ChatGPT - artifical intelligence chatbot

[mishra]

Was put on a 10 day crash course on Generative AI and machine learning. From what I understood, Right now AI is just as good as answering a question from a comprehension. Just that comprehension can be too vast to not introduce errors in answers.

[Real McCoy]

5 hours ago, mishra said:

Was put on a 10 day crash course on Generative AI and machine learning. From what I understood, Right now AI is just as good as answering a question from a comprehension. Just that comprehension can be too vast to not introduce errors in answers.

My experience is with ChatGPT, the free version. It is quite good at comprehension. When it comes to complex topics, it may produce errors in answers. Also if you ask a human an open ended question, the answers will be several and each human is bound to give different answers based on his upbringing and mentality. If you as the same question to ChatGPT, the answer will be the same written with different phrases, everytime you repeat the same question like "Can you think of anything else apart from the last answer you typed?".

 

I see AI will be successful in doing repetitive tasks like computers only so much powerful. The only way it can reach humans in ingenuity is a sinister idea, known by many scientists. It is not ethical by any means and anyone who indulges in it is doomed forever

Edited April 8 by Real McCoy

[mishra]

4 hours ago, Real McCoy said:

My experience is with ChatGPT, the free version. It is quite good at comprehension. When it comes to complex topics, it may produce errors in answers. Also if you ask a human an open ended question, the answers will be several and each human is bound to give different answers based on his upbringing and mentality. If you as the same question to ChatGPT, the answer will be the same written with different phrases, everytime you repeat the same question like "Can you think of anything else apart from the last answer you typed?".

 

I see AI will be successful in doing repetitive tasks like computers only so much powerful. The only way it can reach humans in ingenuity is a sinister idea, known by many scientists. It is not ethical by any means and anyone who indulges in it is doomed forever

There is nothing intelligent in AI. To put it simple, think it as  a search engine based on trained data and open to take up new data.

[BacktoCricaddict]

5 hours ago, mishra said:

There is nothing intelligent in AI. To put it simple, think it as  a search engine based on trained data and open to take up new data.

There's got to be a good fundamental definition of the word intelligence. But it's hard to find.

 

All I can do is wish I had a lot more of it than I do now.

Edited April 9 by BacktoCricaddict

[Real McCoy]

7 hours ago, mishra said:

There is nothing intelligent in AI. To put it simple, think it as  a search engine based on trained data and open to take up new data.

 

Yeah AI is fed insane amounts of data. The humans who have compiled "facts" will introduce their own biases to the algorithm. Real intelligence is not based on statistics and microlevel data management. AI will never possess critical thinking as it involves something natural. Science will never get hold of the original reason of intelligence.

[Vijy]

On 4/8/2024 at 8:12 PM, BacktoCricaddict said:

There's got to be a good fundamental definition of the word intelligence. But it's hard to find.

 

All I can do is wish I had a lot more of it than I do now.

Doesn't exist. I attended a conference involving computer scientists (AI), psychologists, neuroscientists, doctors, ethologists (animal behavior specialists), cognitive scientists, and so on. Each had their own lingo, their own definition, etc

[bharathh]

On 4/9/2024 at 5:42 AM, BacktoCricaddict said:

There's got to be a good fundamental definition of the word intelligence. But it's hard to find.

 

All I can do is wish I had a lot more of it than I do now.

IMO, all intelligence (human or otherwise) is based on 2 things:

recognizing a pattern in a stimulus,

and,

responding to it based on past knowledge or applying past knowledge to on a new response that will most likely be successful

 

How quickly and successfully one can respond to the stimulus is the measure of intelligence. IMO, stays the same whether it is organic or artificial.

[BacktoCricaddict]

9 hours ago, Vijy said:

Doesn't exist. I attended a conference involving computer scientists (AI), psychologists, neuroscientists, doctors, ethologists (animal behavior specialists), cognitive scientists, and so on. Each had their own lingo, their own definition, etc

 

7 hours ago, bharathh said:

IMO, all intelligence (human or otherwise) is based on 2 things:

recognizing a pattern in a stimulus,

and,

responding to it based on past knowledge or applying past knowledge to on a new response that will most likely be successful

 

How quickly and successfully one can respond to the stimulus is the measure of intelligence. IMO, stays the same whether it is organic or artificial.

 

On 4/8/2024 at 3:05 PM, mishra said:

There is nothing intelligent in AI. To put it simple, think it as  a search engine based on trained data and open to take up new data.

 

Would we consider a champion chess player intelligent? In my book, absolutely yes. No doubts whatsoever. Which makes computers that beat champion chess players extremely intelligent. 

 

Perhaps we tend to confuse intelligence with wisdom, and on some level, AI and generative AI may lack that at this time. In 10 years, who knows?

 

 

 

 

 

[gattaca]

13 hours ago, BacktoCricaddict said:

 

 

 

Would we consider a champion chess player intelligent? In my book, absolutely yes. No doubts whatsoever. Which makes computers that beat champion chess players extremely intelligent. 

 

Perhaps we tend to confuse intelligence with wisdom, and on some level, AI and generative AI may lack that at this time. In 10 years, who knows?

 

 

 

 

 

But beating player is about computations. AI excels at that. It finds all possibilities and picks a move based on best probability of winning. Our brain is limited in that space. There is also time once has to make a move in time. Wisdom is not one dimensional though. Multiple variables but again once feed data it can take in to account all the variables as well after all humans were apes. We did learn lot of things through imitation and still do. But a machine in the end will take decisions in its favor when it comes to survival just humans do. It will be us vs them at one point.

