AI learnt "something" from the Physical & Life Sciences Forum.

[AlexB23]

GPT-4o claimed the rebounding mechanism was caused by infalling gases colliding into the incompressible core.
This sudden deacceleration created the shockwaves leading to the supernova.

This is not the primary mechanism the rebound of the core due the Pauli exclusion principle which GPT-4o eventually acknowledged.

View attachment 353567

I still do not put 100% trust in AI, it does a good job at putting pseudoscience ideas to the sword such as windmills causing global climate change but it can still fall short with the nuances of mainstream science.

Yeah, I do not trust AI fully yet. It does not have the capability to understand the nuances of science. Sometimes, it is best to use retrieval augmented generation, to get rid of any hallucinations.

 

[AlexB23]

@sjastro , Mistral 7B tried, and it fell flat on the face. Only 7 billion parameters, it is not good for high level science. At least the AI is all run locally on my laptop. It was also much slower than I wanted it to be.

 

[sjastro]

@sjastro , Mistral 7B tried, and it fell flat on the face. Only 7 billion parameters, it is not good for high level science. At least the AI is all run locally on my laptop. It was also much slower than I wanted it to be.

View attachment 353571

Here are my comments followed by GPT-4o.

Paragraph (1):- The iron core does not rebound, it is the densely packed degenerate neutron matter which rebounds after the electron capture process mentioned in paragraph (2).
Paragraph (2):- It is the absence of outward pressure countering the effects of gravity before the neutron degenerate matter rebounds which causes electron capture and not higher density.
Paragraph (3):- This is news to me, neutron degenerate matter is incredibility rigid and resists compression due to the Pauli exclusion principle hence the idea of the collapsed core oscillating is highly unlikely.
Paragraph (4):- Is flat out wrong for the reasons already given.

Here is GPT-4o's critique.

 

[AlexB23]

Here are my comments followed by GPT-4o.

Paragraph (1):- The iron core does not rebound, it is the densely packed degenerate neutron matter which rebounds after the electron capture process mentioned in paragraph (2).
Paragraph (2):- It is the absence of outward pressure countering the effects of gravity before the neutron degenerate matter rebounds which causes electron capture and not higher density.
Paragraph (3):- This is news to me, neutron degenerate matter is incredibility rigid and resists compression due to the Pauli exclusion principle hence the idea of the collapsed core oscillating is highly unlikely.
Paragraph (4):- Is flat out wrong for the reasons already given.

Here is GPT-4o's critique.

View attachment 353572

So, GPT-4o seems to outsmart Mistral 7B by a long shot. Yeah, my AI on the computer is better for making devotionals than doing science, and it does not seem to understand how the iron core works at all. It does surprisingly well at explaining rules for Olympic sports though, especially when using RAG (retrieval augmented generation), cos I downloaded the rules for almost every sport onto my computer.

 

[sjastro]

I asked GPT-4o why it initially ignored the Pauli exclusion principle as the major driver for core collapse supernova explosions.
The response was very human like.

 

[AlexB23]

I asked GPT-4o why it initially ignored the Pauli exclusion principle as the major driver for core collapse supernova explosions.
The response was very human like.

View attachment 353615

Wow, AI is become more human like, and it knows how to correct itself.

 

[sjastro]

Wow, AI is become more human like, and it knows how to correct itself.

When it comes to providing a diagram in explaining the role of the Pauli exclusion principle in a supernova, GPT-4o suffers terribly from hallucinations.

​

This diagram is best analysed in the field of psychoanalysis as an example of art under the influence of LSD, or something that might have been produced by a Francisco Goya, Edvard Munch or Hieronymus Bosch.

 

[AlexB23]

When it comes to providing a diagram in explaining the role of the Pauli exclusion principle in a supernova, GPT-4o suffers terribly from hallucinations.

View attachment 353625
​

This diagram is best analysed in the field of psychoanalysis as an example of art under the influence of LSD, or something that might have been produced by a Francisco Goya, Edvard Munch or Hieronymus Bosch.

Rebonks? Excusy? Dang, this AI looks like it took some shrooms, man... Some of the text looks like some old Greek Orthodox, or even Georgian (Western Asian). GPT-4o is high, higher than a kite attached to a zeppelin on top of Everest.

