- Get link
- X
- Other Apps
Google has always been for artificial intelligence, so no one was surprised when Ray Kurzweil , one of the leading experts in this field, joined the company. It happened at the end of last year. Nevertheless, many workers of the search giant eyebrows crawled up when Kurzweil said he could create an artificial mind.
Add to all this the fact that Jeffrey Hinton, the “godfather” of computer neural networks, became the next employee of the search engine. It seems that Google has acquired the most courageous developers of artificial intelligence. Someone finds it exciting, someone - anxious. Someone - both.
On Tuesday, Kurzweil made a small gathering at Google, timed to coincide with Will Smith’s upcoming film “After Our Era” (After Earth), apparently linking the futuristic concept of the film with real futurism. The discussion raised questions about the need for space travel and the resolution of global energy problems with the help of solar energy. After all the participants dispersed, Kurzweil gave a short interview to Wired magazine, which we present to you.
Wired : The “google” party has just ended, and Will Smith said that he has a copy of your book, because he is acting in all sorts of science fiction films. How do you see science fiction?
Ray Kurzweil : Science fiction is a great opportunity to speculate on what can happen. To me, as a futurist, she throws up scripts. But the creators of fiction poorly evaluate the time frame and other things. In this film, for example, the characters return to Earth after a thousand years and see that biological evolution has gone so far that animals have become completely different. This is not realistic. In addition, there is a bad tradition to extol the dangers of the development of science over its advantages, hence the dramatic plot. Many films about artificial intelligence suggest that AI will be very smart, but it will lack the key human emotions, therefore, AI will be dangerous.
What is the secret of future predictions?
30 years ago, I realized that the secret of success would be a matter of time. They send me different technological ideas, and I can say in 95% that teams will be able to implement them if they have the resources. But 95% of these projects will fail, because time is not how I can guess. For example, search engines - I knew that they would begin to develop. Fifteen years ago, Larry Page and Sergey Brin were in the right place at the right time with the right idea.
Have you predicted search technology?
Yes. I wrote about it back in the 1980s. [The book was published in 1990]
Could you foresee that you will work in a company that started with search technology?
This is exactly what you can not predict. It was very difficult to guess that these two Stanford children would take over the world of search. But what I learned for sure is that if you evaluate the key indicators of price performance and the capacity of information technology, they will line up in a surprisingly predictable smooth exponential curve. The cost of computer performance has increased exponentially since the 1980s. She passed through fire and water, through war and peace, and nothing stopped her. I think it will last until 2050. In 2013, we are on this curve where we should.
What are you working on at Google?
My mission to Google is to develop natural language understanding with a team of other Google developers. The search went beyond the simple search for keywords, but it still searches through billions of pages in search of semantic content. If you wrote a blog post and you have something to say, you do not just write out the words and synonyms. We want computers to understand semantics, meaning. If this happens, and I believe that this is possible, people will be able to ask more complex questions.
Are you participating in the Jeff Dean program to create an artificial "Google Brain"?
Well, Jeff Dean is one of my employees. He is a leader among researchers. We will use its systems and techniques of deep learning. The reason I'm at Google is a resource like this. In addition, the body of knowledge and advanced parsing system, as well as other advanced technologies, are what I need for a project in which I will develop an understanding of natural language. It's much easier for Google to do this because of their technology.
If your system really understands complex natural language, will you call it reasonable?
Yes, I will. I have a steady date - 2029 - for this prediction. And this includes not only the concept of logical intelligence. This includes emotional intelligence, which can be fun, playful, sweet, loving, understanding. This is the most difficult thing we can do. This is what separates computers and people now. But I believe that by 2029 this abyss will disappear.
Will everything be simplified with better computing and software, or are there some unresolved issues, obstacles?
There are requirements, both hardware and software. I believe that we are pretty close to the required level of software. In part, we are helped by a deep understanding of how the human brain works, and it is constantly growing. Now we can look inside the living brain and see how thoughts form in real time. We can see how the brain thinks and how thoughts shape the mind. Many of these studies show how the cortex mechanism works, that is, where thinking takes place. Biology inspires us to develop computers. We already emulate thinking on computers. The deep learning techniques I mentioned use multilayer neural networks that are based on real ones. Using these models borrowed from biology, plus everything that developers have been doing for decades in the field of artificial intelligence, plus constantly improving hardware - we will get to the level of a person in twenty years.
We really understand why this or that brain leads to completely different expressions of human nature? Take Einstein's integrated thinking, the productivity of Steve Jobs, or work on Larry Page's result. What makes these people so special? Do you have any thoughts on this?
In fact, I asked this question, especially about Einstein, in my latest book, How to Create a Mind.
Tell me about it.
There are a few things. First, our brain is created by our thoughts. We have a limited cortex that can accommodate approximately 300 million hierarchical pattern recognizers. We create this hierarchy by our thinking. This does not mean that the genius of Einstein is due to the fact that he had 350 million or 400 million of them. All about equal opportunities. But he organized his brain so that he would think deeply about a specific topic. He was fond of the violin, but did not become Yasha Kheifets. And Yasha Kheyfets was interested in physics, but he was not Einstein. We have the potential to work for the whole world, but in a certain segment. And this imposes certain limitations on the brain, but Einstein was deeply concentrated on one thing.
But after all, many physicists are heading their way, and only one has become Einstein.
I did not finish. The second aspect is the courage to follow your own thought experiment and not fall off the horse due to the fact that the conclusions are very different from your previous assumptions or disagree with the beliefs of society. People are so unable to defend independent thinking next to their peers, that they immediately give up their beliefs as soon as they arrive at absurd conclusions. We need a certain rod to defend their point of view. Obviously, he had Steve Jobs. He had a vision and he followed it. Call it courage convictions.
What is the biological basis of such courage? If you had an endless opportunity to analyze the brain, you could say: “Oh, here you are, courage!”.
This is the cerebral cortex, and people who persistently fill it with their peer beliefs will probably not be the next Einstein or Steve Jobs.
Can this be controlled?
Good question. I thought about it. I also thought about why some people readily accept the exponential growth of information technologies and their consequences, while others are immune. I came to the conclusion that the connections in our head are connected with linear expectations, because 1000 years ago in the wild it helped me track down meat for lunch. Some people are easily led by the future when you present evidence to them, while others are not. I'm looking for the answer to the question: what causes this? It is not related to achievements, intelligence, level of education or social status. The cerebral cortex of some people is arranged in such a way that they can accept the consequences that await them, without worrying about the opinions of others. Can this be learned? I think so, but I can’t prove it.
Since we are talking about Steve Jobs, let me give you one of his famous quotes from a Stanford performance. He said: “Death is probably the best invention of life. She is the cause of change. ” You are famously trying to extend your life to infinity, so you probably will not agree with that, right?
Yes, this is what I call the “death row statement”, who have taken death for thousands of years as a good thing. Once this was the point, because until recently you could not pick up a substantial argument for which you would need immortality. Therefore, the religion that dominated when science, in fact, was not, stated the following: “Oh, the tragic outcome? Well, that’s good. ” We found the point because we took it. But in my understanding death is a tragedy. Our first reaction to someone’s death is accompanied by regret for ruined knowledge, skills, talents and relationships. It is foolish to think that there are a certain number of places, and if the old people do not die, the young will not be able to take their place, because we are constantly increasing knowledge. Larry Page and Sergey Brin did not replace anyone - they created a completely new field. Knowledge grows exponentially. Double every year.
And you believe that a seriously extended lifespan is possible?
I think we are fifteen years from the destination - longevity.
Technological singularity is close.
The article is based on materials .
Comments
Post a Comment