ended.” According to Vinge, superintelligent machines will take charge of their own evolution, creating ever smarter successors. Humans will become bystanders in history, too dull in comparison with their devices to make any decisions that matter.

Kurzweil transformed the singularity from an interesting speculation into a social movement. His best-selling books The Age of Spiritual Machines and The Singularity Is Near cover everything from unsolved problems in neuroscience to the question of whether intelligent machines should have legal rights. But the crucial thing that Kurzweil did was to make the end of the human era seem actionable: He argues that while artificial intelligence will render biological humans obsolete, it will not make human consciousness irrelevant. The first AIs will be created, he says, as add-ons to human intelligence, modeled on our actual brains and used to extend our human reach. AIs will help us see and hear better. They will give us better memories and help us fight disease. Eventually, AIs will allow us to conquer death itself. The singularity won’t destroy us, Kurzweil says. Instead, it will immortalize us.

There are singularity conferences now, and singularity journals. There has been a congressional report about confronting the challenges of the singularity, and late last year there was a meeting at the NASA Ames Research Center to explore the establishment of a singularity university. The meeting was called by Peter Diamandis, who established the X Prize. Attendees included senior government researchers from NASA, a noted Silicon Valley venture capitalist, a pioneer of private space exploration, and two computer scientists from Google.

At this meeting, there was some discussion about whether this university should avoid the provocative term singularity, with its cosmic connotations, and use a more ordinary phrase, like accelerating change. Kurzweil argued strongly against backing off. He is confident that the word will take hold as more and more of his astounding predictions come true.

Kurzweil does not believe in half measures. He takes 180 to 210 vitamin and mineral supplements a day, so many that he doesn’t have time to organize them all himself. So he’s hired a pill wrangler, who takes them out of their bottles and sorts them into daily doses, which he carries everywhere in plastic bags. Kurzweil also spends one day a week at a medical clinic, receiving intravenous longevity treatments. The reason for his focus on optimal health should be obvious: If the singularity is going to render humans immortal by the middle of this century, it would be a shame to die in the interim. To perish of a heart attack just before the singularity occurred would not only be sad for all the ordinary reasons, it would also be tragically bad luck, like being the last soldier shot down on the Western Front moments before the armistice was proclaimed.

Photo: Garry McLeod

In his childhood, Kurzweil was a technical prodigy. Before he turned 13, he’d fashioned telephone relays into a calculating device that could find square roots. At 14, he wrote software that analyzed statistical deviance; the program was distributed as standard equipment with the new IBM 1620. As a teenager, he cofounded a business that matched high school students with colleges based on computer evaluation of a mail-in questionnaire. He sold the company to Harcourt, Brace & World in 1968 for $100,000 plus royalties and had his first small fortune while still an undergraduate at MIT.

Though Kurzweil was young, it would have been a poor bet to issue him life insurance using standard actuarial tables. He has unlucky genes: His father died of heart disease at 58, his grandfather in his early forties. He himself was diagnosed with high cholesterol and incipient type 2 diabetes — both considered to be significant risk factors for early death — when only 35. He felt his bad luck as a cloud hanging over his life.

Still, the inventor squeezed a lot of achievement out of these early years. In his twenties, he tackled a science fiction type of problem: teaching computers to decipher words on a page and then read them back aloud. At the time, common wisdom held that computers were too slow and too expensive to master printed text in all its forms, at least in a way that was commercially viable.

But Kurzweil had a special confidence that grew from a habit of mind he’d been cultivating for years: He thought exponentially. To illustrate what this means, consider the following quiz: 2, 4, ?, ?.

What are the missing numbers? Many people will say 6 and 8. This suggests a linear function. But some will say the missing numbers are 8 and 16. This suggests an exponential function. (Of course, both answers are correct. This is a test of thinking style, not math skills.)

Human minds have a lot of practice with linear patterns. If we set out on a walk, the time it takes will vary linearly with the distance we’re going. If we bill by the hour, our income increases linearly with the number of hours we work. Exponential change is also common, but it’s harder to see. Financial advisers like to tantalize us by explaining how a tiny investment can grow into a startling sum through the exponential magic of compound interest. But it’s psychologically difficult to heed their advice. For years, an interest-bearing account increases by depressingly tiny amounts. Then, in the last moment, it seems to jump. Exponential growth is unintuitive, because it can be imperceptible for a long time and then move shockingly fast. It takes training and experience, and perhaps a certain analytical coolness, to trust in exponential curves whose effects cannot be easily perceived.

Moore’s law — the observation by Intel cofounder Gordon Moore that the number of transistors on an integrated circuit doubles roughly every 18 months — is another example of exponential change. For people like Kurzweil, it is the key example, because Moore’s law and its many derivatives suggest that just about any limit on computing power today will be overcome in short order. While Kurzweil was working on his reading machine, computers were improving, and they were indeed improving exponentially. The payoff came on January 13, 1976, when Walter Cronkite’s famous sign-off — “and that’s the way it is” — was read not by the anchorman but by the synthetic voice of a Kurzweil Reading Machine. Stevie Wonder was the first customer.

The original reader was the size of a washing machine. It read slowly and cost $50,000. One day late last year, as a winter storm broke across New England, I stood in Kurzweil’s small office suite in suburban Boston, playing with the latest version. I hefted it in my hand, stuck it in my pocket, pulled it out again, then raised it above a book flopped open on the table. A bright light flashed, and a voice began reading aloud. The angle of the book, the curve of its pages, the uneven shadows — none of that was a problem. The mechanical voice picked up from the numerals on the upper left corner — … four hundred ten. The singularity is near. The continued opportunity to alleviate human distress is one key motivation for continuing technological advancement — and continued down the page in an artificial monotone. Even after three decades of improvement, Kurzweil’s reader is a dull companion. It expresses no emotion. However, it is functionally brilliant to the point of magic. It can handle hundreds of fonts and any size book. It doesn’t mind being held at an angle by an unsteady hand. Not only that, it also makes calls: Computers have become so fast and small they’ve nearly disappeared, and the Kurzweil reader is now just software running on a Nokia phone.

