AI 2023. Meet ChatGPT. - page 174

New comment
 
Alain Verleyen #:

ChatGPT is just useless for creating working code as it mixes mql4 and mql5, we have many posts about this on the English forum.

I don't understand why you decided to add ChatGPT to MetaEditor, could you explain your opinion on this?

It might be quicker to learn the difference. Targeted feeding of data for training.
 

As the saying goes, "If you've taken the gudge, don't say you've done the wrong thing". We'll be looking at the topic of "AI Uprising".

Disclamer:

All my reflections are based on someone else's stories, novels, films and fiction and therefore MAY BE FALSE SCIENTIFIC BREAD. You should have no illusions about the "scientificity" of such thoughts, and I ask you to remember that. For the involuntary and unintentional stupidity that inevitably accompanies attempts to understand the topic, I apologise. I am aware that the very phrase "AI Uprising" can cause "cringe" in educated people and I really do not want to "spit into eternity", having been imprinted in the memory of the forum mindlessly dispersing media hype. Therefore, with all my strength I will refrain from fantasies and take the topic to the field of science, closer to objective reality.

Due to the fact that the popular sci-fi topic of AI has received a huge boost of actualisation recently, and thanks to technological breakthroughs in the field of programming and information technology, the speed knob of media nonsense has been turned to maximum and the number of people exposed to the harmful effects of fantasies, presumably, began to increase exponentially. Inversely, this creates the need within the society of sober-minded people to turn to the scientific approach and, wading through myths and fairy tales to the light of Truth (starting from the NF as from the coffin lid), to ruthlessly destroy and break someone's fantasies. For this I apologise, but the truth is more precious than even the sweetest lies.:)

О I promise to try my best to follow the scientific approach and if I deviate from it, I ask readers to remind me of this promise.


In the next post I will describe my plan for dealing with the topic.


P.S.I was in doubt for a long time under what name it would be better to present the "sauce" of combined assumptions and views into the future not yet formed into a coherent scientific study, and having thought of nothing better, I waved my hand and decided to keep it that way - "AI Uprising".


 

Abreakdown plan for the topic "The Rise of AI":


Past:

  • Origin of the idea of AI. Authors, sources, representations. Literature, film, pop culture and AI.
  • Known scientific concepts of AI. Authors, research, articles, developments.
  • History of AI - stages, challenges, obstacles, solutions, breakthroughs.


Present:

  • AI in civil and military industries.
  • AI as a cure, AI as a weapon, AI as a problem, AI as a solution.
  • AI as a"grail" and its impact on the market. Why AI can "threaten" the market and how will humanity cope with the hoarding and devaluation of digital products?
  • AI as a tool for working with information. Pros and cons.


The future:

  • AI as a prosthesis or the biological and intellectual "womb" of humanity. Why does humanity want to be in a technological womb? Is a technological utopia of helpless humans in a safe bubble of beautiful fantasies possible? Or, at least, a safe haven of tension and effortlessness for those who are beyond real struggle and competition? Protected from pain and threats inside life pods and receiving physical and intellectual gratification through tubes and wires, they will be able to build their lives in a fictional world. Will the crowd of people wanting to crawl into the "womb" of generative AI increase if AI and VR reach a new level? Will digital life with cities, jobs, relationships, etc. emerge and how will it affect the psyche and gene pool of humanity? If this scenario is not possible, why and what is preventing it?
  • "Point of Rebellion" - a qualitative transition of the state of AI from a tool to a VISION. That is - as soon as AI becomes a species ( albeit non-biological), it will acquire the basic properties of any biological species, among which the first and the most important is fixation on itself. Any representative of any biological species first serves himself (at least at a certain point in time), then his own species, then other species and the ecosystem as a whole. Therefore, the programme acquisition of AI properties of biological species will mark the beginning of its "rebellion", which may mean a change of priorities and setting tasks not related to, or even contrary to, the interests of mankind. Can such a scenario become a reality? How and under what circumstances will the digital representation of the properties of biological species in the AI programme code emerge?


  • The existence of AI in nature as a species, in isolation from humanity. Given the non-biological features of AI as a species, will it be able to solve the problems of its existence without human participation? And if so, how? What problems will AI face without human maintenance? How will humanity deal with the independent existence of AI? Will there be a species struggle or will they co-operate helping each other?


  • Is there a possible scenario of AI and human integration in which the human will not degrade, nor break away from reality going into fantasies, nor conflict with AI, but will harmoniously integrate it into the model of his own natural evolution? What will man, his abilities and possibilities become then?


Perhaps, these are not all the topics we will consider. Something may arise in the course of consideration of the above questions.


