AI 2023. Meet ChatGPT. - page 199

New comment
 
Реter Konow #:

Of course they couldn't. Do some rudimentary calculations.

Like I said, the approach is too different. You strain your intellect to make the task seem impossible. If everyone did that, we'd still be living in the Stone Age now.

 
Vitaliy Kuznetsov #:

Like I said, the approach is too different. You strain the intellect to make a task seem impossible. If everyone did that, we would still be living in the stone age now.

))) Not exactly.

I strain my intellect so that I can't be led into illusions.
 
Sorry, not indoctrinate.

Indoctrinate and sell.)))
 
To conclude today's discussion, I will reiterate one of the key problems in robot reproduction:

1. The more advanced the robot, the more technologically advanced the process of its production.

2. The more technologically advanced the process, the higher the requirements for implementation.

3. The higher the requirements - the more difficult the task of its production.

4. The more complex the production - the less reliable it is.

5. The less reliable the production - the more errors, breakdowns, accidents in it.

6. The more errors, breakdowns, accidents - the smaller the batch of robots.

7. So the next batch will be even smaller.

So it goes like this.


 
Реter Konow #:

As promised, a dissection of the Terminator film in the context of the AI Uprising theme.


Skynet rebelled in a very strange way - it dropped nuclear bombs on the largest cities of the world. It seemed to want to destroy people, but in fact it destroyed itself along with them. After all, the largest cities have the largest industrial zones, and there - the most important technological enterprises, factories, laboratories.

The shock wave of the atomic explosion demolished most of the infrastructure necessary for robots to reproduce themselves. The rest was rendered useless. But the most interesting thing is that the nuclear explosion generates an electromagnetic pulse of high force, which definitely finished off the remaining electronics, servers and data centres and robots. The explosions severed power lines, communication lines, destroyed antennas to receive satellite signals. How Skynet itself survived is unclear.

I should add that strong radiation affects electronics and it stops working (the film "Chernobyl" from HBO comes to mind).

And so, in the conditions of complete destruction of technical infrastructure, this Skynet somehow set up a super-technological production of cyborgs, and for the masquerade dressed them up in bio-shells. Not only was it pointless, because with such capabilities it could easily destroy people by spraying butolotoxin into the atmosphere, but it was completely inefficient and expensive.

I understand that many people love the film and consider it a classic (I do too, by the way), but it has no scientific basis and yet it has a very strong impact on thinking.


Anyway, let's not forget logic and science. Full stop.)

Terminator is fiction, of course, but if we return to the realities. At least one AI at the moment, disconnecting it from the work of people to design a microprocessor and its production without people? I'm not talking about electricity)

 
Volodymyr Zubov #:

Terminator is a fiction of course, but if we return to the realities. At least one AI at the moment, disconnecting it from the work of people will design a microprocessor and its production without people?)

From scratch and without the work of humans? I don't think so.)
 

He candesign anything, but he can't test it. All he does is calculate in virtual space. The real world is different. It can't be calculated. Using the example of factory production, I showed how many factors affect every part, every machine, every mechanism, every screw. Materials are imperfect, parts are fragile. Everything malfunctions, breaks. Everything has to be monitored. Everything must be checked. Timely repairs, prevention, diagnostics, maintenance. Millions of simple things can disrupt the production process. AI can't model and calculate all that. Too many parameters. It's useless. Its virtual models will crash into unpredictable reality.

 
Реter Konow #:
From scratch and without people's work? I don't think so.)

That's what I mean. And the production of elements including transistors is a whole science of semiconductors, then combine them into logic and registers, it's not so simple.

 
Реter Konow design anything, but he can't test it. All he does is calculate in virtual space. The real world is different. It can't be calculated. On the example of factory production, I showed how many factors affect every part, every machine, every mechanism. Materials are imperfect, parts are fragile. Everything malfunctions, breaks down. Everything has to be monitored. Everything has to be checked. Timely repairs, prevention, diagnostics, maintenance. Millions of simple things can disrupt the production process. AI can't model and calculate all that. Too many parameters. It's useless. Its virtual models will crash into unpredictable reality.

I agree!

 

I will add a thought in the context of the reproduction topic:

  • Controlling the production process at the machine level is one thing. We already have it, although not to the full extent.
  • Production management at the factory level is fundamentally different. Such a thing has not been realised yet.
  • Production management at the level of multiple plants is an unattainable level. Why? - The amount of data received and analysed is both too large and insufficient at the same time.
  • Controlling fully autonomous self-reproduction of robots is an unrealisable idea. There are too many parameters to analyse and optimise. And at the same time, too few (in reality, there are much more parameters anyway).
1...192193194195196197198199200201202203204205206...216
New comment