In the late ’70s, Kurzweil’s character-recognition algorithms were used to scan legal documents and articles from newspapers and magazines. The result was the Lexis and Nexis databases. And a few years later, Kurzweil released speech recognition software that is the direct ancestor of today’s robot customer service agents. Their irritating mistakes taking orders and answering questions would seem to offer convincing evidence that real AI is still many years away. But Kurzweil draws the opposite conclusion. He admits that not everything he has invented works exactly as we might wish. But if you will grant him exponential progress, the fact that we already have virtual robots standing in for retail clerks, and cell phones that read books out loud, is evidence that the world is about to change in even more fantastical ways.

Look at it this way: If the series of numbers in the quiz mentioned earlier is linear and progresses for 100 steps, the final entry is 200. But if progress is exponential, then the final entry is upwards of a nonillion, computers will soon be smarter than humans, and nobody has to die.

In a small medical office on the outskirts of Denver, with windows overlooking the dirty snow and the golden arches of a fast-food mini-mall, one of the world’s leading longevity physicians, Terry Grossman, works on keeping Ray Kurzweil alive. Kurzweil is not Grossman’s only client. The doctor charges $6,000 per appointment, and wealthy singularitarians from all over the world visit him to plan their leap into the future.

Grossman’s patient today is Matt Philips, 32, who became independently wealthy when Yahoo bought the Internet advertising company where he worked for four years. A young medical technician is snipping locks of his hair, and another is extracting small vials of blood. Philips is in good shape at the moment, but he is aware that time marches on. “I’m dying slowly. I can’t feel it, but I know its happening, little by little, cell by cell,” he wrote on his intake questionnaire. Philips has read Kurzweil’s books. He is a smart, skeptical person and accepts that the future is not entirely predictable, but he also knows the meaning of upside. At worst, his money buys him new information about his health. At best, it makes him immortal.

“The normal human lifespan is about 125 years,” Grossman tells him. But Philips wasn’t born until 1975, so he starts with an advantage. “I think somebody your age, and in your condition, has a reasonable chance of making it across the first bridge,” Grossman says.

According to Grossman and other singularitarians, immortality will arrive in stages. First, lifestyle and aggressive antiaging therapies will allow more people to approach the 125-year limit of the natural human lifespan. This is bridge one. Meanwhile, advanced medical technology will begin to fix some of the underlying biological causes of aging, allowing this natural limit to be surpassed. This is bridge two. Finally, computers become so powerful that they can model human consciousness. This will permit us to download our personalities into nonbiological substrates. When we cross this third bridge, we become information. And then, as long as we maintain multiple copies of ourselves to protect against a system crash, we won’t die.

Kurzweil himself started across the first bridge in 1988. That year, he confronted the risk that had been haunting him and began to treat his body as a machine. He read up on the latest nutritional research, adopted the Pritikin diet, cut his fat intake to 10 percent of his calories, lost 40 pounds, and cured both his high cholesterol and his incipient diabetes. Kurzweil wrote a book about his experience, The 10% Solution for a Healthy Life. But this was only the beginning.

Kurzweil met Grossman at a Foresight Nanotech Institute meeting in 1999, and they became research partners. Their object of investigation was Kurzweil’s body. Having cured himself of his most pressing health problems, Kurzweil was interested in adopting the most advanced medical and nutritional technologies, but it wasn’t easy to find a doctor willing to tolerate his persistent questions. Grossman was building a new type of practice, focused not on illness but on the pursuit of optimal health and extreme longevity. The two men exchanged thousands of emails, sharing speculations about which cutting-edge discoveries could be safely tried.

Though both Grossman and Kurzweil respect science, their approach is necessarily improvisational. If a therapy has some scientific promise and little risk, they’ll try it. Kurzweil gets phosphatidylcholine intravenously, on the theory that this will rejuvenate all his body’s tissues. He takes DHEA and testosterone. Both men use special filters to produce alkaline water, which they drink between meals in the hope that negatively charged ions in the water will scavenge free radicals and produce a variety of health benefits. This kind of thing may seem like quackery, especially when promoted by various New Age outfits touting the “pH miracle of living.” Kurzweil and Grossman justify it not so much with scientific citations — though they have a few — but with a tinkerer’s shrug. “Life is not a randomized, double-blind, placebo-controlled study,” Grossman explains. “We don’t have that luxury. We are operating with incomplete information. The best we can do is experiment with ourselves.”

Obviously, Kurzweil has no plan for retirement. He intends to sustain himself indefinitely through his intelligence, which he hopes will only grow. A few years ago he deployed an automated system for making money on the stock market, called FatKat, which he uses to direct his own hedge fund. He also earns about $1 million a year in speaking fees.

Meanwhile, he tries to safeguard his well-being. As a driver he is cautious. He frequently bicycles through the Boston suburbs, which is good for physical conditioning but also puts his immortality on the line. For most people, such risks blend into the background of life, concealed by a cheerful fatalism that under ordinary conditions we take as a sign of mental health. But of course Kurzweil objects to this fatalism. He wants us to try harder to survive.

His plea is often ignored. Kurzweil has written about the loneliness of being a singularitarian. This may seem an odd complaint, given his large following, but there is something to it. A dozen of his fans may show up in Denver every month to initiate longevity treatments, but many of them, like Matt Philips, are simply hedging their bets. Most health fanatics remain agnostic, at best, on the question of immortality.

