Have things changed since 1955?

[jacks]

I was reading The End of Eternity by Isaac Asimov. (Great book by the way.) In it he makes the following statement about computers determining what is good.
“What is that? Your machines tell you. Your Computaplexes. But who adjusts the machines and tells them what to weigh in the balance? The machines do not solve problems with greater insight than men do, only faster. Only faster!”

Now he wrote this back in 1955, but my question is does it still hold true? That is can the only thing computers or AI can do is solve problems faster, than people. Or do they have greater insight?

 

[Mark Quayle]

I was reading The End of Eternity by Isaac Asimov. (Great book by the way.) In it he makes the following statement about computers determining what is good.
“What is that? Your machines tell you. Your Computaplexes. But who adjusts the machines and tells them what to weigh in the balance? The machines do not solve problems with greater insight than men do, only faster. Only faster!”

Now he wrote this back in 1955, but my question is does it still hold true? That is can the only thing computers or AI can do is solve problems faster, than people. Or do they have greater insight?

They can gather more data and look at (listen to) more points-of-view. But their insight is still bound by value weights that are not their own, and even if a supercomputer was, over many years of learning, able to assign their own value-weights, it would still be according to their programming and lacking what man has —intuition that transcends logic. I don't mean that intuition is better than logic —goodness knows we could all use logic when our intuition goes wrong!— but logic alone is a cold and inevitably mistaken pursuit because of false premises and a false mindset (programming), limited data (even for the most able supercomputer) and ultimately useless pursuits to search out.

 

[AlexB23]

I was reading The End of Eternity by Isaac Asimov. (Great book by the way.) In it he makes the following statement about computers determining what is good.
“What is that? Your machines tell you. Your Computaplexes. But who adjusts the machines and tells them what to weigh in the balance? The machines do not solve problems with greater insight than men do, only faster. Only faster!”

Now he wrote this back in 1955, but my question is does it still hold true? That is can the only thing computers or AI can do is solve problems faster, than people. Or do they have greater insight?

Nothing has changed, though AI is getting really good nowadays, and one can hold a conversation with an AI. GIGO: Garbage In, Garbage Out. AI is not capable of sentience, yet. Though, an AI might have more knowledge compared to a person in some areas, simply cos it was trained on so much data.

 

[jacks]

I tend to agree that the programming itself is a constraint on the ability of a computer to think freely. I thought perhaps that speed alone may lead to insight. i.e. if a computer can quickly assimilate and categorize information, it may come up with "new" ideas. However, I couldn't think of an example of this. I know the AI currently included in search engines is handy, but doesn't do anything more than conveniently assemble relevant posts.
Asimov was very prophetic about science in his writings.

 

[AlexB23]

I tend to agree that the programming itself is a constraint on the ability of a computer to think freely. I thought perhaps that speed alone may lead to insight. i.e. if a computer can quickly assimilate and categorize information, it may come up with "new" ideas. However, I couldn't think of an example of this. I know the AI currently included in search engines is handy, but doesn't do anything more than conveniently assemble relevant posts.
Asimov was very prophetic about science in his writings.

Agreed. We will wait and see where tech is in 2030

 

[AlexB23]

Guys, so in 2024, AI can be run privately and offline on laptops, and it is scary how good it is at analyzing verses. What are your thoughts on AI (when used with proper discernment from a human) to assist in writing devotionals? I try to use discernment, and add a personal anecdote when making devotionals using Mistral 7B. I would never trust an AI generated sermon though, or an AI priest/pastor. That is why I double check the AI on my computer.

Imagine where AI would be in 2035, which would be 80 years after 1955.

Hebrews Devotional Draft

 

[jacks]

What are your thoughts on AI (when used with proper discernment from a human) to assist in writing devotionals?

I think assisting humans in creating is a fine use of AI and something it can excel at. It simplifies the research process of creation. Whether it is writing or developing any other type of design, construction, etc. In this way it can be very helpful.
With your above example the interpretation of Bible passages is clearly just paraphrasing, the part that begins to look like "intelligence" is when it attempts to apply them to a different situation; Star Trek. So yeah, in a number of years, it may not reach self-awareness, but it's creativity will undoubtedly improve.

Perhaps the problem is the very term Artificial Intelligence. It is definitely "artificial" but "intelligence" applies an awareness or creative process that it currently lacks. Maybe something mundane like Computer Assisted Creativity would be more descriptive.

 

[Tuur]

Back in the day, AI referred to learning machines. These didn't have to be computers. Martin Gardner wrote about (and may have invented) a learning machine that played Hexpawn. You made diagrams of possible moves with colored arrows, put a diagram of each possible situation and moves on a matchbox, then beads the same color as the arrows inside. Hazy on the rules, but I think if the machine lost the game, you took away the bead representing that color from the box with the last move. Eventually this AI made from matchboxes would defeat humans after it "learned:" to play the game.