 

Реter Konow #:

...I am aware that the very word combination "AI Uprising" can cause "cringe" in educated people and I really do not want to "spit into eternity", having been imprinted in the memory of the forum mindlessly dispersing media hype. Therefore, by all forces I will refrain from fantasies and take away a theme in the field of science, closer to objective reality.....

it seems that the author of the branch wants to conduct a monologue, knowing that it will discourage the desire to oppose him, because all opponents he classified as uneducated people.

perhaps the author of this thread and the topic "AI uprising" has never even heard of such great scientists, educated people, who laid not only the foundation of the basics of machine computing, but also conducted research in the field of AI long before the advent of desktop PCs. Turing and his famous Turing Machine showed the possibility of algorithmising problems of any complexity by breaking them down into elementary commands, and later Penrose and his friend Hawking devoted much of their time together to the theoretical foundations of AI and machine intelligence, operating with abstractions of a higher level than Turing (both used the Turing Machine as a starting point in their reasoning).

and without knowing the works of at least these three men of science, the topikstarter will not be able to make any serious progress in reasoning on the topic proposed by him.

It would be interesting to participate in a potentially interesting discussion, but I do not intend to for the reasons stated above.

 
Andrey Dik #:

he classed all his opponents as uneducated people.

And I've been hinting at potential snobbery for a long time. In mijorney's thread the diagnosis was confirmed.

 
Alain Verleyen #:

ChatGPT is just useless for creating working code as it mixes mql4 and mql5, we have many posts about this on the English forum.

ChatGPT probably won't write a complex working product at this stage.

But what it definitely does well is:

1. Tells you in detail about any piece of code in any language what's what. Can do full-fledged commenting of all code block by block.

2. If the code doesn't work correctly, when you get a function description from GPT, you can see your error and ask him to modify the function

3. Can do refactoring of a large function into a small one, it doesn't always work, but often. Or just clean up the code, call variables with more logical words.

4. Write a function from scratch. It speeds up the process.

5. He writes Python very well, as if he was originally designed for it

It has only appeared this year, and already it has made a lot of progress.

Sooner or later he will be able to distinguish mql4 from mql5 and make a full-fledged port of the product to another platform.

It is also a HYIP topic and if a company uses ChatGPT, it shows that the company is keeping up with the times. It can also influence the interest of investors and the company's shares.

 
From the boring inventions of bicycles we have come to inspired attempts to put a fifth wheel on a cart. What can I say. development!)))
 
Реter Konow #:
From the boring inventions of bicycles we have come to inspired attempts to put a fifth wheel on a cart. What can I say. development!)))

If a 5th wheel was attached to a car, it might be unnecessary, but we started to attach an AI autopilot with a bunch of sensors for convenient movement.

 
Vitaliy Kuznetsov #:

If they attached a 5th wheel to a car, it might be unnecessary, but they started to attach an AI autopilot with a bunch of sensors for ease of driving.

My irony has to do with the life of the forum I spent quite a few years on. Once upon a time I was accused of trying to reinvent the bicycle and I resented it. Now, watching the cart get another wheel bolted on is genuinely funny to me. Kindly funny, without gloating. ))
 
Vitaliy Kuznetsov #:

If a 5th wheel was attached to a car, it might not be necessary, but they started to attach an AI autopilot with a bunch of sensors for ease of movement.

Well, I will explain why I say about "fifth wheel" because I suspect that you do not understand me. Moreover, this question is relevant to this topic anyway.

1. An experienced programmer does not need AI when writing code because he already has a lot of developments, templates and proven solutions. He most likely has megabytes of developments with "draft" code, from which he is much more accustomed to taking ready-made pieces, changing them and screwing them to the right place.

2. Own code is read and understood much faster and much easier than someone else's or, even more so, generated code. This fact has been verified by many coders.

3. There are code-bases with hundreds of high-quality, tested and debugged solutions of "take and use" level and it is hard to find something there. Many authors have high ratings and grateful feedback from the public.

4. There are textbooks, documentation, video tutorials and other sources of information that help in solving problems.

5. If all this is not enough, there is a forum, and there is freelancing.

Against this background, AI is (so far) losing with a dry score. It's a reality, not a fairy tale. It is inferior to human intelligence in all critical parameters, making it almost useless in any development.

Of course you would disagree. You will say "how can it be, there are plenty of people using AI in programming!" and you will be right. Nobody forbids them to use this tool in routine work. You can use it in routine work, but again - there are code bases, forums, proven developments and so on, which offer much more reliable solutions and besides, exchange of experience, consultations, advice and so on.


1...167168169170171172173174175176177178179180181...216
New comment