Kurzweil predicts that by the early 2030s, most of our fallible internal organs will have been replaced by tiny robots. We’ll have “eliminated the heart, lungs, red and white blood cells, platelets, pancreas, thyroid and all the hormone-producing organs, kidneys, bladder, liver, lower esophagus, stomach, small intestines, large intestines, and bowel. What we have left at this point is the skeleton, skin, sex organs, sensory organs, mouth and upper esophagus, and brain.”

In outlining these developments, Kurzweil’s tone is so calm and confident that he seems to be describing the world as it is today, rather than some distant, barely imaginable future. This is because his prediction falls out cleanly from the equations he’s proposed. Knowledge doubles every year, Kurzweil says. He has estimated the number of computations necessary to simulate a human brain. The rest is simple math.

But wait. There may be something wrong. Kurzweil’s theory of accelerating change is meant to be a universal law, applicable wherever intelligence is found. It’s fine to say that knowledge doubles every year. But then again, what is a year? A year is an astronomical artifact. It is the length of time required by Earth to make one orbit around our unexceptional star. A year is important to our nature, to our biology, to our fantasies and dreams. But it is a strange unit to discover in a general law.

“Doubling every year,” I say to Kurzweil, “makes your theory sound like a wish.”

He’s not thrown off. A year, he replies, is just shorthand. The real equation for accelerating world knowledge is much more complicated than that. (In his book, he gives it as: .)

He has examined the evidence, and welcomes debate on the minor details. If you accept his basic premise of accelerating growth, he’ll yield a little on the date he predicts the singularity will occur. After all, concede accelerating growth and the exponential fuse is lit. At the end you get that big bang: an explosion in intelligence that yields immortal life.

Despite all this, people continue to disbelieve. There is a lively discussion among experts about the validity of Moore’s law. Kurzweil pushes Moore’s law back to the dawn of time, and forward to the end of the universe. But many computer scientists and historians of technology wonder if it will last another decade. Some suspect that the acceleration of computing power has already slowed.

There are also philosophical objections. Kurzweil’s theory is that super-intelligent computers will necessarily be human, because they will be modeled on the human brain. But there are other types of intelligence in the world — for instance, the intelligence of ant colonies — that are alien to humanity. Grant that a computer, or a network of computers, might awaken. The consciousness of the this fabulous AI might remain as incomprehensible to us as we are to the protozoa.

Other pessimists point out that the brain is more than raw processing power. It also has a certain architecture, a certain design. It is attached to specific type of nervous system, it accepts only particular kinds of inputs. Even with better computational speed driving our thoughts, we might still be stuck in a kind of evolutionary dead end, incapable of radical self-improvement.

And these are the merely intellectual protests Kurzweil receives. The fundamental cause for loneliness, if you are a prophet of the singularity, is probably more profound. It stems from the simple fact that the idea is so strange. “Death has been a ubiquitous, ever-present facet of human society,” says Kurzweil’s friend Martine Rothblatt, founder of Sirius radio and chair of United Therapeutics, a biotech firm on whose board Kurzweil sits. “To tell people you are going to defeat death is like telling people you are going to travel back in time. It has never been done. I would be surprised if people had a positive reaction.”

To press his case, Kurzweil is writing and producing an autobiographical movie, with walk-ons by Alan Dershowitz and Tony Robbins. Kurzweil appears in two guises, as himself and as an intelligent computer named Ramona, played by an actress. Ramona has long been the inventor’s virtual alter ego and the expression of his most personal goals. “Women are more interesting than men,” he says, “and if it’s more interesting to be with a woman, it is probably more interesting to be a woman.” He hopes one day to bring Ramona to life, and to have genuine human experiences, both with her and as her. Kurzweil has been married for 32 years to his wife, Sonya Kurzweil. They have two children — one at Stanford University, one at Harvard Business School. “I don’t necessarily only want to be Ramona,” he says. “It’s not necessarily about gender confusion, it’s just about freedom to express yourself.”

Kurzweil’s movie offers a taste of the drama such a future will bring. Ramona is on a quest to attain full legal rights as a person. She agrees to take a Turing test, the classic proof of artificial intelligence, but although Ramona does her best to masquerade as human, she falls victim to one of the test’s subtle flaws: Humans have limited intelligence. A computer that appears too smart will fail just as definitively as one that seems too dumb. “She loses because she is too clever!” Kurzweil says.

The inventor’s sympathy with his robot heroine is heartfelt. “If you’re just very good at doing mathematical theorems and making stock market investments, you’re not going to pass the Turing test,” Kurzweil acknowledged in 2006 during a public debate with noted computer scientist David Gelernter. Kurzweil himself is brilliant at math, and pretty good at stock market investments. The great benefits of the singularity, for him, do not lie here. “Human emotion is really the cutting edge of human intelligence,” he says. “Being funny, expressing a loving sentiment — these are very complex behaviors.”

One day, sitting in his office overlooking the suburban parking lot, I ask Kurzweil if being a singularitarian makes him happy. “If you took a poll of primitive man, happiness would be getting a fire to light more easily,” he says. “But we’ve expanded our horizon, and that kind of happiness is now the wrong thing to focus on. Extending our knowledge and casting a wider net of consciousness is the purpose of life.” Kurzweil expects that the world will soon be entirely saturated by thought. Even the stones may compute, he says, within 200 years.

Every day he stays alive brings him closer to this climax in intelligence, and to the time when Ramona will be real. Kurzweil is a technical person, but his goal is not technical in this respect. Yes, he wants to become a robot. But the robots of his dreams are complex, funny, loving machines. They are as human as he hopes to be.

