This is a matter of opinion.
Kurzweil pointed out there are six epochs, and similar views are actually standing on the macro laws of historical development. But can laws really predict the future? here I’d like to discuss Three issues.
first one, The illusion of technological acceleration
In the past 40 years, IT has exploded and mankind has entered the Internet age. But the progress of IT technology does not represent the development of overall technology. Many of mankind's great inventions, such as electricity, telephone, automobile, airplane and so on, were completed before 1940. After that, in addition to the rapid development of IT, other technologies have not made epoch-making breakthroughs. Besides, Scientific revolutions are rare, not cumulative.
The next question is Computational ability
The development of AI is now largely based on computing power. But is the connection between computational speed and singularities reasonable? Let's assume that the gorilla brain can work fast, but we know that it still can't think at the human level, because the essential difference between human and gorilla is that the human brain has uniquely complex cognitive modules. The same is true of artificial intelligence, which is faster, but the nature of the human mind has not been solved.
The third issue,The "consciousness" problem
Reaching the singularity means having machines that can match human intelligence, something we call artificial general intelligence. Such AI has true consciousness. But now we can't explain how "human consciousness" came into being, so how can we inform the machine of "consciousness", "emotion" and so on in an algorithmic way? It is obvious that "consciousness" cannot be accurately described or measured by science.
In short, to reach a singularity requires a fundamental breakthrough in principle. Because according to the current principles of ARTIFICIAL intelligence, we need to digitize human life scenes to model machines, and many things we can’t digitize. Of course, It’s not impossible to do everything by numbers, but it’s very unlikely.
*伦理pre