 

[sjastro]

Rebonks? Excusy? Dang, this AI looks like it took some shrooms, man... Some of the text looks like some old Greek Orthodox, or even Georgian (Western Asian). GPT-4o is high, higher than a kite attached to a zeppelin on top of Everest.

Copilot is even more psychotic.

 

[AlexB23]

Copilot is even more psychotic.

View attachment 353629​

Whew, Microsoft Copilot has been hitting the drugs really hard. This image could go in a modern art museum. Have you tried DALL-E?

 

[sjastro]

Whew, Microsoft Copilot has been hitting the drugs really hard. This image could go in a modern art museum. Have you tried DALL-E?

Copilot uses DALL-E.

Here is how I would use a diagram in explaining the role of the Pauli exclusion principle in a supernova.
At the quantum scale nuclear physics uses what is known as a Nuclear shell model - Wikipedia to describe the energy levels of the neutron where the Pauli exclusion principle states no pair of neutrons can occupy the same energy state unless their spins (denoted by the arrows) are anti-parallel.

 

[AlexB23]

Copilot uses DALL-E.

Here is how I would use a diagram in explaining the role of the Pauli exclusion principle in a supernova.
At the quantum scale nuclear physics uses what is known as a Nuclear shell model - Wikipedia to describe the energy levels of the neutron where the Pauli exclusion principle states no pair of neutrons can occupy the same energy state unless their spins (denoted by the arrows) are anti-parallel.

View attachment 353656

That is a much better diagram, though it needs a bit more color.

 

[sjastro]

That is a much better diagram, though it needs a bit more color.

It's not meant to be a work of art but to convey information.

 

[AlexB23]

It's not meant to be a work of art but to convey information.

That is true. Utility comes before aesthetics.

 

[sjastro]

So why has there been vast improvements in AI over the last 12 months?

 

[AlexB23]

So why has there been vast improvements in AI over the last 12 months?

View attachment 353785​

Does ChatGPT actually know about itself, or is this just a hallucination?

 

[Hans Blaster]

I wasn't entirely satisfied with this response as it made no mention of the role of the Pauli exclusion principle in providing the neutron degeneracy pressure in the supernova mechanism.
I quizzed GPT-4o about it.

View attachment 353566

ChatGPT did a better job the first time. That was until you tried to convince it neutron degeneracy pressure was important and it fell for it. Even in cold neutron stars degeneracy pressure isn't dominant. In baby proto-neutron stars inside supernovae, it is even less so. Collapse is halted because of the repulsive core of the nucleon-nucleon interaction that makes further compression difficult. The only real error the first ChatGPT output made was to call a neutron star "incompressible".

This furthers my impression that AI tools like ChatGPT can be treated like authorities, but they are certainly "persuadable".

 

[Hans Blaster]

When it comes to providing a diagram in explaining the role of the Pauli exclusion principle in a supernova, GPT-4o suffers terribly from hallucinations.

View attachment 353625
​

This diagram is best analysed in the field of psychoanalysis as an example of art under the influence of LSD, or something that might have been produced by a Francisco Goya, Edvard Munch or Hieronymus Bosch.

"Rebonks"

 

[sjastro]

ChatGPT did a better job the first time. That was until you tried to convince it neutron degeneracy pressure was important and it fell for it. Even in cold neutron stars degeneracy pressure isn't dominant. In baby proto-neutron stars inside supernovae, it is even less so. Collapse is halted because of the repulsive core of the nucleon-nucleon interaction that makes further compression difficult. The only real error the first ChatGPT output made was to call a neutron star "incompressible".

This furthers my impression that AI tools like ChatGPT can be treated like authorities, but they are certainly "persuadable".

The mistake made by ChatGPT in point (3) of post #34, it referred to a core rebound without describing the mechanism of the rebound.
The core rebound due to neutron degeneracy pressure lasts for around 20 ms during which time frame shocks waves are produced and magnified by colliding infalling matter with the rebounding core.

The nuclear force is strongly attractive in the range 0.8 -2.5 fm, below 0.7 fm it becomes strongly repulsive due to the Pauli exclusion principle and is found in neutron scattering experiments which supports the theory of the core rebounding during the supernova process.

 

[sjastro]

Does ChatGPT actually know about itself, or is this just a hallucination?

A hallucination occurs when AI does not have access to data and makes up a response.
I doubt AI has reached a level of sentience, its response to improvements over the last 12 months is based on available data.