Gary Wolf (gary@aether.com) wrote about productivity guru David Allen in issue 15.10.

“Kalt? Kalt?” one of them calls out. The man gives a polite but vague answer, then turns and dives into the waves. After swimming back and forth in the 40-degree water for a few minutes, he emerges from the surf and jogs briefly along the shore. The wind is strong, but the man makes no move to get dressed. Passersby continue to comment and stare. “This is one of the reasons I prefer anonymity,” he tells me in English. “You do something even slightly out of the ordinary and it causes a sensation.”

Piotr Wozniak’s quest for anonymity has been successful. Nobody along this string of little beach resorts recognizes him as the inventor of a technique to turn people into geniuses. A portion of this technique, embodied in a software program called SuperMemo, has enthusiastic users around the world. They apply it mainly to learning languages, and it’s popular among people for whom fluency is a necessity — students from Poland or other poor countries aiming to score well enough on English-language exams to study abroad. A substantial number of them do not pay for it, and pirated copies are ubiquitous on software bulletin boards in China, where it competes with knockoffs like SugarMemo.

SuperMemo is based on the insight that there is an ideal moment to practice what you’ve learned. Practice too soon and you waste your time. Practice too late and you’ve forgotten the material and have to relearn it. The right time to practice is just at the moment you’re about to forget. Unfortunately, this moment is different for every person and each bit of information. Imagine a pile of thousands of flash cards. Somewhere in this pile are the ones you should be practicing right now. Which are they?

Fortunately, human forgetting follows a pattern. We forget exponentially. A graph of our likelihood of getting the correct answer on a quiz sweeps quickly downward over time and then levels off. This pattern has long been known to cognitive psychology, but it has been difficult to put to practical use. It’s too complex for us to employ with our naked brains.

Twenty years ago, Wozniak realized that computers could easily calculate the moment of forgetting if he could discover the right algorithm. SuperMemo is the result of his research. It predicts the future state of a person’s memory and schedules information reviews at the optimal time. The effect is striking. Users can seal huge quantities of vocabulary into their brains. But for Wozniak, 46, helping people learn a foreign language fast is just the tiniest part of his goal. As we plan the days, weeks, even years of our lives, he would have us rely not merely on our traditional sources of self-knowledge — introspection, intuition, and conscious thought — but also on something new: predictions about ourselves encoded in machines.

Given the chance to observe our behaviors, computers can run simulations, modeling different versions of our path through the world. By tuning these models for top performance, computers will give us rules to live by. They will be able to tell us when to wake, sleep, learn, and exercise; they will cue us to remember what we’ve read, help us track whom we’ve met, and remind us of our goals. Computers, in Wozniak’s scheme, will increase our intellectual capacity and enhance our rational self-control.

The reason the inventor of SuperMemo pursues extreme anonymity, asking me to conceal his exact location and shunning even casual recognition by users of his software, is not because he’s paranoid or a misanthrope but because he wants to avoid random interruptions to a long-running experiment he’s conducting on himself. Wozniak is a kind of algorithmic man. He’s exploring what it’s like to live in strict obedience to reason. On first encounter, he appears to be one of the happiest people I’ve ever met.

In the late 1800s, a German scientist named Hermann Ebbinghaus made up lists of nonsense syllables and measured how long it took to forget and then relearn them. (Here is an example of the type of list he used: bes dek fel gup huf jeik mek meun pon daus dor gim ke4k be4p bCn hes.) In experiments of breathtaking rigor and tedium, Ebbinghaus practiced and recited from memory 2.5 nonsense syllables a second, then rested for a bit and started again. Maintaining a pace of rote mental athleticism that all students of foreign verb conjugation will regard with awe, Ebbinghaus trained this way for more than a year. Then, to show that the results he was getting weren’t an accident, he repeated the entire set of experiments three years later. Finally, in 1885, he published a monograph called Memory: A Contribution to Experimental Psychology. The book became the founding classic of a new discipline.

Ebbinghaus discovered many lawlike regularities of mental life. He was the first to draw a learning curve. Among his original observations was an account of a strange phenomenon that would drive his successors half batty for the next century: the spacing effect.

Ebbinghaus showed that it’s possible to dramatically improve learning by correctly spacing practice sessions. On one level, this finding is trivial; all students have been warned not to cram. But the efficiencies created by precise spacing are so large, and the improvement in performance so predictable, that from nearly the moment Ebbinghaus described the spacing effect, psychologists have been urging educators to use it to accelerate human progress. After all, there is a tremendous amount of material we might want to know. Time is short.

How Supermemo Works
SuperMemo is a program that keeps track of discrete bits of information you’ve learned and want to retain. For example, say you’re studying Spanish. Your chance of recalling a given word when you need it declines over time according to a predictable pattern. SuperMemo tracks this so-called forgetting curve and reminds you to rehearse your knowledge when your chance of recalling it has dropped to, say, 90 percent. When you first learn a new vocabulary word, your chance of recalling it will drop quickly. But after SuperMemo reminds you of the word, the rate of forgetting levels out. The program tracks this new decline and waits longer to quiz you the next time.
How Supermemo Works

However, this technique never caught on. The spacing effect is “one of the most remarkable phenomena to emerge from laboratory research on learning,” the psychologist Frank Dempster wrote in 1988, at the beginning of a typically sad encomium published in American Psychologist under the title “The Spacing Effect: A Case Study in the Failure to Apply the Results of Psychological Research.” The sorrrowful tone is not hard to understand. How would computer scientists feel if people continued to use slide rules for engineering calculations? What if, centuries after the invention of spectacles, people still dealt with nearsightedness by holding things closer to their eyes? Psychologists who studied the spacing effect thought they possessed a solution to a problem that had frustrated humankind since before written language: how to remember what’s been learned. But instead, the spacing effect became a reminder of the impotence of laboratory psychology.

