Artificial Intelligence: Ideological Collision Across Time and Space

2020-01-03

As an IT news optimization translation expert and a pseudo-original SEO optimization expert, I will rewrite and translate the Chinese parts of the title and content you provided into professional English. The requirement is to optimize the original text and meet the originality standards. Note! If there are HTML `

` tags in the original text, they must be retained, while removing the `style` and `class` attributes. If there are multiple empty `

` tags, they should also be deleted. However, the HTML tags themselves should not be modified or deleted, and should be kept as they are. This is very important! Do not return any extra content, and only output the English content. Do not output any other excessive content. Do not include quotation marks before or after the content. Here is the content you provided: "It is well known that artificial intelligence is very popular now and is considered one of the two technologies that are changing the world, along with big data. However, in reality, artificial intelligence is also divided into different levels, namely weak artificial intelligence, strong artificial intelligence, and super artificial intelligence. This is a progressive process. On September 26, 1985, Richard Feynman, the Nobel Prize winner in physics and the most intelligent theoretical physicist after Einstein, who was also the first person to propose the concept of nanotechnology, proposed the idea of strong artificial intelligence in a lecture. Today, many years later, with the emergence of machine learning through artificial neural networks, Feynman's ideas and their proximity to the present, even though they were proposed 35 years ago, are still fascinating.

Stages of Artificial Intelligence

Weak artificial intelligence does not have the ability to think. It is essentially achieved through statistics and function fitting, and cannot truly reason or solve problems. It does not have its own worldview or values. It only appears intelligent, but it is not. To give an analogy, weak artificial intelligence is like telling it: if I wave at you, you should wave back and greet me. After this training, it can learn to wave and greet. However, there are significant drawbacks. It only follows commands and does not consider the environment or the level of danger. Even if a knife is swung at it, it will faithfully complete the preset instruction to wave, in other words, it is stubborn! Strong artificial intelligence has its own thoughts and can think independently. It has its own values and worldview. It has biological instincts, such as the need for survival, safety, and rest when tired, and so on. Using the waving analogy again, if you only teach it to wave, it will only wave. But with strong artificial intelligence, when you wave at it, it will judge whether waving back is dangerous, for example, if there is a wire above. If weak artificial intelligence does not have a program to detect wires, it will wave without hesitation. However, strong artificial intelligence will understand that waving when there is danger will lead to trouble, so it will not wave and will wait until it is in a safe area before waving. Super artificial intelligence has its own worldview and values, can establish its own rules, possesses the instincts that humans have, and has the creativity of humans. It also has a brain with thinking efficiency and quality that is countless times higher than that of humans. Examples include the Terminator, Red Queen, and Ultron. Currently, artificial intelligence is still in the stage of weak artificial intelligence. Even the so-called strong artificial intelligence that is expected to develop in 20 years is only pseudo-strong artificial intelligence, which is essentially weak artificial intelligence. It just has better quality and can do more things, appearing to be infinitely close to strong artificial intelligence, but its essence has not changed much. In summary, the current trend is to make artificial intelligence more human-like.

Imaginations and Thoughts from Over Thirty Years Ago

Over thirty years ago, Richard Feynman raised a question about the model of intelligence by comparing the differences between the motion patterns of natural evolution and mechanical design: Regarding the question of whether to make machines think like "humans," my view is that we try to make these things work as efficiently as possible with the materials we have, but the materials are different from neurons. If we want to create something that can run fast on the ground, we can observe the cheetah's running and try to create a machine that runs like a cheetah, but obviously, it is much easier to make a machine with wheels. Just like airplanes can fly, but they don't fly like birds. They don't flap their wings. They have another device at the front that can fly in all directions, or more modern airplanes have a pipe where you can heat the air at the front and then expel it from the back, as well as jet propulsion, jet engines, internal rotating fans, and so on. They also need to use gasoline. These are all different. So, without a doubt, machines in the future will not think like humans. The same principle applies to intelligence. For example, they won't do arithmetic like we do, but they will do it better. For arithmetic, they do it much better than anyone else. They are much faster and their methods are completely different, but fundamentally the same, because they will eventually arrive at the same result. This is a good example. We will never change the way they do arithmetic to make it more like humans. That would be a step backward. Because the arithmetic done by humans is slow, tedious, messy, and full of errors. So, computers do many things much better than humans. If you want to compare machines with humans, you better remember this... How to make machines think like humans? This is a question that many scientists are struggling with today. But over thirty years ago, someone proposed the idea: why make machines think like humans? This undoubtedly directly contradicted the efforts of many people. But if you think about it from a different perspective, the road to the evolution of artificial intelligence is so dark and long, with countless sparks of thought colliding. Who dares to say that they are completely correct?"

RECENT AI NEWS

RECENT AI TOOLS