November 19, 2011

The benefits of 7 billion people

How many of our technologies and inventions are a result of a large population?

You don't spend $10M on a drug that might cure 50 people...

November 18, 2011

Connections Matter

The Algorithms of Wall Street
On May 6, 2010, the Dow Jones Industrial Average fell nearly 1000 points - the largest single-day drop in history - only to regain nearly everything that was lost within minutes.

So what happened? Was there some sort of dramatic news that was reported, and then recanted? A war that almost happened? An election reversed? These are the sorts of dramatic events that seem necessary for such an extreme trading pattern.

In fact, there was no such root cause. While the details are still sketchy, one theory for the cause of the "Flash Crash" is that a single large sell order triggered high-frequency trading algorithms (i.e., computer-directed trades) to start selling. The algorithms were optimized to deal with small changes in price, and as prices started plummeting, the computers continued to trade with each other, driving prices down even further. Where a professional trader could tell that prices were too low, there were no people in the loop - it was just computers buying and selling to each other.

While this is a dramatic example of how eliminating people from a process can prove disastrous, it highlights just how important computer-to-computer interactions are. In fact, I believe that computer-to-computer communication will turn out to be one of a (very) long string of "Connection Revolutions", which I would like to briefly outline. As usual, I will blur the line between biological and technological advances, because I see them as being very similar.

I want to quickly outline each of these revolutions, and possibly follow up with a blog post about each one.

Connection Revolutions

Environment Sensing
Very early in the history of life, organisms gained the ability to sense their environment, and respond to it. These first senses were incredibly rudimentary, and chemical in nature. An organism that recognizes the chemicals that occur near its food source, for example, will be more successful in finding food, and will pass that trait on to its offspring.

Eventually, organisms not only sensed the environment, but intentionally altered the environment by emitting chemicals, thus communicating with other organisms. These communication channels can be surprisingly complex, allowing even one-celled bacteria to determine important information about their environment, and to coordinate their actions.

This coordination allowed some organisms to become so closely linked that they could specialize, subdivide their tasks, and become multicellular. At first, their larger size allowed them to become bigger, and thus move up the food chain. In time, specialization and intimate intercellular communication made specialized organs possible, allowing bodies to grow much larger, and much, much more complex.

One of those organs was the brain, composed of cells whose only job is communication. Brains take in stimuli collected from all over the body, process it, and send instructions for how to react. By centralizing this information processing, relations between stimuli could be discovered and responded to.

Some animals learned to signal directly, chemically, visually (think peacocks feathers or flowers), and most importantly for our timeline, aurally. Some animals learned to produce and sense sounds. In the case of humans, our supersized brains allowed us to progress beyond simple grunts and communicate whatever we wanted.

This allowed people to communicate about things other than the immediate present. Concepts and ideas could be discussed and examined. Instructions about how to hunt, plant crops, etc. could be relayed unambiguously. It is possible that language is what enabled our ancestors great migration out of Africa.

Writing allowed language to be conveyed not just to those immediately present, but across time and space. Vastly more knowledge could be preserved and distributed, making complex societies and technologies possible.

Printing Press
The printing press seems like an incremental improvement to hand-written writing, but its effects were much more important than the technological improvement would predict. By making books financially accessible to everyone, I believe the printing press was a necessary condition for universal literacy, democracy, the Renaissance, and the Industrial Revolution. By making it possible to (nearly) perfectly duplicate even seemingly unimportant books, knowledge about disparate subjects could be gathered, assimilated, and serve as inputs for further knowledge and technology creation.

Mass Media
The rise of mass media (i.e., TV, radio) was another important change. For the first time, media could be produced and distributed simultaneously to entire nations. This had important impact in creating a shared culture, and a shared way of seeing the world. Its masters held great power, and some, like Joseph Goebbels, understood just how revolutionary and powerful it could be.

The Internet is the next great communication revolution. I definitely want to write another post about this, but I want to outline a few of the important trends that I see from the Internet.

First, the ability to communicate one-to-one with anyone else, immediately. Every person is connected to every other person, and theoretically could communicate with any of them, without intermediaries.

Second, we are storing even unimportant things, digitally, where they can be preserved and searched forever. By aggregating seemingly unimportant things (e.g., which web pages link to which other web pages), amazing things can be built and discovered (e.g., Google).