As a student at the Poznan University of Technology in western Poland in the 1980s, Wozniak was overwhelmed by the sheer number of things he was expected to learn. But that wasn’t his most troubling problem. He wasn’t just trying to pass his exams; he was trying to learn. He couldn’t help noticing that within a few months of completing a class, only a fraction of the knowledge he had so painfully acquired remained in his mind. Wozniak knew nothing of the spacing effect, but he knew that the methods at hand didn’t work.

The most important challenge was English. Wozniak refused to be satisfied with the broken, half-learned English that so many otherwise smart students were stuck with. So he created an analog database, with each entry consisting of a question and answer on a piece of paper. Every time he reviewed a word, phrase, or fact, he meticulously noted the date and marked whether he had forgotten it. At the end of the session, he tallied the number of remembered and forgotten items. By 1984, a century after Ebbinghaus finished his second series of experiments on nonsense syllables, Wozniak’s database contained 3,000 English words and phrases and 1,400 facts culled from biology, each with a complete repetition history. He was now prepared to ask himself an important question: How long would it take him to master the things he wanted to know?

The answer: too long. In fact, the answer was worse than too long. According to Wozniak’s first calculations, success was impossible. The problem wasn’t learning the material; it was retaining it. He found that 40 percent of his English vocabulary vanished over time. Sixty percent of his biology answers evaporated. Using some simple calculations, he figured out that with his normal method of study, it would require two hours of practice every day to learn and retain a modest English vocabulary of 15,000 words. For 30,000 words, Wozniak would need twice that time. This was impractical.

Wozniak’s discouraging numbers were roughly consistent with the results that Ebbinghaus had recorded in his own experiments and that have been confirmed by other psychologists in the decades since. If students nonetheless manage to become expert in a few of the things they study, it’s not because they retain the material from their lessons but because they specialize in a relatively narrow subfield where intense practice keeps their memory fresh. When it comes to language, the received wisdom is that immersion — usually amounting to actual immigration — is necessary to achieve fluency. On one hand, this is helpful advice. On the other hand, it’s an awful commentary on the value of countless classroom hours. Learning things is easy. But remembering them — this is where a certain hopelessness sets in.

As Wozniak later wrote in describing the failure of his early learning system: “The process of increasing the size of my databases gradually progressed at the cost of knowledge retention.” In other words, as his list grew, so did his forgetting. He was climbing a mountain of loose gravel and making less and less progress at each step.
Photo: Patrick Voigt

The problem of forgetting might not torment us so much if we could only convince ourselves that remembering isn’t important. Perhaps the things we learn — words, dates, formulas, historical and biographical details — don’t really matter. Facts can be looked up. That’s what the Internet is for. When it comes to learning, what really matters is how things fit together. We master the stories, the schemas, the frameworks, the paradigms; we rehearse the lingo; we swim in the episteme.

The disadvantage of this comforting notion is that it’s false. “The people who criticize memorization — how happy would they be to spell out every letter of every word they read?” asks Robert Bjork, chair of UCLA’s psychology department and one of the most eminent memory researchers. After all, Bjork notes, children learn to read whole words through intense practice, and every time we enter a new field we become children again. “You can’t escape memorization,” he says. “There is an initial process of learning the names of things. That’s a stage we all go through. It’s all the more important to go through it rapidly.” The human brain is a marvel of associative processing, but in order to make associations, data must be loaded into memory.

Once we drop the excuse that memorization is pointless, we’re left with an interesting mystery. Much of the information does remain in our memory, though we cannot recall it. “To this day,” Bjork says, “most people think about forgetting as decay, that memories are like footprints in the sand that gradually fade away. But that has been disproved by a lot of research. The memory appears to be gone because you can’t recall it, but we can prove that it’s still there. For instance, you can still recognize a ‘forgotten’ item in a group. Yes, without continued use, things become inaccessible. But they are not gone.”

After an ichthyologist named David Starr Jordan became the first president of Stanford University in the 1890s, he bequeathed to memory researchers one of their favorite truisms: Every time he learned the name of a student, Jordan is said to have complained, he forgot the name of a fish. But the fish to which Jordan had devoted his research life were still there, somewhere beneath the surface of consciousness. The difficulty was in catching them.

During the years that Wozniak struggled to master English, Bjork and his collaborator, Elizabeth Bjork (she is also a professor of psychology; the two have been married since 1969), were at work on a new theory of forgetting. Both were steeped in the history of laboratory research on memory, and one of their goals was to get to the bottom of the spacing effect. They were also curious about the paradoxical tendency of older memories to become stronger with the passage of time, while more recent memories faded. Their explanation involved an elegant model with deeply counterintuitive implications.

Long-term memory, the Bjorks said, can be characterized by two components, which they named retrieval strength and storage strength. Retrieval strength measures how likely you are to recall something right now, how close it is to the surface of your mind. Storage strength measures how deeply the memory is rooted. Some memories may have high storage strength but low retrieval strength. Take an old address or phone number. Try to think of it; you may feel that it’s gone. But a single reminder could be enough to restore it for months or years. Conversely, some memories have high retrieval strength but low storage strength. Perhaps you’ve recently been told the names of the children of a new acquaintance. At this moment they may be easily accessible, but they are likely to be utterly forgotten in a few days, and a single repetition a month from now won’t do much to strengthen them at all.

