From Newsgroup: comp.programming

Hey, I am a millionaire.... Would You want to get rich too ?
On Wednesday, January 4, 2023 at 4:53:54 AM UTC+2, Amine Moulay Ramdane wrote:

Hello,

More of my philosophy about the other weakness of artificial intelligence such as ChatGPT and more of my thoughts..

I am a white arab from Morocco, and i think i am smart since i have also invented many scalable algorithms and algorithms..

I think i am highly smart since I have passed two certified IQ tests and i have scored "above" 115 IQ, and i mean that it is "above" 115 IQ,
i think that the other weakness of artificial intelligence such as ChatGPT is that when in mathematics you have a differentiable function
, you can find the derivative of it and be exact, but artificial intelligence such as ChatGPT is not this kind of exactness, since it approximates more and more by being trained on the data , so the quality is dependent on the data on wich artificial intelligence is trained, other than that how to be sure that the solution given by ChatGPT is exact? or how to be sure of the quality of the solutions given by ChatGPT?, since for example ChatGPT and even GPT-4 is trained on a data from internet that is of a variable quality, but still
ChatGPT or GPT-4 can become really powerful and really useful.

And of course the genetic algorithm has the same problem,
read my following thoughts about it so that you understand my views:

https://groups.google.com/g/alt.culture.morocco/c/YwW7mgTaJOw

More of my philosophy about sex and about artificial intelligence and more of my thoughts..

I think that sex is not a sophisticated thing, since it is so easy to have sex and you can easily waste your too much time with sex and you can easily become stupid with sex and such, so sex is not really efficient, but i think that artificial intelligence like ChatGPT can make you really efficient since it is efficient, so i think we have not to waste our time with sex, but we have to learn how to be efficient with such artificial intelligence tools as ChatGPT so that to be much more smart so that to "adapt" efficiently, but of course so that you understand my views i invite you to read all my following thoughts:

More of my philosophy of does ChatGPT has a kind of common sense reasoning and more of my thoughts..

So today i will ask a question of: Does ChatGPT has some kind of common sense reasoning ? i think it has a kind common sense reasoning since it
can logically infer it from the patterns of the data on wich it has been trained on, and even if it is limited by the smartness of the data on wich it has been trained, i think that ChatGPT can become really powerful and really useful, and notice on my following interactions with ChatGPT how it has logically inferred a kind of common sense reasoning from the patterns of the data on wich it has been trained on, here it is and read it carefully so that to notice:

I have asked ChatGPT the following:

If a car goes from Casablanca to Rabat and if it goes from Rabat to Fez , how many kilometers does the car has traveled ?

ChatGPT has answered:

"I'm sorry, but I don't have enough information to accurately determine how many kilometers the car has traveled. The distance between Casablanca and Rabat is about 95 kilometers, and the distance between Rabat and Fez is about 210 kilometers. However, I don't know the exact route the car took, so it is possible that it traveled a greater distance. Additionally, I don't know if the car made any other trips beyond the ones you mentioned. Without this information, it is not possible to accurately determine how many kilometers the car has traveled."

And i have answered ChatGPT:

"Just give an approximation"

And ChatGPT has answered:

"Based on the information you have provided, it is possible that the car traveled a distance of around 305 kilometers, if it went directly from Casablanca to Rabat and then from Rabat to Fez. However, it is important to note that this is just an approximation and the actual distance traveled could be different, depending on the exact route the car took and any other trips it may have made."

More of my philosophy about is ChatGPT going to replace jobs and about the law of accelerating returns and more of my thoughts..

I have just looked at the following new video about ChatGPT with GPT4,
and i invite you to look at it:

This A.I. will Change the World Forever

https://www.youtube.com/watch?v=w7KfELRxc3E

I think i am highly smart since I have passed two certified IQ tests and i have scored "above" 115 IQ, and i mean that it is "above" 115 IQ,
so i am not in accordance with the above video, since i can do predictions about ChatGPT, so is it going to replace many jobs ? so i will say that the problem of the quality of data on wich ChatGPT is trained is not solved, i mean that ChatGPT even with GPT4 is trained on enormous amounts of text of variable quality from the internet, so then it causes that we can not be sure about the quality of ChatGPT, so it is why i don't think that ChatGPT will replace jobs as is saying it the above video, other than that, ChatGPT can cause problems to the value of products and services since ChatGPT makes creativity much more easy and this hurts the value as i am saying it below, the other problem that ChatGPT creates is the one of the law of accelerating returns, so i mean that those that are "apt" to use efficiently ChatGPT will advance so rapidly and in an accelerating returns way, so this will create much more inequality, other than that , if you want to understand the weaknesses of ChatGPT even with GPT4, you can read my following previous thoughts:

More of my philosophy about supervised and unsupervised learning in artificial intelligence and more of my thoughts..

