The prediction of the tsunami in Japan - and the ``Atid Mokelt'' company

After the strong earthquake and tsunami in Japan, I received three calls from three different sources, and they all asked the same question: Roy, how did you not foresee the tsunami in Japan? And everyone got the same answer from me: because we didn't get enough funding 

After the strong earthquake and tsunami in Japan, I received three calls from three different sources, and they all asked the same question: Roy, how did you not foresee the tsunami in Japan? And everyone got the same answer from me: because we didn't get enough funding.

But if we are serious, we must admit that there will always be events that are very difficult to predict ahead of time. Still, even one day's notice could have saved hundreds of lives in the case of the recent tsunami. This is the power of predicting the future, and it is not limited to earthquakes and natural disasters.

Can we ever achieve technology sophisticated enough to fully predict the future?

well no. not fully. But a company called Recorded Future is currently making one of the most impressive attempts in the field.

The concept of a recorded future is quite simple. They believe that almost all current and contemporary human information is circulating on the web. This belief has a serious basis. As of February 2011, there are over 156 million public blogs across the Internet, 'personal' diaries open to any viewer. Even if we assume that most of them are closed and dreary, this is still almost a whole percentage of the population of the earth that uploads its thoughts and experiences to the Internet in a non-stop manner. Facebook managed to acquire 600 million enthusiastic participants, and almost fifty percent of the population of the United States has a Facebook account. Many of the participants record every detail of their lives, every emotion and every thought, on their private walls.

What if we could unify and cross-reference all this information and derive meaning from it?

I already told you in a previous blog article about a study that tried to measure the sentiments of Twitter users, which found that the calmer the users (by tracking words that indicated calmness in their tweets), the less likely the Dow Jones index was to fall. What if we could visit all the messages from Twitter, Facebook, blogs, the press releases of large companies, the guesses of speculators, the statements of policymakers - and combine all of these into one big statistical forecast about the coming days and years to come?

A recorded future is trying to do this almost impossible task. The company, which was founded less than two years ago, created a sophisticated algorithm that scans (as of now) news websites, blogs, public information sources, government websites, financial databases and more. He extracts information from the text about future events that are expected to happen and when they will happen, and tries to create a future picture of the following years based on the information he receives.

So, for example, if you want to know what the Toyota car market in Israel will be like in the coming years, you can enter your query (Toyota cars in Israel) into the search engine and get graphs that will show you the production plans for the new Toyota cars according to the company, the forecasts of its managers for penetration for the market in Israel, the statements of the Ministry of Finance on the tax that will be imposed on the import of Toyota cars in Israel starting from 2012, and so on.

This is a limited type of forecasting that may fail in many cases. Most likely he would not have predicted, for example, the great economic crisis of recent years, for the simple reason that almost no one thought about it or wrote about it in advance. The few who wrote were swallowed up in the vast text mass of the enthusiastic majority. On the other hand, the power of an algorithm that can cross-reference many clues and reach a conclusion that relies on messages coming from all over the world incessantly cannot be underestimated.

And what does all this have to do with the tsunami in Japan? Well, most natural disasters don't happen in seconds. The source of the tsunami was an earthquake that occurred in the ocean at a distance of 400 kilometers from the city of Tokyo. Already today, there are seismic sensors located in the sea and on land that are supposed to warn of small earthquakes and subtle hints that predict the arrival of more massive earthquakes. There are satellites in space that monitor sea level. Even humans can quickly spread the information about unusual events in their environment, as evidenced by the case where an earthquake started sixty kilometers from San Francisco, and Twitter users in the area sent warning tweets to San Francisco, which reached the city a few seconds before the earthquake itself reached her.

The more the means of monitoring and imaging expand, and the more the information network expands to include them in its forecasts for the coming minutes and days, the closer we will get to a reliable forecasting ability in a wider variety of fields. It is not impossible that a recorded future company will lead the trend in creating such a global algorithm.

Today, the Recorded Future algorithm is only able to provide the data from all over the world and display it graphically and cannot yet formulate its own conclusions. This ability remains, for the time being, within the limits of the skull of the human brain only. At the same time, it is possible to think of a future in the coming decades, in which the ability of computers to link billions of different items of information will exceed that of humans, and will reach the point of creating a kind of artificial god. That he can predict most world events every second by second - and if necessary also influence them.

Anyone want to start worshiping Google?

http://www.youtube.com/watch?v=fm3BwqueTAk