The Bjorks were not the first psychologists to make this distinction, but they and a series of collaborators used a broad range of experimental data to show how these laws of memory wreak havoc on students and teachers. One of the problems is that the amount of storage strength you gain from practice is inversely correlated with the current retrieval strength. In other words, the harder you have to work to get the right answer, the more the answer is sealed in memory. Precisely those things that seem to signal we’re learning well — easy performance on drills, fluency during a lesson, even the subjective feeling that we know something — are misleading when it comes to predicting whether we will remember it in the future. “The most motivated and innovative teachers, to the extent they take current performance as their guide, are going to do the wrong things,” Robert Bjork says. “It’s almost sinister.”

The most popular learning systems sold today — for instance, foreign language software like Rosetta Stone — cheerfully defy every one of the psychologists’ warnings. With its constant feedback and easily accessible clues, Rosetta Stone brilliantly creates a sensation of progress. “Go to Amazon and look at the reviews,” says Greg Keim, Rosetta Stone’s CTO, when I ask him what evidence he has that people are really remembering what they learn. “That is as objective as you can get in terms of a user’s sense of achievement.” The sole problem here, from the psychologists’ perspective, is that the user’s sense of achievement is exactly what we should most distrust.

The battle between lab-tested techniques and conventional pedagogy went on for decades, and it’s fair to say that the psychologists lost. All those studies of human memory in the lab — using nonsense syllables, random numbers, pictures, maps, foreign vocabulary, scattered dots — had so little influence on actual practice that eventually their irrelevance provoked a revolt. In the late ’70s, Ulric Neisser, the pioneering researcher who coined the term cognitive psychology, launched a broad attack on the approach of Ebbinghaus and his scientific kin.

“We have established firm empirical generalizations, but most of them are so obvious that every 10-year-old knows them anyway,” Neisser complained. “We have an intellectually impressive group of theories, but history offers little confidence that they will provide any meaningful insight into natural behavior.” Neisser encouraged psychologists to leave their labs and study memory in its natural environment, in the style of ecologists. He didn’t doubt that the laboratory theories were correct in their limited way, but he wanted results that had power to change the world.

Many psychologists followed Neisser. But others stuck to their laboratory methods. The spacing effect was one of the proudest lab-derived discoveries, and it was interesting precisely because it was not obvious, even to professional teachers. The same year that Neisser revolted, Robert Bjork, working with Thomas Landauer of Bell Labs, published the results of two experiments involving nearly 700 undergraduate students. Landauer and Bjork were looking for the optimal moment to rehearse something so that it would later be remembered. Their results were impressive: The best time to study something is at the moment you are about to forget it. And yet — as Neisser might have predicted — that insight was useless in the real world. Determining the precise moment of forgetting is essentially impossible in day-to-day life.

Obviously, computers were the answer, and the idea of using them was occasionally suggested, starting in the 1960s. But except for experimental software, nothing was built. The psychologists were interested mainly in theories and models. The teachers were interested in immediate signs of success. The students were cramming to pass their exams. The payoff for genuine progress was somehow too abstract, too delayed, to feed back into the system in a useful way. What was needed was not an academic psychologist but a tinkerer, somebody with a lot of time on his hands, a talent for mathematics, and a strangely literal temperament that made him think he should actually recall the things he learned.

The day I first meet Wozniak, we go for a 7-mile walk down a windy beach. I’m in my business clothes and half comatose from jet lag; he’s wearing a track suit and comes toward me with a gait so buoyant he seems about to take to the air. He asks me to walk on the side away from the water. “People say that when I get excited I tend to drift in their direction, so it is better that I stand closer to the sea so I don’t push you in,” he says.

Wozniak takes an almost physical pleasure in reason. He loves to discuss things with people, to get insight into their personalities, and to give them advice — especially in English. One of his most heartfelt wishes is that the world have one language and one currency so this could all be handled more efficiently. He’s appalled that Poland is still not in the Eurozone. He’s baffled that Americans do not use the metric system. For two years he kept a diary in Esperanto.

Although Esperanto was the ideal expression of his universalist dreams, English is the leading real-world implementation. Though he has never set foot in an English-speaking country, he speaks the language fluently. “Two words that used to give me trouble are perspicuous and perspicacious,” he confessed as we drank beer with raspberry syrup at a tiny beachside restaurant where we were the only customers. “Then I found a mnemonic to enter in SuperMemo: clear/clever. Now I never misuse them.”

Wozniak’s command of English is the result of a series of heroic experiments, in the tradition of Ebbinghaus. They involved relentless sessions of careful self-analysis, tracked over years. He began with the basic conundrum of too much to study in too little time. His first solution was based on folk wisdom. “It is a common intuition,” Wozniak later wrote, “that with successive repetitions, knowledge should gradually become more durable and require less frequent review.”

This insight had already been proven by Landauer and Bjork, but Wozniak was unaware of their theory of forgetting or of any of the landmark studies in laboratory research on memory. This ignorance was probably a blessing, because it forced him to rely on pragmatic engineering. In 1985, he divided his database into three equal sets and created schedules for studying each of them. One of the sets he studied every five days, another every 18 days, and the third at expanding intervals, increasing the period between study sessions each time he got the answers right.

This experiment proved that Wozniak’s first hunch was too simple. On none of the tests did his recall show significant improvement over the naive methods of study he normally used. But he was not discouraged and continued making ever more elaborate investigations of study intervals, changing the second interval to two days, then four days, then six days, and so on. Then he changed the third interval, then the fourth, and continued to test and measure, measure and test, for nearly a decade. His conviction that forgetting could be tamed by following rules gave him the intellectual fortitude to continue searching for those rules. He doggedly traced a matrix of paths, like a man pacing off steps in a forest where he is lost.

