Uncle Bob Martin: “Take me to Toronto, HAL”

I do not cease to read in the news about the imminent onset of unmanned vehicles. Financial news all the time popping about this idea, and technological too. Over the next five years, it is predicted that truck drivers, taxi drivers and uber drivers will be left without work.

I have one word for all these predictions: Toronto!

Do you remember the moment when IBM Watson played and won in Jeopardi? [Jeopardy - TV game in the United States, similar to "His game." - approx. lane.] This happened in February 2011. Success was spectacular. Watson was so far superior to human rivals that it was somewhat sad. After that, Ken Jennings, one of Watson’s opponents, who had won 74 games in a row before, reluctantly admitted defeat, welcoming “our new computer overlords.”
')
Despite IBM's outstanding success, it was not without misunderstandings. And they were very revealing. The mistakes that Watson made were not what a person would have done. Indeed, from them the eyes on the forehead were climbing - and, if you imagine the consequences, then the hair will stand on end.

One such case occurred when rivals were asked to name a city in the United States, in which there is one airport named after the hero of World War II, and another airport named after the battle in it. Think for a second. Run through the three largest cities in the United States. New York? No, John F. Kennedy Airport, La Guardia and Newark are not suitable. Los Angeles? No, Los Angeles Airport, John Wayne Airport, Ontario are not suitable. Aha O'Hara and Midway! It's them. So which city of the United States chose Watson?

Toronto.

Of course, it is now clear that there was a reason why Watson chose Toronto instead of Chicago. In the end, Watson was a computer and computers always have completely discrete and unambiguous reasons why they do something. A series of if expressions comparing weighted values ​​using a complex associative tree gave a final, definite result. That is, of course, there was a reason. Good reason.

I do not know the details. I would like to know, because I think the answer would be interesting from a technical point of view. On the other hand, I don’t really care what the reasons were, because, whatever they were, the answer that Watson gave was extremely foolish.

No one with average intelligence would have made such a mistake. And no one could understand how another person could be wrong. Indeed, anyone who insisted on such an answer (as Watson did) could easily be declared incapacitated.

So, there is a dilemma: Watson outperforms rivals in “Jeopardi” with a significant margin. It is claimed that cars without drivers will also significantly surpass human drivers. Unmanned vehicles will reduce the number of accidents and victims, make the roads safer. So they say.

Over time, the inevitable tragedy will occur. We can imagine it. One day, an unmanned vehicle can knock down a two-year-old child who is lost on the street.

The car will have a good reason to kill him. After all, a car is a computer, and computers always have a completely discrete and unambiguous reason for what they do. And believe me, everyone will want to know the reason why a small child had to die.

Imagine a courtroom. Mad parents, angry press, depressed lawyers representing the company that produced the car. On-board computer is on the rack. He is about to answer the question. There is silence, because everyone is inclined to him. The prosecutor asks the question essentially: "Why did this child have to die?" And the computer, having parsed all the data from the memory, runs along the chain of if expressions, comparing all the carefully weighted values, and finally answers: “The reason was Toronto .”