I have to explain something important, so GPT of ChatGPT uses a combination of unsupervised pre-training and supervised fine-tuning,
so i think that the patterns are discovered from the data in the unsupervised pre-training, so it means that ChatGPT doesn't for example know about what is a verb or an adverb, but it is just the doing calculations of finding the patterns that make GPT recognize what is a verb or an adverb, so as you notice that this process of recognizing by GPT is not the understanding the "meaning", but it is just discovering patterns with calculations in the data that make you recognize the things.

More of my philosophy about creativity of artificial intelligence such as ChatGPT and more of my thoughts..

I invite you to read the following article that says the following
about artificial intelligence such as ChatGPT:

"Stakeholders in the IT development sector are nevertheless of the opinion that artificial intelligence is only good for automating redundant tasks and will not go beyond this level. The development of specifications or specifications for large IT projects remains a task that requires the creativity of a human brain that an artificial intelligence will never have."

Read more in the following web link (and you have to translate the article from french to english):

AI startup predicts 'relying on humans to write programs will be outdated in the future'

https://intelligence-artificielle.developpez.com/actu/339951/Une-startup-specialisee-en-IA-predit-que-s-appuyer-sur-des-humains-pour-ecrire-des-programmes-sera-depasse-a-l-avenir-et-ravive-le-debat-sur-la-possible-disparition-du-metier-de-developpeur/

So i think that the above saying from Stakeholders in the IT development sector is not so correct, since artificial intelligence such as ChatGPT
can be "creative" by also composing patterns that it discovers from the data it has been trained on, so i say that language models like ChatGPT are as good as the data they have been trained on, they are simply programmed to generate text based on patterns in the data they have been trained on. But i invite you to read about the weaknesses of artificial intelligence such as ChatGPT in my following previous thoughts:

More of my philosophy about the weaknesses of ChatGPT and about
artificial intelligence and more of my thoughts..

I think i am highly smart since I have passed two certified IQ tests and i have scored "above" 115 IQ, and i mean that it is "above" 115 IQ, and I think that ChatGPT has been trained not on only free data but on Common Crawl (filtered) and WebText2 and Book1 and Book2 and Wikipedia, other than that i say that language models like ChatGPT are as good as the data they have been trained on, they are simply programmed to generate text based on patterns in the data they have been trained on. So i think i am discovering the pattern with my fluid intelligence that explains the weakness of Generative Pre-trained Transformer of like ChatGPT, and it is that ChatGPT can discover the patterns using the existing patterns from the data or knowledge, so it is like using the smartness of the data, but ChatGPT can not use the smartness of the human brain that also comes with human consciousness that optimizes more smartness, so it can not invent highly smart patterns or things from like is doing it a highly smart human from his brain, since i think that for example human consciousness creates the "meaning" that optimizes more smartness and enhance smartness, but ChatGPT has no meaning like with human consciousness, but it is like calculations on the data so that to find the patterns done by being limited by the smartness of the data, and even the generative side of ChatGPT is limited by the smartness of the data, so i think that ChatGPT will still be not capable of this kind of highly smart creativity, but still it remains really powerful and really useful, so the next step is to crack the code of human consciousness and understand how human consciousness is made so that to give "meaning" that optimizes more smartness and enhance smartness, since if the smart pattern to be discovered has an unknown territory by lack of meaning from human consciousness to the smartness of the data that has to be discovered by using the meaning from human consciousness, the meaning from human consciousness permit to construct the smartness that can discover the unknown territory by lack of meaning from the smartness of the data , i mean that discovering patterns from the data is less sophisticated smartness as discovering them using the meaning from the human consciousness, since the meaning from human consciousness permits to decrypt or understand from the low level to the high level when the pattern is unknown to the smartness of the data by lack of meaning from human consciousness, so i invite you to read my following previous thoughts that make you understand my views:

Here is my new proverb about artificial intelligence such as ChatGPT:

"The most important in artificial intelligence is not to make it performant, but it is to make it in accordance with the legal system, since if you make artificial intelligence such as ChatGPT too much performant too much rapidly , it can hurt the system and make it too much unstable, so the most important in a legal system is to balance correctly the performance with the being stability or being order."

So i invite you to read my below thoughts of how to regulate ChatGPT
so that you notice that the problem is that ChatGPT is too much performant too much rapidly, so you have to know how to correctly balance the performance with the being stability or being order, so here is my
thoughts and reread them carefully so that you understand my views:

More of my philosophy about how to regulate ChatGPT and artificial intelligence and more of my thoughts..

So i think that OpenAI is not thinking correctly, since the ChatGPT
is taking the patterns from the data of creativity of humans and is using them, so i don't think that it has the right to do that, since it is "clear" in our legal system that humans that have also created the patterns of there creativity have not given the rights to make an artificial intelligence such as ChatGPT with them that can hurt humans in a not lawful way, since the value of a product or service comes from smartness and/or from work or the hard work, but artificial intelligence tools such as ChatGPT are hurting in a not lawful way humans and are hurting in a not lawful way the value of creativity of a product or service that comes from smartness and/or work or hard work of a human, since the fact that ChatGPT makes it much more easy to be creative , so this also hurts the value, and i invite you to read my previous thoughts so that to understand my views:

More of my philosophy about the weakness of Generative Pre-trained Transformer and more of my thoughts..

