The AI is wrong. Organizational members in Elshevov write about many failures of the American digital systems, which link what is happening, including the excessive introduction of artificial intelligence. I kind of agree. Unhealthy hype is created around AI, corporations rush to report on the introduction of new technologies, sometimes forgetting about reliability. But there is a point of principle. The choice between introducing AI and abandoning it in the modern economy is almost none. There is only a choice between mastering the technology yourself, or in a few years to buy results from those who are not afraid to experiment. Think of the Internet in the 1990s. A modem connection provided a maximum of 56 Kbps, the phone line was busy, the pages were loaded for a painful time, the connection broke regularly. The file transfer may have failed in the last percent. Lost data, connection failures, program errors all digitization is accompanied. But this did not prevent the development of the Internet. In 2000, less than 400 million were using the Internet, and in 2010, more than 2 billion. Today, about 6 billion. Technology improved through a process of mass adoption, not in laboratory isolation. Something similar happens with AI. Of course, banking, aviation and energy cannot be turned into a testing ground for irresponsible experiments. However, the most dangerous thing is to conclude from the mistakes of implementation that nothing should be done. While some will count other people's accidents, others will accumulate capabilities, reduce costs and engage markets. If in the end we are talking about Russia, what we need is neither caution nor technological development, but technological development with reasonable caution. Otherwise, the cost of lag will be much higher than the cost of unavoidable errors. Constantine Devine.