All of his early work was done on paper. In the computer science department at the Poznan University of Technology, “we had a single mainframe of Polish-Russian design, with punch cards,” Wozniak recalls. “If you could stand in line long enough to get your cards punched, you could wait a couple of days more for the machine to run your cards, and then at last you got a printout, which was your output.”

The personal computer revolution was already pretty far along in the US by the time Wozniak managed to get his hands on an Amstrad PC 1512, imported through quasi-legal means from Hamburg, Germany. With this he was able to make another major advance in SuperMemo — computing the difficulty of any fact or study item and adjusting the unique shape of the predicted forgetting curve for every item and user. A friend of Wozniak’s adapted his software to run on Atari machines, and as access to personal computers finally spread among students, so did SuperMemo.

After the collapse of Polish communism, Wozniak and some fellow students formed a company, SuperMemo World. By 1995, their program was one of the most successful applications developed by the country’s fledgling software industry, and they were searching for funding that would allow them to relocate to Silicon Valley. That year, at Comdex in Las Vegas, 200,000 people got a look at Sony’s new DVD technology, prototypes of flatscreens, and Wozniak’s SuperMemo, which became the first Polish product shown at the great geek carnival, then at the height of its influence. In Europe, the old communist experiment in human optimization had run its course. Wozniak believed that in a world of open competition, where individuals are rewarded on merit, a scientific tool that accelerated learning would find customers everywhere.

Wozniak’s chief partner in the campaign to reprogram the world’s approach to learning through SuperMemo was Krzysztof Biedalak, who had been his classmate at the University of Technology. The two men used to run 6 miles to a nearby lake for an icy swim. Biedalak agrees with Wozniak that winter swimming is good for mental health. Biedalak also agrees with Wozniak that SuperMemo produces extreme learning. But Biedalak does not agree with Wozniak about everything. “I don’t apply his whole technique,” he says. “In my context, his technique is inapplicable.”

What Biedalak means by Wozniak’s technique is the extension of algorithmic optimization to all dimensions of life. Biedalak is CEO of SuperMemo World, which sells and licenses Wozniak’s invention. Today, SuperMemo World employs just 25 people. The venture capital never came through, and the company never moved to California. About 50,000 copies of SuperMemo were sold in 2006, most for less than $30. Many more are thought to have been pirated.

Biedalak and I meet and talk in a restaurant in downtown Warsaw where the shelves are covered in gingham and the walls are lined with jars of pickled vegetables. He has an intelligent, somewhat hangdog expression, like a young Walter Matthau, and his tone is as measured as Wozniak’s is impulsive. Until I let the information slip, he doesn’t even know the exact location of his partner and friend.

“Piotr would never go out to promote the product, wouldn’t talk to journalists, very rarely agreed to meet with somebody,” Biedalak says. “He was the driving force, but at some point I had to accept that you cannot communicate with him in the way you can with other people.”

The problem wasn’t shyness but the same intolerance for inefficient expenditure of mental resources that led to the invention of SuperMemo in the first place. By the mid-’90s, with SuperMemo growing more and more popular, Wozniak felt that his ability to rationally control his life was slipping away. “There were 80 phone calls per day to handle. There was no time for learning, no time for programming, no time for sleep,” he recalls. In 1994, he disappeared for two weeks, leaving no information about where he was. The next year he was gone for 100 days. Each year, he has increased his time away. He doesn’t own a phone. He ignores his email for months at a time. And though he holds a PhD and has published in academic journals, he never attends conferences or scientific meetings.

Instead, Wozniak has ridden SuperMemo into uncharted regions of self-experimentation. In 1999, he started making a detailed record of his hours of sleep, and now he’s working to correlate that data with his daily performance on study repetitions. Psychologists have long believed there’s a correlation between sleep and memory, but no mathematical law has been discovered. Wozniak has also invented a way to apply his learning system to his intake of unstructured information from books and articles, winnowing written material down to the type of discrete chunks that can be memorized, and then scheduling them for efficient learning. He selects a short section of what he’s reading and copies it into the SuperMemo application, which predicts when he’ll want to read it again so it sticks in his mind. He cuts and pastes completely unread material into the system, assigning it a priority. SuperMemo shuffles all his potential knowledge into a queue and presents it to him on a study screen when the time is right. Wozniak can look at a graph of what he’s got lined up to learn and adjust the priority rankings if his goals change.

These techniques are designed to overcome steep learning curves through automated steps, like stairs on a hill. He calls it incremental reading, and it has come to dominate his intellectual life. Wozniak no longer wastes time worrying that he hasn’t gotten to some article he wants to read; once it’s loaded into the system, he trusts his algorithm to apportion it to his consciousness at the appropriate time.

The appropriate time, that is, for him. Having turned over his mental life to a computerized system, he refuses to be pushed around by random inputs and requests. Naturally, this can be annoying to people whose messages tend to sift to the bottom. “After four months,” Biedalak says sadly, “you sometimes get a reply to some sentence in an email that has been scrambled in his incremental reading process.”

For Wozniak, these misfires were less a product of scrambling than of an inevitable clash of goals. A person who understands the exact relationship between learning and time is forced to measure out his hours with a certain care. SuperMemo was like a genie that granted Wozniak a wish: unprecedented power to remember. But the value of what he remembered depended crucially on what he studied, and what he studied depended on his goals, and the selection of his goals rested upon the efficient acquisition of knowledge, in a regressive function that propelled him relentlessly along the path he had chosen. The guarantee that he would not forget what he learned was both a gift and a demand, requiring him to sacrifice every extraneous thing.