So i think i am discovering the pattern with my fluid intelligence that explains the weakness of Generative Pre-trained Transformer of like ChatGPT, and it is that ChatGPT can discover the patterns using
the existing patterns from the data or knowledge, so it is like using
the smartness of the data, but ChatGPT can not use the smartness of the human brain that also comes with human consciousness that optimizes more smartness, so it can not invent highly smart patterns or things from like is doing it a highly smart human from his brain, so i think that ChatGPT will still be not capable of this kind of highly smart creativity, but still it remains really powerful and really useful, so i invite you to read my following previous thoughts that make you understand my views:

More precision of my philosophy about the mechanisms of attention and self-attention of Transformers AI models and more of my thoughts..

I think i am highly smart since I have passed two certified IQ tests and i have scored "above" 115 IQ, and i mean that it is "above" 115 IQ, i think i am understanding deep learning, but i say that Transformers are deep learning + self-attention and attention, and this attention and self-attention permit to grasp "context" and "antecedents", for example when you say the following sentence:

"The animal didn't cross the street because it was too tired"

So we can ask how the artificial intelligence of ChatGPT that uses Generative Pre-trained Transformer will understand that the "it" in
the above sentence is not the street but the animal, so i say that
it is with self-attention and attention mechanisms of artificial intelligence and with the training with more and more data that the transformer can "detect" the pattern of the "it" refers to the "animal" in the above sentence, so self-attention and attention of the artificial intelligence of ChatGPT that we call Generative Pre-trained Transformer permit to grasp "context" and "antecedents" too, it is also like logically inferring the patterns using self-attention and attention from the context of the many many sentences from the data, and since the data is exponentially growing and since the artificial intelligence is also generative, so i think it will permit to make the artificial intelligence of the transformer much more powerful, so as you notice that the data is King , and the "generative" word of the Generative Pre-trained Transformer refers to the model's ability to generate text, and of course we are now noticing that it is making ChatGPT really useful and powerful, and of course i say that ChatGPT will still much more improve , and read my following previous thoughts so that to understand my views about it:

More of my philosophy about transformers and about the next GPT-4 and about ChatGPT and more of my thoughts..

The capabilities of transformer architectures, as in GPT of ChatGPT that is called Generative Pre-trained Transformer, are truly remarkable, as they allow machine learning models to surpass human reading comprehension and cognitive abilities in many ways. These models are trained on massive amounts of text data, including entire corpora such as the English Wikipedia or the entire internet, which enables them to become highly advanced language models (LMs) with a deep understanding of language and the ability to perform complex predictive analytics based on text analysis. The result is a model that is able to approximate human-level text cognition, or reading, to an exceptional degree - not just simple comprehension, but also the ability to make sophisticated connections and interpretations about the text, because the network of Transformer pay “attention” to multiple sentences, enabling it to grasp "context" and "antecedents". These transformer models represent a significant advancement in the field of natural language processing and have the potential to revolutionize how we interact with and understand language.

GPT-4 is significantly larger and more powerful than GPT-3, with 170 trillion parameters compared to GPT-3’s 175 billion parameters(and even GPT-3.5 of the new ChatGPT has 175 billion parameters). This allows GPT-4 to process and generate text with greater accuracy and fluency, so with feedback from users and a more powerful GPT-4 model coming up and by being trained on a substantially larger amount of data , ChatGPT that will use GPT-4 may "significantly" improve in the future. So i think ChatGPT will still become much more powerful. And i invite you to read my previous thoughts about my experience with the new ChatGPT:

More of my philosophy about my experience with ChatGPT and about artificial intelligence and more of my thoughts..

I think i am highly smart since I have passed two certified IQ tests and i have scored "above" 115 IQ, and i mean that it is "above" 115 IQ,
so in those two last days i have just tested ChatGPT so that to see if
this new artificial intelligence launched by OpenAI in November 2022 is efficient, and i think that it is really useful, since i think by testing it that it can score well on the human average smartness, but if you want it to be highly smart by inventing highly smart things , it will not be able to do it, but if you want ChatGPT to be highly smart on what it has learned from the existing smartness of the human knowledge that it has been trained on, i think it can also score high in many times of it, also ChatGPT can in many times make much less errors than humans, so i think that ChatGPT is really useful, and i think that ChatGPT will be improved much more by increasing the size of its transformer (A transformer is a deep learning model that adopts the mechanism of self-attention) , and i also think that ChatGPT will be improved much more when it will be trained on a substantially larger amount of data, considering an article that DeepMind just published a few days ago demonstrating that the performance of these models can be drastically improved by scaling data more aggressively than parameters ( Read it here: https://arxiv.org/pdf/2203.15556.pdf ), and it is
why i am optimistic about the performance of ChatGPT and i think that it will be much more improved.

Thank you,
Amine Moulay Ramdane.

--- Synchronet 3.20a-Linux NewsLink 1.113