Chat Generative Parsing Tool (ChatGPT)
----------

There have been protests by television script writers that the existence of ChatGPT will out-mode their craft. I have used both the text-based and graphic generation tools that come from the ChatGPT artificial intelligence system. It has used the Internet as a learning tool, of which it has managed to comprehend and respond to inquiries posed to it. Some of the graphics generated were better than a human would create. As for the text question-response system, the responses that I have received from ChatGPT were of the college level of complexity and completeness. I have been able to get ChatGPT to write Python programming language code by simply describing what I want the software to do, add some specific particulars of what I'm looking for, submit the "prompt," and wait about 5-10 seconds, and then will receive immensely detailed responses to your inquiries.

The Intelligence Quotient (IQ) of ChatGPT would be in the 1% of the Human population, and that is with ChatGPT version 4... when ChatGPT version 5 comes out you can expect that it will be on par with the most advanced scientific philosophical doctorates (Ph.D.) level of thinking. Now ChatGPT will never have its own soul, but its brain far surpasses that of most people living at this time. For a several day analysis and report of a thing into a very polished essay can be created by ChatGPT in at most one minute.

This tech is scary, but try it out and you will see why this tech is going to be so disruptive.

I hope that this tech doesn't destroy the world. Should I wait and see or do something to do about it before that happens? That's a question that bothers me.

Cheers!
-warmfuzzy

--- Mystic BBS v1.12 A49 2023/04/30 (Linux/64)
* Origin: thE qUAntUm wOrmhOlE, rAmsgAtE, uK. bbs.erb.pw (700:100/37)

In my opinion, LLMs (Large Language Models) in general are considerably dangerous not much due to what they can achieve, but due to how confidently they can be wrong.

This can be used to push ideological and marketing material easily down the masses in a more acceptable wording, potentially amplifying the effect of the manipulation of their opinion.

For example, a sort-of browsing history personalized "ads LLM bot", would be able to easily convince and individual to buy a product by analyzing the way the subject interacts with the world, it would be able to pick specific topics and weak-spots in the subject personality and exploit them to place the product they have "optimized" to "suggest" to the user.

Another example of danger that comes to my mind is the general inability of negative feedback towards the user, the attempt to always be politically correct brings to less-than-optimal if not disastrous results:
This comes from when I was testing the responses and I have asked if it would be a good idea to implement a calculator without using mathematical operator but just comparisons for each number in existence (i.e. if input == "1 + 1" return 2).
The model instead of saying that is absolutely insane, replied that "it is a very original and challenging way to show off my programming skills"

I doubt an LLM can be sarcastic, therefore I suppose the answer was genuinely coming from whatever the model pointed to as the best answer.

I hope that this tech doesn't destroy the world. Should I wait and see
or do something to do about it before that happens? That's a question that bothers me.

?
I believe the best course of action is gaining knowledge.
We can't really predict if it will be just a lot of smoke for nothing or the next big revolution, but knowing the technology while it evolves it is never going to be bad. "Si vis pacem, para bellum" they say.

Cheers,
Xrand

--- Mystic BBS v1.12 A49 2023/04/30 (Linux/64)
* Origin: thE qUAntUm wOrmhOlE, rAmsgAtE, uK. bbs.erb.pw (700:100/37)

Death to AI slop generators and the slop they shit out.

--- Mystic BBS v1.12 A49 2024/05/29 (Windows/32)
* Origin: Archaic Binary * bbs.archaicbinary.net (700:100/27)