The capabilities of the car exceed most human counterparts, and Wired magazine describes Google's car as "the poster child for self-driving cars." It uses a series of sensors and pattern-recognition algorithms in an on-board computer to navigate an endless sea of possible driving situations.
Sunday, October 7, 2012
Getting Around With Google
– Larry Page, Co-founder, Google incorporated.
Google’s has certainly established itself as a powerhouse in the modern world. With budgets greater than many nation states, and a consistent growth margin of 20-30% annually, Google is demonstrating the resurgence of American multinationals. But that is not to say that we can liken Google to some of the more monstrous corporations, like Haliburton or BP.
But it isn't Google's financial status that makes them so remarkable. They are the gatekeepers of information, and as such are presented with a profound responsibility to human civilization. While this is a post for another day, it illustrates just how powerful Google is becoming. But unlike some other nefarious corporations, Google uses its weight to influence positive change and to provide financial support for businesses who have "a healthy disregard for the impossible," investing in areas as diverse as photovoltaics, genomics, articificial intelligence, and biotechnology. The subject of this post is google's development of a self-driving car. The car has been tested over hundreds of thousands of miles since 2010 when the project was initially announced. Now, the car continues on its meteoric rise, from technological innovation to legislative support. It is now legal in Nevada and California, and when it was signed into action in California by Governor Jerry Brown, he commented that the bill was “turning today’s science fiction into tomorrow’s reality.”
However, it is not the subject of this article to describe the car's current status, but its potential implications for the future. It is not a stretch of the imagination that self-driving cars would be prevalent on the streets of a post-modern society. Science fiction has long anticipated a mechanized society, as evidenced in the Jetsons. The question becomes, how close are we to mass-produced cars of this effect?
The answer is somewhat disheartening when taken on its own. The current cost of Google's self-driving car can be estimated at nearly $300, 000. This is due to the high cost of the very sensitive sensors required to perform necessary measurements. So, what is the solution? There are several that have been proposed, including using cheaper sensors and better software algorithms; subsidies from insurance companies (lower rates because there are no accidents); and my own simple statement of the obvious, which is that computers get more powerful and cheaper over time.
Now that the state of current development has been delineated, it is important to consider the economic, cultural, and technological ramifications. I'll outline them briefly:
Moore's Law dictates that over time computation doubles in power and deflates in price. Therefore, a $1000 laptop a year from now will be twice as powerful as a current $1000 model. For those who have ever seen a graph of an exponential, its pretty obvious that computation is slated to be the linchpin of the modern world's survival and affluence. We will likely see ubiquity of computation as a result of this process, and simply stated, the more time passes, the more economically viable autonomous (anything) becomes.
Further, as more and more self-driving cars are implemented, fewer accidents will occur. This spells fewer costs for consumers, and fewer costs for insurance companies (and as I said above, a subsidy might make the insurance for a Google car very inexpensive). Additionally, because the cars could be coordinated synchronously, all stop-and-go traffic will be utterly negated. That is, until some old lady in a Lincoln merges and causes thousands of dollars in lost fuel efficiency and immeasurable frustration.
Envision: you and your friends have bought tickets to a concert. You're looking forward to a 2 hour drive to the respective city, as well as a ride to the concert, to a bar, to a hotel, or whatever the case may be. One could spend hundreds on cabs while also having to commute. Instead, imagine going in a self-driving car - where drinking and driving isn't taboo - its fashionable, and safe. Safer, in fact, than if you were driving sober. So, crack a beer in the back and the car does all the work. It could even be programmed to pick you up after the concert and take you around the city like a chauffeur. Currently, of course, even where these cars are legal, it is required to have a capable driver in the seat. But as technical advancements and legislative efforts continue, we may see that change.
Also, the creation of the self-driving car is a hallmark of futurism. It is an important symbol which inspires hope, not unlike when Kennedy was president, and described the ascent of mankind into the ether.
As these cars get cheaper, new luxury items will be implemented. I envision limo-type cars running in "autonomous carpool lanes" running from city to city. One could concievably start a commuting business that relies solely on autonomous transportation.
In the History of Science, many discoveries have been made peripherally - meaning that the intention of a particular body of work or expiermient may yield unexpected results, whether positive or negative. Good examples include Faraday's work, Turing's papers on Morphogenesis, belousov's oscillating chemicals, and thousands more. So what does this have to do with Google's car?
First it is necessary to determine what elements of technology are required. More specifically, its important to recognize what aspects of artificial intelligence that this project must employ. Indeed, replacing a human driver is no simple measure. Humans are capable of highly parallel processing, able to perform many tasks at once while amalgamating millions of bytes of data every second. Google's work (in general) on pattern recognition algorithms has allowed computers to recognize cats, dogs, faces, and other objects with consistency, simply by looking at thousands of pictures of each of those things. As the demand for this product increases, the field of AI will advance to serve the economic demand, while also satisfying the Manifest Destiny of modern computer science: the turing machine. Pattern recognition is a crucial element of humanity. What is a sonata? What is a Van Gogh painting? A Shakespeare sonnet? All of these are particular organizations of bits - fundamentally super-tiny bits of information which comprise words, notes, colours and shapes. The conclusion therefore is that the advancement of the self-driving car will eventually lead to smarter computers, possibly capable of creating beautiful works of art.
It doesn't take a drunk driver killing someone you love in order to realize the potential of this technology; but for me personally, that holds an element in my desire to see this technology come to pass. I know that if self-driving cars had been ubiquitous, my brother would be alive. It comes down to human beings. There is a certain humanistic exaggeration which has lingered since the renaissance. We have been in the process of discovering our own tumultuous natures for thousands of years. Yet now, at the precipice of truly understanding ourselves, there is a refusal of acceptance that certain elements of humanity are negative, and others positive. While I don't necessarily advocate genetic engineering to remove these traits, I also think its irresponsible to ignore them. In this respect, self-driving cars are a moral imperative. How many will die this year in collisions? How many lives destroyed? What peripheral suffering will continue to echo in the world while we hesitate, while we vaguely acknowledge our volatile natures and yet refuse to act on that knowledge?