When people talk about the dangers of artificial intelligence (which, in reality, we still haven’t actually achieved), the first thing to consider is not a nuclear war from The Terminator or the extermination of humanity, but simple oblivion.
Technological progress has already left a large number of people behind. For a significant portion of the population, almost everything that happens today is magic: mobile phones, the internet, electricity, cars. With every passing year, people understand less and less how the world around them actually works.
The appearance of neural networks has sharply widened this gap. And now it is hitting the engineers themselves. It is no longer always necessary to write code by hand, create 3D models, or solve equations. On the one hand, this accelerates progress. On the other, knowledge is being lost. People begin to forget skills they once possessed, grow lazier, and increasingly shift the work onto neural networks.
On one side, this looks like the loss of a culture. A decline in the number of engineers who truly understand their subject will lead to the closure of entire industries. Those that remain will start making serious mistakes that gradually become the norm. Planes will crash more often not because of bad autopilots, but because of poor maintenance. The overall quality of almost everything that is produced will noticeably decline.
On the other side, a relatively small number of genuine experts will remain. These people will become economically almost priceless. Only the largest companies will be able to hire and train such specialists — simply because only they will have the necessary financial resources. The majority will consist of people who from the very beginning worked only with the help of neural networks. Less competent people, unable to properly evaluate what the machine is doing. But the cost of their labor will be lower. As a result, we will get a huge volume of mediocre products made possible by yet another reduction in production costs — even though the quality will be worse.
Today, production is often outsourced to developing countries (for example, India) precisely because it is cheap. The average quality of the products is not high. Neural networks perform even worse in this regard, but compensate for it with an extremely low cost of development.
At the same time, neural networks will keep getting better. At some point it will become possible to hand them the entire production cycle and bring the system to a semi-technocratic state. For example, a fully autonomous system for growing potatoes. It could be scaled almost without limits, while people would only handle design and planning. Everything else — from planting to harvesting — would be done by machines. Moreover, the machines could maintain and repair themselves.
Where does this scenario lead? To overpopulation? To real technocracy? Most likely it will become a new stage in the development of humanity. And it is more probable that private companies, rather than states, will succeed in it. Tesla’s robots, for instance, could not only work in factories but also provide people with food — as part of the mission to settle Mars. And these would be specialized robots. Step by step, the basic minimum of human needs could be covered inside a city that Musk built to service his launch pads. A micro-utopia inside a single private company.
On the scale of the entire planet this is hard to imagine — there are too many disagreements, problems, and wars. Besides, it is unclear what people freed from work would actually do. Cause harm? But inside a large private company, where the culture is relatively homogeneous and the average level of intelligence is higher than the planetary average, it is quite realistic.
What awaits us in the end? Even more division, problems, and conflicts. Some will sink into oblivion, others will build their own technocracy. And these two groups will most likely barely intersect.
Comments (0)
No comments yet.