Third, the Internet allows our things to talk to each other. Never before have our tools been able to tell us, or each other, what they need. We are more and more able to craft an environment that understands us, and can interact meaningfully with us. This has some dangers, like the "Flash Crash", but I believe that the benefits will become more and more apparent.

These changes are incredibly important, and I believe that we are like the readers of the first printed books, seeing a cool new technology, but blind to the societal changes (e.g., the Renaissance, democracy, and the Industrial Revolution) waiting in the wings.

October 18, 2011

Latent Celebrity

In the past, society only had the resources to keep track of the currently pertinent. Books, magazines, and newspapers have limited space, and limited resources, so they cover only (what they consider to be) important things and important people.

Now, because so much of our lives are digital, we can store everything, and decide later what is important. Obscure action by obscure people can be stored, and later gain relevance when either the action or the person turns out to have been important, in retrospect. It's as though the data is just waiting for its importance to be revealed - a sort of latent celebrity.

For example, in 1991 a Master's student in Finland was working on a hobby operating system, and asked if anyone wanted to pitch in. The project eventually became Linux. Because the conversation was held on a digital mailing list, we can now study the founding of Linux using the original sources. That is really amazing.

Perhaps just as importantly, there is even a record of similar projects that failed. When publishing is a bottleneck, there is a bias toward focusing on the positive. By having full data on the winners and the losers, we can do much better at finding the source of the success (which might be just random).

There is a dark side to this "latent celebrity" - this storing of everything. As a silly example, digital search and transmission makes it easy for sites like to publicize obscure idiocy for the world to see.

I believe that we are just beginning to explore the implications of a digital culture.

Immortal Artifacts

I found this video very humorous, and probably a very apt description of how our understanding of the past compares to reality.

However, the same lessons don't extrapolate into the future. We are leaving digital artifacts, which are (potentially) immortal. It's very possible that a historian in 3000 A.D. could listen to all of the Beatles' recordings, with exactly the same quality that we can listen to them now. He could watch their movies, and read newspaper articles about them.

Part of what makes another culture feel old or distant is the fact that aging is a lossy process - their paintings fade, their buildings collapse, and their books are lost. How will things change when our artifacts (e.g., the web) can be perfectly preserved indefinitely?

October 14, 2011

The Second Digital Revolution

Image by dullhunk
DNA provides a great distinction between data and medium. While our DNA (the medium) dies with us, our genes (the data) can be passed  down to our posterity. In fact, some of our genes may be billions of years old. This immortality of genes stems from the fact that they are digital. Genes happen to be encoded in DNA, but they are just base-4 information.

Up until very recently, human beings only created analog artifacts. A painting lasts only as long as the canvas (or the cave) it was painted on. They can be reproduced, but  only imperfectly and at great expense.

Books are one step closer to being digital. The English alphabet is, in a sense, a base-26 code. The problem with books is that it is difficult to replicate them.

With computers, we have created a system like DNA - digital information plus a way to replicate that information quickly and faithfully. The first time that sort of system came into being, the output was all of the life on earth. It will be interesting to see what the results are of this second digital revolution.

February 4, 2011

We are almost certainly wrong

As individuals, and as societies, we rightfully dismiss a bunch of crazy theories. However, we can't dismiss them with 100% confidence. We are pretty sure that the moon landing wasn't faked, that UFOs aren't stashed at Roswell, that we are not part of a computer simulation (à la Matrix), etc. Let's say there are 100 of these crackpot theories, and that we are 95% certain that each one is incorrect. That seems pretty good, right? We are 95% sure that all of those theories are incorrect, right?

This is where things get strange. Even though we are 95% sure that each of the crackpot theories is incorrect, we shouldn't be 95% confident that all of the theories are wrong. If there is a 5% chance that each one is correct, then the chance that all 100 theories are incorrect is actually only 0.6%! In other words, at least one of the theories is almost certainly correct, and probably quite a few of them are correct.

So does this mean that you should start believing crackpot theories? No - and that's the craziest part of all. The key is that we don't know which of our beliefs are wrong. It is rational to disbelieve all of the crackpot theories and believe that some of your beliefs are wrong.