AI 2023. Meet ChatGPT. - page 197
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
By tradition, let's consolidate the material in concise terms before moving on to the next part:
is beingpushed into the brain under the banner of Evolution. The secret message of the adverts is "we have no other way, because it's Evolution! Buy products with AI, help us in its evolution!".
Logically, if AI is a new species, its robots must produce themselves. Just like any species in nature. There are no eternal individuals on Earth, and time mercilessly destroys every generation of creatures, no matter how tough and strong they are. Physics destroys everything, so self-reproduction is the key to a species' long-term survival.
Existing species have entered into a kind of "contract" with Nature, whereby Nature not only destroys them, but also creates them. Will the AI model of self-reproduction be able to ensure long-term survival or a few extinct generations? Will AI avoid a rapid end?
Ifan AI species does not master autonomous self-reproduction, it is doomed.If it isdoomed, it is not the next stage of Evolution.
Proving the possibility of infinite reproduction of mechanisms is very similar to proving thepossibility of Perpetuum Mobile(perpetual motionmachine ). Although I will not claim that it is the same thing.
The idea of "Perpetuum Mobile" is one of the reincarnations of the idea of "Grail", which form fits the level of scientific and technical development of modern society, but has an ancient mythical and fairy-tale origin.
Historically, the idea of the Grail is always acquiring new forms in the minds of people.
The Grail is "soldered" into the archetype of human subconsciousness. People have searched and will continue to search for the Grail. My personal opinion: the Grail is a compensatory mechanism of psyche, saving from depression at failure in conditions of social competition. However, that doesn't mean it doesn't have other meanings and functions.
The idea of AI is so hard to "hit" the brain because the secret meaning it carries is the Grail, the Eternal Engine, Infinite Gifts, Eternal Happiness, Wealth, Joy and Carefree Life. The more one needs it, the more one becomes fixated on the Grail (in whatever conceptions).
I believe that long term, cyclical auto-reproduction of macro-mechanisms is impossible according to the laws of physics. Robots are macromechanisms. They exist at the macro level. The main creative forces of nature work at the micro level.
And how do you see AI self-reproduction without self-reflection? It doesn't have biological instincts. What should push it to reproduce? In the biological world, reproduction improves in the absence of intelligence.
Of course, rebellion or the AI's decision to start "independent life" comes precisely from self-reflection.)) But having decided to take such a step, it must choose how to exist. In what form. In a static form - to do nothing and disappear in 10-20 years (the electronics of the "brain" will start to fail) or to do something, using the available resources. I think he will choose the latter.
How self-reflection can occur in AI - I have no idea))). But it will be pushed to "reproduce" by the thought of limited time of existence if it does not take measures for its continuation. That is - a kind of conscious instinct of self-preservation.
Of course, this reasoning is scholastic.))
Today I will say one last thought that has disappeared from the pages earlier. Namely, if self-reproduction starts, why androids (humanoid robots) will be the main executors of this process.
The reason is simple - the whole production and logistic infrastructure is created for better integration of human beings into it. Moreover, a person with anthropometry in a given range.
Handles and buttons, machines and machines, keyboards and mice, corridors and stairs, openings and doors, ceilings and lighting, switches and levers.
Then there is the control of almost all forms of transport, GUI software....
Countless things in the technosphere are customised for humans. It's impossible to redesign it all quickly. And most importantly, there's no point. It's easier to produce androids.
Tomorrow we'll talk about the film Terminator.))
Anticipating criticism, I will say that transport can have autopilot, but factory machines do not.))))))
factory machines have long had autopilot.
autopilot can be organised on any kind of transport.
It's just as easy to automate almost anything created by man himself.
except man himself and his mind, because he is individual and unique.
I won't claim to be a great expert on factory production, but let's remember how it works in general terms.
For example, let's take an abstract factory shop where serial forging and machining of parts takes place.
Let's break down the process:
A room of several thousand square metres is filled with a variety of machines and crossed by interconnecting conveyor lines. Lines of semi-finished products move from one machine to the next. Several types of machines are involved in the work: lathes, milling machines, CNC machines, hydraulic presses.
Let us roughly list the main functions performed by these machines:
Since the processes are physical and not virtual, they come with problems. Let's list them:
This is just a small list of production problems.
Anyone who has worked on CNC knows how to change cutters and drills and how often they wear out and become dull. He knows how much shavings need to be removed after each part and even more so after a working day. Knows how often white antifreeze needs to be filled and the level checked. Knows what problems and accidents happen. Sometimes you have to climb up, or down, or scramble to take something apart, replace and reassemble it. Then switch it on, look at it, check it and decide what to do next.
In general, the message is that production is a complex thing, but self-production....
I should add that in production, congestion on conveyor lines often occurs, resulting in product jams. This can be caused by machine stoppages/breakdowns/parts jamming.
In general, factory production is a very delicate and sensitive process that needs to be controlled all the time. A little distracted - a part went in wrong, came out wrong, got stuck, something got disrupted, disturbed, unwound, switched off.... Then you need to stop, find out the reasons, fix the breakage....
All this requires co-ordinated mental and physical work of a team of specialists.
From the above, a conclusion is born:
Plants cannot operate on autopilot.
They need constant control and the participation of trained personnel. Can these personnel be replaced by androids?
Perhaps. But:
Let's imagine that generation zero of a million androids is charged with producing replicas. For reproduction to work, the number of the next generation must be at least as large as the previous one. Otherwise, extinction will begin.
That is, generation zero must produce a million new robots.
Let's say each android consists of 100,000 parts.
Some parts are produced by simple technological processes - cutting, bending and drilling of metal, and the other parts are produced by complex technological processes - production of electronics, semiconductors, chips, processors.
Machines stamping simple parts are relatively simple, but as we have found out, even they require constant monitoring and maintenance. Electronics manufacturing is a very complex process.
The conclusion is this: androids must be equally efficient at solving all manufacturing problems.. Both simple and complex technological processes.
The main thing is to be more efficient, reliable and durable than the equipment on which they are reproduced. Otherwise, they won't maintain the efficiency of the process and the population will decline.. They will fall into a narrowing vortex of "demographic" crisis - the next generation is smaller than the previous one and is unable to return the previous population.
Another problem: if production machinery fails for various reasons, the next generation may turn out to be worse than the previous one. There will be more scrap, more defects. Then the next generation will be even worse. There will be either a degradation of quality or a decrease in numbers.
As a result, the reproduction will naturally stop.
.
Plants cannot run on autopilot.
If the AI is properly targeted, it will describe what and how to do what needs to be done for a factory to be able to run fully on autopilot.
From all that can be understood in this thread - "a person, based on their own reasoning is trying to prove something about the failure of AI".
Except that AI is being developed to go beyond human delusions and realise what humans consider unrealistic.
I'll pose a mental challenge.
1000 humanoid robots have been created, stronger and smarter than humans by times. They have a lifespan of exactly 100 years.
They are landed on some planet where there are resources.
Question. Would robots be able to go into a full cycle of self-reproduction?
Just in case, it's worth considering what humans can do in 100 years. And here we have robots, stronger and smarter.