Neural nets work in a similar fashion, except with weights and layers, at least they did back in the day. You train the neural net by running inputs against desired outputs and let it adjust the weights until you get the desired results. Then you apply new data and see what it comes up with. One example of an AI image processing program turned up that it had "cheated" by incorporating the original image in the output. The machine didn't know it cheated; the neural nets only produced the desired results.

I don't think AI could thus give ups "insights" into data. They could run different scenarios and we can study the output, but you can do the same with traditional computer modeling. The only difference is that with traditional computer modeling, you are making assumptions about the process. With neural nets, it's trying to "learn" how to deliver the output from the input. It's not even making assumptions, only adjusting weights.

 

[jacks]

Back in the day, AI referred to learning machines. These didn't have to be computers. Martin Gardner wrote about (and may have invented) a learning machine that played Hexpawn. You made diagrams of possible moves with colored arrows, put a diagram of each possible situation and moves on a matchbox, then beads the same color as the arrows inside. Hazy on the rules, but I think if the machine lost the game, you took away the bead representing that color from the box with the last move. Eventually this AI made from matchboxes would defeat humans after it "learned:" to play the game.

Neural nets work in a similar fashion, except with weights and layers, at least they did back in the day. You train the neural net by running inputs against desired outputs and let it adjust the weights until you get the desired results. Then you apply new data and see what it comes up with. One example of an AI image processing program turned up that it had "cheated" by incorporating the original image in the output. The machine didn't know it cheated; the neural nets only produced the desired results.

I don't think AI could thus give ups "insights" into data. They could run different scenarios and we can study the output, but you can do the same with traditional computer modeling. The only difference is that with traditional computer modeling, you are making assumptions about the process. With neural nets, it's trying to "learn" how to deliver the output from the input. It's not even making assumptions, only adjusting weights.

Very interesting. Knowing how AI actually works would be very helpful in answering my original question. i.e. can AI provide greater insight than a human. I have to admit I've only looked at the results of AI programing and have no idea what that programming actually entails. Something to explore!

 

[AlexB23]

I think assisting humans in creating is a fine use of AI and something it can excel at. It simplifies the research process of creation. Whether it is writing or developing any other type of design, construction, etc. In this way it can be very helpful.
With your above example the interpretation of Bible passages is clearly just paraphrasing, the part that begins to look like "intelligence" is when it attempts to apply them to a different situation; Star Trek. So yeah, in a number of years, it may not reach self-awareness, but it's creativity will undoubtedly improve.

Perhaps the problem is the very term Artificial Intelligence. It is definitely "artificial" but "intelligence" applies an awareness or creative process that it currently lacks. Maybe something mundane like Computer Assisted Creativity would be more descriptive.

Agreed. I just used AI today to find the Dewey number for astronomy (523). Yes, AI paraphrases stuff, and processes stuff in a way we do not know fully yet.

 

[Diamond72]

Now he wrote this back in 1955, but my question is does it still hold true? That is can the only thing computers or AI can do is solve problems faster, than people. Or do they have greater insight?

AI can answer that question for you.

1. Speed of Computation:
   • Computers and AI excel at processing vast amounts of data and performing complex calculations rapidly. They can analyze patterns, recognize correlations, and make decisions in milliseconds, tasks that would take humans significantly longer or might be impossible due to computational limitations1.
   • For instance, IBM’s Watson can read and remember extensive research on cancer, providing insights that no human could match in terms of speed2.
2. Insight and Creativity:
   • While computers excel at processing information, human insight and creativity remain unparalleled. Humans can draw from intuition, empathy, and contextual understanding to solve complex problems.
   • Insight involves connecting disparate pieces of information, understanding nuance, and making leaps of logic. Creativity allows us to generate novel solutions, think outside the box, and adapt to new situations.
   • AI lacks true consciousness and creativity. It operates based on predefined algorithms and data patterns, without genuine understanding or original thought.

 

[jacks]

AI can answer that question for you.

1. Speed of Computation:
   • Computers and AI excel at processing vast amounts of data and performing complex calculations rapidly. They can analyze patterns, recognize correlations, and make decisions in milliseconds, tasks that would take humans significantly longer or might be impossible due to computational limitations1.
   • For instance, IBM’s Watson can read and remember extensive research on cancer, providing insights that no human could match in terms of speed2.
2. Insight and Creativity:
   • While computers excel at processing information, human insight and creativity remain unparalleled. Humans can draw from intuition, empathy, and contextual understanding to solve complex problems.
   • Insight involves connecting disparate pieces of information, understanding nuance, and making leaps of logic. Creativity allows us to generate novel solutions, think outside the box, and adapt to new situations.
   • AI lacks true consciousness and creativity. It operates based on predefined algorithms and data patterns, without genuine understanding or original thought.

It's refreshing to have something that admits to their weaknesses. This makes me like AI even more.

I guess Isaac Asimov still remains correct!