Rather the approximate time when artificial
Posted: Sun Feb 02, 2025 9:38 am
On artificial intelligence, internet marketing, real estate, singularity theory, problems of the welfare state
- A question about artificial intelligence. When will it be invented?
I'll disappoint you, it's been around for a long time, you just don't know about it.
intelligence will reach the level of an adult. Now, in my opinion, it has reached the level of a 3-year-old child. This is the first question. The second question is, will robots receive money at that moment? The luxembourg number data third question is, how will a person earn money?
Denis Dovgopoly: Sasha Belousov answered this question very well. He said the magic phrase: "What kind of artificial intelligence can we talk about when we don't know how human intelligence works?"
Alexander Olshansky: I advise everyone to read Penrose's "The Emperor's New Mind". It is very difficult to read, but I strongly recommend it... There is a lot on this topic. The problem is so deep that it is almost impossible to cover it now. But, formally, today the only criterion for creating artificial intelligence is passing the Turing test. The problem is that, having, for example, the Google information base, you can easily pass the Turing test. But for some reason, this does not produce artificial intelligence. That is, it is not that we do not have artificial intelligence, we do not even have a formally good metric to measure whether it exists or not, what it is, and how intelligent it is. By the way, look for a newer edition of this book, because this book has been reprinted several times. There are very interesting things there about the general possibility of creating artificial intelligence and some fundamental problems that lie in this area. The book, for example, can be perceived as proof that modern computers will never have artificial intelligence due to the principle of their operation. And with quantum machines it is possible, but we know so little about it now. That is, we have made 64 qubits today - that is 64 cells. In a modern processor there are some billions, and there are 64 qubits and even then these are not quite real qubits.
Denis Dovgopoly: There is a theory that artificial intelligence will become such when it develops a sense of humor.
Alexander Olshansky: There is such a rich field for science fiction here, but it seems to me that futurism is still very weak here. There is no solid ground for any predictions at all.
– I understand the answer is that we won’t live to see it?
Never say never. When I was studying at the institute in 1990, during my lecture a professor, not a fool, but actually a very smart person who really understood the issue, said this phrase: "You are young and you are lucky, you may live to see live video on the screen of a computer." This was in 1990. And so never say never. And so far I do not see any harbingers. This is a black swan in its purest form.
- A question about artificial intelligence. When will it be invented?
I'll disappoint you, it's been around for a long time, you just don't know about it.
intelligence will reach the level of an adult. Now, in my opinion, it has reached the level of a 3-year-old child. This is the first question. The second question is, will robots receive money at that moment? The luxembourg number data third question is, how will a person earn money?
Denis Dovgopoly: Sasha Belousov answered this question very well. He said the magic phrase: "What kind of artificial intelligence can we talk about when we don't know how human intelligence works?"
Alexander Olshansky: I advise everyone to read Penrose's "The Emperor's New Mind". It is very difficult to read, but I strongly recommend it... There is a lot on this topic. The problem is so deep that it is almost impossible to cover it now. But, formally, today the only criterion for creating artificial intelligence is passing the Turing test. The problem is that, having, for example, the Google information base, you can easily pass the Turing test. But for some reason, this does not produce artificial intelligence. That is, it is not that we do not have artificial intelligence, we do not even have a formally good metric to measure whether it exists or not, what it is, and how intelligent it is. By the way, look for a newer edition of this book, because this book has been reprinted several times. There are very interesting things there about the general possibility of creating artificial intelligence and some fundamental problems that lie in this area. The book, for example, can be perceived as proof that modern computers will never have artificial intelligence due to the principle of their operation. And with quantum machines it is possible, but we know so little about it now. That is, we have made 64 qubits today - that is 64 cells. In a modern processor there are some billions, and there are 64 qubits and even then these are not quite real qubits.
Denis Dovgopoly: There is a theory that artificial intelligence will become such when it develops a sense of humor.
Alexander Olshansky: There is such a rich field for science fiction here, but it seems to me that futurism is still very weak here. There is no solid ground for any predictions at all.
– I understand the answer is that we won’t live to see it?
Never say never. When I was studying at the institute in 1990, during my lecture a professor, not a fool, but actually a very smart person who really understood the issue, said this phrase: "You are young and you are lucky, you may live to see live video on the screen of a computer." This was in 1990. And so never say never. And so far I do not see any harbingers. This is a black swan in its purest form.