From the business side of SuperMemo, Wozniak’s priorities can sometimes look selfish. Janusz Murakowski, one of Wozniak’s friends who worked as a manager at the company during its infancy, thinks that Wozniak’s focus on his own learning has stunted the development of his invention. “Piotr writes this software for himself,” says Murakowski, now a professor of electrical engineering at the University of Delaware. “The interface is just impossible.” This is perhaps a bit unfair. SuperMemo comes in eight flavors, some of which were coded by licensees: SuperMemo for Windows, for Palm devices, for several cell phones, even an Internet version. It’s true that Wozniak is no Steve Jobs, and his software has none of the viral friendliness of a casual game like Brain Age for Nintendo DS. Still, it can hardly be described as the world’s most difficult program. After all, photographers can learn to produce the most arcane effects in Photoshop. Why shouldn’t more people be able to master SuperMemo?

“It was never a feel-good product,” Murakowski says, and here he may be getting closer to the true conflict that lies at the heart of the struggle to optimize intelligence, a conflict that transcends design and touches on some curious facts about human nature. We are used to the idea that normal humans can perform challenging feats of athleticism. We all know someone who has run a marathon or ridden a bike cross-country. But getting significantly smarter — that seems to be different. We associate intelligence with pure talent, and academic learning with educational experiences dating far back in life. To master a difficult language, to become expert in a technical field, to make a scientific contribution in a new area — these seem like rare things. And so they are, but perhaps not for the reason we assume.

The failure of SuperMemo to transform learning uncannily repeats the earlier failures of cognitive psychology to influence teachers and students. Our capacity to learn is amazingly large. But optimal learning demands a kind of rational control over ourselves that does not come easily. Even the basic demand for regularity can be daunting. If you skip a few days, the spacing effect, with its steady march of sealing knowledge in memory, begins to lose its force. Progress limps. When it comes to increasing intelligence, our brain is up to the task and our technology is up to the task. The problem lies in our temperament.

The Baltic Sea is dark as an unlit mirror. Wozniak and I walk along the shore, passing the wooden snack stands that won’t be open until spring, and he tells me how he manages his life. He’s married, and his wife shares his lifestyle. They swim together in winter, and though Polish is their native language, they communicate in English, which she learned with SuperMemo. Wozniak’s days are blocked into distinct periods: a creative period, a reading and studying period, an exercise period, an eating period, a resting period, and then a second creative period. He doesn’t get up at a regular hour and is passionate against alarm clocks. If excitement over his research leads him to work into the night, he simply shifts to sleeping in the day. When he sits down for a session of incremental reading, he attends to whatever automatically appears on his computer screen, stopping the instant his mind begins to drift or his comprehension falls too low and then moving on to the next item in the queue. SuperMemo graphs a distribution of priorities that he can adjust as he goes. When he encounters a passage that he thinks he’ll need to remember, he marks it; then it goes into a pattern of spaced repetition, and the information it contains will stay in his brain indefinitely.

“Once you get the snippets you need,” Wozniak says, “your books disappear. They gradually evaporate. They have been translated into knowledge.”

As a science fiction fan, I had always assumed that when computers supplemented our intelligence, it would be because we outsourced some of our memory to them. We would ask questions, and our machines would give oracular — or supremely practical — replies. Wozniak has discovered a different route. When he entrusts his mental life to a machine, it is not to throw off the burden of thought but to make his mind more swift. Extreme knowledge is not something for which he programs a computer but for which his computer is programming him.

I’ve already told Wozniak that I am not optimistic about my ability to tame old reading habits in the name of optimized knowledge. Books, for me, are not merely sources of information I might want to load into memory but also subjective companions, almost substitute people, and I don’t see why I would want to hold on to them in fragments. Still, I tell him I would like to give it a shot.

“So you believe in trying things for yourself?” he asks.

“Yes.”

This provides his opening. “In that case, let’s go swimming.”

At the edge of the sea, I become afraid. I’m a strong swimmer, but there’s something about standing on the beach in the type of minuscule bathing suit you get at the gift shop of a discount resort in Eastern Europe, and watching people stride past in their down parkas, that smacks of danger.

“I’m already happy with anticipation,” Wozniak says.

“Will I have a heart attack?”

“There is less risk than on your drive here,” he answers.

I realize he must be correct. Poland has few freeways, and in the rural north, lines of cars jockey behind communist-era farm machinery until they defy the odds and try to pass. There are spectacular wrecks. Wozniak gives close attention to the qualitative estimate of fatal risks. By graphing the acquisition of knowledge in SuperMemo, he has realized that in a single lifetime one can acquire only a few million new items. This is the absolute limit on intellectual achievement defined by death. So he guards his health. He rarely gets in a car. The Germans on the beach are staring at me. I dive in.

Philosopher William James once wrote that mental life is controlled by noticing. Climbing out of the sea and onto the windy beach, my skin purple and my mind in a reverie provoked by shock, I find myself thinking of a checklist Wozniak wrote a few years ago describing how to become a genius. His advice was straightforward yet strangely terrible: You must clarify your goals, gain knowledge through spaced repetition, preserve health, work steadily, minimize stress, refuse interruption, and never resist sleep when tired. This should lead to radically improved intelligence and creativity. The only cost: turning your back on every convention of social life. It is a severe prescription. And yet now, as I grin broadly and wave to the gawkers, it occurs to me that the cold rationality of his approach may be only a surface feature and that, when linked to genuine rewards, even the chilliest of systems can have a certain visceral appeal. By projecting the achievement of extreme memory back along the forgetting curve, by provably linking the distant future — when we will know so much — to the few minutes we devote to studying today, Wozniak has found a way to condition his temperament along with his memory. He is making the future noticeable. He is trying not just to learn many things but to warm the process of learning itself with a draft of utopian ecstasy.

[First published in Wired 16.05, 4.21.08]