Dan’s review of The Closing of the Western Mind: The Rise of Faith and the Fall of Reason > Likes and Comments
6 likes · Like
I understand what Platonism means in mathematics. But what does it mean in AI?
Liedzeit wrote: "I understand what Platonism means in mathematics. But what does it mean in AI?"
It postulates that reality is nothing else but mathematics (i.e., numbers are the building blocks of reality) and that programming is the way to get to this reality and not thinking or similar. For example, Max Tegmark in his “Our Mathematical Universe” elaborated this - along with his understanding of physics and of the AI project. These days the AI project got derailed by the craziness of these LLM models; but I expect that once they realize that there is not much future into them, the scientists will go back to more serious and fundamental projects where some kind of ontology is needed again - and a Platonic ontology is the most obvious, practical, and promising for them. But we will see :)
Really? You mean an ontology that states that only ideas are real? No wonder you have to believe that LLMs are craziness. Because if they‘re not idealism is disproved. 🤓
Liedzeit wrote: "Really? You mean an ontology that states that only ideas are real? No wonder you have to believe that LLMs are craziness. Because if they‘re not idealism is disproved. 🤓"
LLM are just language models. Since they assume nothing about reality, they have nothing to do with idealism, Platonism, materialism, and with reality in general; that is, there is no ontology behind them.
Before LLMs, people like Tegmark assumed that reality is fundamentally numbers; and this goes back to Plato. Tegmark explicitly defines himself as Platonist. They did not build any AI; but I expect that the next generation of AI will drop LLMs and will return to some ontology; and again I expect that this ontology will be a Platonic one. But we will see.
According to Plato we live in a dualistic. People need a soul that allows them access to the real world (of forms) by amnesis. Since an LLM is a purely materialistic thing it means it has no soul. So if it can produce ideas no dualism is needed and the whole Platonic world collapses.
Liedzeit wrote: "According to Plato we live in a dualistic. People need a soul that allows them access to the real world (of forms) by amnesis. Since an LLM is a purely materialistic thing it means it has no soul. ..."
LLMs are not either materialistic or idealistic things; they are technological things. LLMs do not produce ideas; they just simply re-arrange language as triggered. The companies creating the LLMs present them as if they create new ideas - but this is just old marketing. Seeing them as the first signs of the arrival of some technological-god, is just old Christianity.
As far as I am concerned, the entire materialistic/idealistic structure is bogus - in its dualistic or idealistic or materialistic form. The problem here is that modern science depends on this structure; and consequently stresses the importance of either one of the 3: dualistic, materialistic, or idealistic. Recently, scientists rely more and more on the idealistic structure. If they are capable of questioning their own foundations - and they are rarely capable of something like this - they acknowledge it. Tegmark did it - and I mentioned him because he is involved in current and advanced mathematics, fundamental physics, and AI research.
Well, you brought up Plato’s ontology. I just said that if LLMs had a mind it would make the concept of a soul obsolete. And dualism and idealism both ground on the belief that mind cannot be something materialistic. You may be absolutely right that LLMs do not have a mind and can never get one. I just happen to think you might not be.
Liedzeit wrote: "Well, you brought up Plato’s ontology. I just said that if LLMs had a mind it would make the concept of a soul obsolete. And dualism and idealism both ground on the belief that mind cannot be somet..."
Liedzeit; there is no way we will agree on the nature of LLMs. I also know that we have our differences regarding materialism/idealism/dualism :). The second issue will probably never be resolved in a definitive way. However, I hope that the first may be settled in the next few years - that is when the LLMs hit the proverbial wall, people start abandoning them as useless, and the AI community will restart research in fundamental and serious directions and away from LLMs.
You are right, Dan. There is definitely a chance that we will agree on LLMs in the not too distant future. 🙂 I am trying to write a book with the help of Chat at the moment and it is both a joyful and a very frustrating experience.
back to top
date
newest »
newest »
message 1:
by
Liedzeit
(new)
Jan 24, 2026 06:49AM
I understand what Platonism means in mathematics. But what does it mean in AI?
reply
|
flag
Liedzeit wrote: "I understand what Platonism means in mathematics. But what does it mean in AI?"It postulates that reality is nothing else but mathematics (i.e., numbers are the building blocks of reality) and that programming is the way to get to this reality and not thinking or similar. For example, Max Tegmark in his “Our Mathematical Universe” elaborated this - along with his understanding of physics and of the AI project. These days the AI project got derailed by the craziness of these LLM models; but I expect that once they realize that there is not much future into them, the scientists will go back to more serious and fundamental projects where some kind of ontology is needed again - and a Platonic ontology is the most obvious, practical, and promising for them. But we will see :)
Really? You mean an ontology that states that only ideas are real? No wonder you have to believe that LLMs are craziness. Because if they‘re not idealism is disproved. 🤓
Liedzeit wrote: "Really? You mean an ontology that states that only ideas are real? No wonder you have to believe that LLMs are craziness. Because if they‘re not idealism is disproved. 🤓"LLM are just language models. Since they assume nothing about reality, they have nothing to do with idealism, Platonism, materialism, and with reality in general; that is, there is no ontology behind them.
Before LLMs, people like Tegmark assumed that reality is fundamentally numbers; and this goes back to Plato. Tegmark explicitly defines himself as Platonist. They did not build any AI; but I expect that the next generation of AI will drop LLMs and will return to some ontology; and again I expect that this ontology will be a Platonic one. But we will see.
According to Plato we live in a dualistic. People need a soul that allows them access to the real world (of forms) by amnesis. Since an LLM is a purely materialistic thing it means it has no soul. So if it can produce ideas no dualism is needed and the whole Platonic world collapses.
Liedzeit wrote: "According to Plato we live in a dualistic. People need a soul that allows them access to the real world (of forms) by amnesis. Since an LLM is a purely materialistic thing it means it has no soul. ..."LLMs are not either materialistic or idealistic things; they are technological things. LLMs do not produce ideas; they just simply re-arrange language as triggered. The companies creating the LLMs present them as if they create new ideas - but this is just old marketing. Seeing them as the first signs of the arrival of some technological-god, is just old Christianity.
As far as I am concerned, the entire materialistic/idealistic structure is bogus - in its dualistic or idealistic or materialistic form. The problem here is that modern science depends on this structure; and consequently stresses the importance of either one of the 3: dualistic, materialistic, or idealistic. Recently, scientists rely more and more on the idealistic structure. If they are capable of questioning their own foundations - and they are rarely capable of something like this - they acknowledge it. Tegmark did it - and I mentioned him because he is involved in current and advanced mathematics, fundamental physics, and AI research.
Well, you brought up Plato’s ontology. I just said that if LLMs had a mind it would make the concept of a soul obsolete. And dualism and idealism both ground on the belief that mind cannot be something materialistic. You may be absolutely right that LLMs do not have a mind and can never get one. I just happen to think you might not be.
Liedzeit wrote: "Well, you brought up Plato’s ontology. I just said that if LLMs had a mind it would make the concept of a soul obsolete. And dualism and idealism both ground on the belief that mind cannot be somet..."Liedzeit; there is no way we will agree on the nature of LLMs. I also know that we have our differences regarding materialism/idealism/dualism :). The second issue will probably never be resolved in a definitive way. However, I hope that the first may be settled in the next few years - that is when the LLMs hit the proverbial wall, people start abandoning them as useless, and the AI community will restart research in fundamental and serious directions and away from LLMs.
You are right, Dan. There is definitely a chance that we will agree on LLMs in the not too distant future. 🙂 I am trying to write a book with the help of Chat at the moment and it is both a joyful and a very frustrating experience.
