Log24

Monday, November 27, 2023

A Confused AI Asks for Help

Filed under: General — Tags: , , — m759 @ 9:14 am

Monday, November 20, 2023

The Tumultuous Weekend

Filed under: General — Tags: , — m759 @ 9:13 am

Thursday, November 9, 2023

The Enormous Mess

Filed under: General — Tags: — m759 @ 1:02 am
 

"Symmetry is the concept that something can undergo a series of transformations—spinning, folding, reflecting, moving through time—and, at the end of all those changes, appear unchanged. It lurks everywhere in the universe, from the configuration of quarks to the arrangement of galaxies in the cosmos.

The Enormous Theorem demonstrates with mathematical precision that any kind of symmetry can be broken down and grouped into one of four families, according to shared features. For mathematicians devoted to the rigorous study of symmetry, or group theorists, the theorem is an accomplishment no less sweeping, important, or fundamental than the periodic table of the elements was for chemists. In the future, it could lead to other profound discoveries about the fabric of the universe and the nature of reality.

Except, of course, that it is a mess: the equations, corollaries, and conjectures of the proof have been tossed amid more than 500 journal articles, some buried in thick volumes, filled with the mixture of Greek, Latin, and other characters used in the dense language of mathematics. Add to that chaos the fact that each contributor wrote in his or her idiosyncratic style.

That mess is a problem because without every piece of the proof in position, the entirety trembles. For comparison, imagine the two-million-plus stones of the Great Pyramid of Giza strewn haphazardly across the Sahara, with only a few people who know how they fit together. Without an accessible proof of the Enormous Theorem, future mathematicians would have two perilous choices: simply trust the proof without knowing much about how it works or reinvent the wheel. (No mathematician would ever be comfortable with the first option, and the second option would be nearly impossible.)"

Ornes, Stephen (2015). "The Whole Universe Catalog : Before they die, aging mathematicians are racing to save the Enormous Theorem's proof, all 15,000 pages of it, which divides existence four ways." Scientific American,  July 2015: 313 (1), 68–75. Reprinted in Stewart, Amy; Folger, Tim. The Best American Science and Nature Writing 2016  (The Best American Series) (pp. 222-230). Houghton Mifflin Harcourt. Kindle Edition.

    Compare and contrast with the ChatGPT version.

Wednesday, November 8, 2023

The Theorem as Big as the Ritz

Filed under: General — Tags: , , , — m759 @ 8:56 am

From this  journal on Monday, Nov. 6, 2023 — ChatGPT DevDay —

Monday, May 15, 2023

Chatbot Review

Filed under: General — Tags: , , — m759 @ 9:49 pm
 

By ChatGPT* — 

What is the Cullinane diamond theorem? 
5/10/2023, revised as below on 5/15/2023

The Cullinane Diamond Theorem, also known as the Diamond Theorem, is a result in combinatorial mathematics proposed by Steven H. Cullinane. The theorem explores connections between finite geometries, algebraic structures, and combinatorial designs. 

The theorem is primarily concerned with four related structures:
 
1. Finite geometries: Specifically, the theorem relates to finite projective geometries over finite fields, such as projective planes and projective spaces. 

2. Vector spaces: The theorem connects finite vector spaces over finite fields to the aforementioned finite geometries. 

3. Boolean functions:* The theorem establishes relationships between the finite geometries and certain Boolean functions derived from them. 

4. Symmetry groups: The theorem involves symmetry groups associated with the finite geometries.
 
The precise details and implications of the Cullinane Diamond Theorem are beyond the scope of this response. It involves abstract algebra, group theory, combinatorial designs, and finite fields. The theorem has been the subject of various publications and research discussions in mathematical literature. If you are interested in delving deeper into the Cullinane Diamond Theorem, I recommend referring to the original works by Steven H. Cullinane and related research articles on the topic.

* With one revision by Cullinane: "Boolean functions" replaces "Boolean algebras." See (for instance) M. A. Harrison, “On the classification of Boolean functions by the general linear and affine groups,” Journal of the Society for Industrial and Applied Mathematics  1964 12:2, 285-299.

Wednesday, May 10, 2023

ChatGPT Improves

Filed under: General — Tags: , — m759 @ 2:19 am

Previous ChatGPT responses to questions like those below
have been, to put it mildly, lacking in accuracy. But there has
lately been considerable improvement . . .

Sunday, April 30, 2023

For Harlan Kane: The Walpurgisnacht Hallucination

Note that if the "compact Riemann surface" is a torus formed by
joining opposite edges of a 4×4 square array, and the phrase
"vector bundle" is replaced by "projective line," and so forth,
the above ChatGPT hallucination is not completely unrelated to
the following illustration from the webpage "galois.space" —

See as well the Cullinane  diamond theorem.

Monday, March 13, 2023

Frontiers of Artificial Mathematics

Filed under: General — Tags: , , — m759 @ 11:57 am

Previous posts have shown ChatGPT answering the question
"What is the diamond theorem?" with references to Thurston
and, later, to Conway. Today it is Penrose's turn.

Related search results (click to enlarge) —

 

Sunday, March 5, 2023

Annals of Artificial Mathematics

Filed under: General — Tags: , , — m759 @ 9:11 pm

The response of ChatGPT to a question about my work
continues to evolve.  It now credits Conway, not Thurston,*
for the diamond theorem.

The paragraph beginning "The theorem states" appears** to be based
on the following 24 patterns — which number only 8, if rotated or
reflected patterns are considered equivalent.

* For Thurston in an earlier ChatGPT response to the same question,
see a Log24 post of Feb. 25.

** The illustration above is based on the divison of a square into
four smaller subsquares.  If the square is rotated by 45 degrees,
it becomes a diamond  that can be, in the language of ChatGPT,
divided into "four smaller diamonds ."

Saturday, February 25, 2023

The Diamond Theorem according to ChatGPT

Filed under: General — Tags: , , — m759 @ 2:38 am

ChatGPT on the diamond theorem, Feb. 25, 2023.

The part about tilings, group actions, and the diamond-shaped
pattern is more or less OK.  The parts about Thurston and 
applications are utterly false. 

Compare and contrast . . .

Friday, February 24, 2023

Red Lines: ChatGPT vs. ChatGPS

Filed under: General — Tags: , , — m759 @ 8:48 am

"Recently, several leading academic journals and publishers
updated their submission guidelines to explicitly ban researchers
from listing ChatGPT as a co-author, or using text copied from a
ChatGPT response. Some professors have criticized these bans
as shortsightedly resistant to an inevitable technological change.
We shouldn’t be surprised at the disagreement. This is a new
ethical space that only roughly follows the outlines of our existing
agreements on plagiarism, authorship criteria, and fraud.
Precisely where to draw red lines is not clear."

— Ben Chrisinger, Feb. 22, 2023, in The Chronicle of Higher Education

Flashback to Log24 on Dec. 11, 2022

Compare and Contrast:


Digitally-Assisted Art from … 
 

Chat 'N' Chill ®
 

The Source:

A perhaps more interesting digital assistant
now offers Netflix 'N' Chill —


See as well "red line overload" in the previous post.

Sunday, December 11, 2022

Camera-Ready Copy

Filed under: General — Tags: , — m759 @ 1:16 am

Compare and Contrast:

Digitally-Assisted Art from …

Chat 'N' Chill ®

 

The Source:

A perhaps more interesting digital assistant
now offers Netflix 'N' Chill

Powered by WordPress