⬆️ ⬇️

This year there will be no robots, whatever Ilon tells

"Prediction" from Robert Krinzhli (real name - Mark Stevens). Journalist, writer, business consultant, documentary filmmaker. The author of the best-selling story about Silicon Valley, “Accidental Empires: A Battle of Competition, and Still Can't Get a Date” [Random empires: how Silicon Valley boys earn their millions, fight foreign competitors and still not capable of stirring up with a girl]. Repeatedly published in the New York Times, Newsweek, Forbes and other publications.



We all know people who do not like anything. They are successful people who, as it sometimes seems, have succeeded in answering no to all questions. I'm not like that. I'm an optimist. I even like to take a little risk. But I will not argue that in 2019 we will see something more than beta testing of robots. Therefore, my prediction number 4 [ from a series of technological predictions from the author’s website / approx. trans. ] will be as follows: this year, robomobils will not appear in retail sale in any form. We are simply not ready, and probably we will not be ready for this for several more years.



Robo-mobile problems are not in the field of technology. We have almost completed this technology over the past decade. Dock there fresh data collected by Google, and, especially, with all these adzes on autopilot, and it turns out that almost all the problems with the machines, leading themselves, will end. However, this will not be allowed to happen, because otherwise people will die - mainly because of idiots.



The problem is not roboMobiles, it is in machines that do not behave themselves, and that drive such idiots like me.

')

I first encountered robots in 1995, and I already wrote about this . Then the idea was not to have the car bring your children from school, but to push more cars on the Los Angeles highway. They would be controlled remotely, from the moment of entering the highway to the moment of exit, they would have traveled one meter from each other, saving fuel due to the reduced air resistance.



And this system could work if only cars under direct control from outside could drive on the highway. Let them travel everywhere at a speed of 120 km / h even during peak hours, and there will be enough economic pressure to implement this system.



But today everyone wants to eliminate drivers completely, door to door, and it is much more difficult. But we would have succeeded if it were not necessary to interact with the millions of drivers who still control their cars in the old fashioned way, which often means with barely enough competence. Therefore, we abandoned the concept of superhighways and switched to another form of road wars.



Machines have always served for 10 years, but over the past couple of decades, this period has increased due to a couple of techniques: 1) plating of steel bodies that do not rot like your dad's Oldsmobile, and 2) a significant increase in car prices, even taking inflation into account. We use cars longer because they do not rot, and we cannot afford to change them often.



As a result, if earlier we could expect dramatic changes in car technology every ten years, now it takes twenty. I have a 2006 GMC Yukon XL Denali with a mileage of 264,000 km. My next car, if possible, will be the 2006 Mercedes-Benz E320 CDI (if you have a good copy, let me know). Therefore, I can’t do anything to achieve a robust singularity.



It will eventually happen. When half of the car park is replaced by cars that can drive on their own, if allowed, there will be a great financial incentive to remove the other half from the streets. It will be especially powerful if the acceleration of climate change continues further.



I think that most of the cars of the 2020s could become robotic after updating the software, which is why Ilon Musk predicts Tesla will have full autonomy by the end of 2019. But note that Ilon does not predict that Tesla will be allowed to drive on their own everywhere.



First we will see it on the highway, where the interaction between different machines is easier to handle. But a few failures can easily put an end to this undertaking, and then only in the 2030s we will have the feeling that all cars suddenly became automatic, and this will happen purely on the possibilities of the fleet.



So what is the world so actively discussing robomobils and complete autonomy? Part of this is the hype around Tesla, and part of it is marketing, because automakers want to give us new cars that can eventually become robotic cars, but perhaps not before they get a second owner. Other reasons for such an abundance of talk related to the fact that Uber plans to go public this year.



In Uber, they say that last year they spent $ 750 million on research in the field of ro-mobiles. I have no idea if this is true. Maybe money spent on robomobili, maybe for something else. Maybe they bought a huge amount of delicious dinners for engineers who are developing robo-mobile. I know that they have spent a lot, because Uber is going to get another billion in investments from SoftBank and other companies only to continue working on robo mobile technology, which, by their hints, can cost $ 5- $ 10 billion by itself



And what will it cost if Uber can not use it until the 2030s?



I do not think that the management of the company cares about it. This is still a long time to come, and they need to worry about an IPO, which will be much better if motorists believe that we will be able to see robobiles very soon. Uber has a problem with the staff. If she can convince everyone that uncouth and expensive human drivers will soon give way to electrons, this will convince Wall Street. But, as I explained, it will not be true.



The world has not yet come to this - and Uber, Tesla, and everyone else at some point suddenly recognize this, somewhere a year after entering the stock exchange.



At the same time, it cannot be said that there was no progress on various types of robility vehicles. I recently discussed this topic with a reader named Chris Edwards working for Buffalo Automation , a company that develops automated vessels. As for me, this is a much more meaningful exercise, given that the traffic at sea is less and the traffic lanes are wider. I especially like the idea of ​​automatically mooring boats.



“Personally, I believe in the possibility of solving all simple problems, but not in solving the problem of idiots,” says Chris. - The huge potential lies in areas such as a tractor (I can assure you that Case New Holland is hiring people with my experience) and mining (I love Komatsu, she is a true pioneer of autonomous vehicles). Please note that in such enterprises the problem of idiots is completely absent.



We in Buffalo are also interested in improving the blind zones (sometimes literally) that the robots industry has in conditions like snow and winter (with some exceptions ). I personally have seen how idiots get into accidents on snow many times, and I can assure you that 99.9% of such accidents are caused by a problem to which the car has immunity - namely, too much driving speed. There is great potential here if we can overcome this obstacle in the form of idiots. ”



In a conversation with Chris, I forgot to mention the story I wrote about 25 years ago for Forbes about an automatic harvester from Caterpillar. At about the same time, I was studying information about California computer-controlled highway research. Caterpillar, a major producer of agricultural machinery, was worried about leaving the farm workers, and was looking for ways to produce equipment that could drive on its own. They hired an engineer who worked with neural networks and bungled a Macintosh II-based system that actually worked. She studied independently, was able to process any field of any size and shape, and independently calculate how to harvest the most efficiently from him - usually corn or wheat.



The only reason Caterpillar did not bring autonomous equipment to the market in the 1990s was a problem with the laws - and with this problem, manufacturers of robots will butt for many more years. Lawyers did not allow Caterpillar to sell the combine. They had to figure out what to do with the self-learning crop optimization system, instead. In a fit of genius, the company commissioned an engineer with Mac II to convert the system from driving a combine to optimizing the income of their pension fund, which independently brought itself to record profits in the next few years.

Source: https://habr.com/ru/post/447096/



All Articles