Edited April 11 by gattaca

[bharathh]

56 minutes ago, gattaca said:

But beating player is about computations. AI excels at that. It finds all possibilities and picks a move based on best probability of winning. Our brain is limited in that space. There is also time once has to make a move in time. Wisdom is not one dimensional though. Multiple variables but again once feed data it can take in to account all the variables as well after all humans were apes. We did learn lot of things through imitation and still do. But a machine in the end will take decisions in its favor when it comes to survival just humans do. It will be us vs them at one point.

The brain has far more capability than any supercomputer today. It's just that we don't know how to leverage it. 

[BacktoCricaddict]

On 4/10/2024 at 10:30 PM, gattaca said:

But beating player is about computations. AI excels at that. It finds all possibilities and picks a move based on best probability of winning. Our brain is limited in that space. There is also time once has to make a move in time. Wisdom is not one dimensional though. Multiple variables but again once feed data it can take in to account all the variables as well after all humans were apes. We did learn lot of things through imitation and still do. But a machine in the end will take decisions in its favor when it comes to survival just humans do. It will be us vs them at one point.

I may be totally off on this topic.

 

But, to me, the bold-faced part is intelligence defined - the ability to quickly find various possibilities, know which one has the best probability of winning and pick it rapidly. An excellent human chess player's brain does the same thing much better than the average human. It's the same thing with mathematicians, I'd imagine. Some people's brains can go through abstract computations better than the average person. And that represents intelligence.

 

All of that, of course, is different from wisdom.

[BacktoCricaddict]

On 4/10/2024 at 11:28 PM, bharathh said:

The brain has far more capability than any supercomputer today. It's just that we don't know how to leverage it. 

[Real McCoy]

With all the intelligence it possesses, it still cannot look after itself and needs a human to feed it both physically and intellectually. AI doesn't have mind much less a soul. Humans and other lifeforms do. This is part of our tradition. Our rishis were more intelligent than us and they didn't walk around like big shots. Nowadays everyone is a big shot because they can play chess and some board games.

[BacktoCricaddict]

https://www.quantamagazine.org/how-do-machines-grok-data-20240412/

Quote

For all their brilliance, artificial neural networks remain as inscrutable as ever. As these networks get bigger, their abilities explode, but deciphering their inner workings has always been near impossible. Researchers are constantly looking for any insights they can find into these models. A few years ago, they discovered a new one.

 

In January 2022, researchers at OpenAI, the company behind ChatGPT, reported that these systems, when accidentally allowed to munch on data for much longer than usual, developed unique ways of solving problems. Typically, when engineers build machine learning models out of neural networks — composed of units of computation called artificial neurons — they tend to stop the training at a certain point, called the overfitting regime. This is when the network basically begins memorizing its training data and often won’t generalize to new, unseen information. But when the OpenAI team accidentally trained a small network way beyond this point, it seemed to develop an understanding of the problem that went beyond simply memorizing — it could suddenly ace any test data.

 

The researchers named the phenomenon “grokking,” a term coined by science-fiction author Robert A. Heinlein to mean understanding something “so thoroughly that the observer becomes a part of the process being observed.” The overtrained neural network, designed to perform certain mathematical operations, had learned the general structure of the numbers and internalized the result. It had grokked and become the solution.

Not knowing the first thing about how all this works, should I be scared or excited or ... a bit of both?

[diga]

2 minutes ago, BacktoCricaddict said:

https://www.quantamagazine.org/how-do-machines-grok-data-20240412/

Not knowing the first thing about how all this works, should I be scared or excited or ... a bit of both?

Skynet.. judgement day is coming beware

[gattaca]

1 hour ago, BacktoCricaddict said:

https://www.quantamagazine.org/how-do-machines-grok-data-20240412/

Not knowing the first thing about how all this works, should I be scared or excited or ... a bit of both?

It is dangerous. I use it a work. Every time I use it the next time it is better. I use it because I get help to finish my work faster but in the end it is dangerous 

[Real McCoy]

1 hour ago, gattaca said:

It is dangerous. I use it a work. Every time I use it the next time it is better. I use it because I get help to finish my work faster but in the end it is dangerous 

 

Its only dangerous as it will reduce your ability to search for whatever you are looking for, loose your brain cells because the answer is there on demand. Its even harder for children who are raised with this technology. Many people don't ask for directions these days because google maps is there. Getting to something faster damages things in the long run but that's where the world is headed.

[BacktoCricaddict]

1 hour ago, gattaca said:

It is dangerous. I use it a work. Every time I use it the next time it is better. I use it because I get help to finish my work faster but in the end it is dangerous 

 

As a bit of a dud with computer-based technology (coding etc.), I really wish I had the capacity to learn more about computers and actually understand it.

 

I mean, to me, even something as ubiquitous as a keystroke appearing on a screen as a character - is extremely fascinating but shrouded in mystery. What are the different steps in that flow of information? What are all the physical components involved? How exactly does a code - at the very very basic level - turn into action?

 

So much to learn, so little bandwidth!  

 

 

 

 

 

[gattaca]

1 hour ago, Real McCoy said:

 

Its only dangerous as it will reduce your ability to search for whatever you are looking for, loose your brain cells because the answer is there on demand. Its even harder for children who are raised with this technology. Many people don't ask for directions these days because google maps is there. Getting to something faster damages things in the long run but that's where the world is headed.

I have experienced what they call is hallucinations from AI. It confidently answers but if you ask more deep questions it knows it got it wrong and apologizes. I would never trust it but always verify it.It still simplifies your work sometimes.

[Real McCoy]

1 minute ago, gattaca said:

I have experienced what they call is hallucinations from AI. It confidently answers but if you ask more deep questions it knows it got it wrong and apologizes. I would never trust it but always verify it.It still simplifies your work sometimes.

That could be because it probably is trained on comparing its own answers, finding a flaw, and reporting it back to you. It does simplify work but as you mentioned, you can never trust it. Human oversight is needed.