[boris8826]

Stephen Wolfram, a genius mathematician, creator of Mathematica and Wolfram Alpha, wrote an article that explains how ChatGPT works. It is oriented for a layman reading.

Besides a thorough explanation of the engineering, I found it interesting to ponder this view of his [emphasis is mine]:

But in the end, the remarkable thing is that all these operations—individually as simple as they are—can somehow together manage to do such a good “human-like” job of generating text. It has to be emphasized again that (at least so far as we know) there’s no “ultimate theoretical reason” why anything like this should work. And in fact, as we’ll discuss, I think we have to view this as a—potentially surprising—scientific discovery: that somehow in a neural net like ChatGPT’s it’s possible to capture the essence of what human brains manage to do in generating language.

I think this fits with his theory of automata, published in his book New Kind of Science, in which he claims that simple operations generate the complexities of physical reality. In that book, early in its opening, he proposes a deterministic theory of Free Will, which says that it can’t be predicted what a person will do, because there is no shortcut to a “computation”—operations in reality that must be performed—to know the future.

https://writings.stephenwolfram.com/2023/02/what-is-chatgpt-doing-and-why-does-it-work/

Tip: If you type three regular dashes in the editor, it turns into a long-dash after the post is rendered.

/sb

[Jim Hanson]

Re: Boris Reitman’s post 104479 of 12/29/24

Be careful with Wolfram. I have a lot of experience with his technical work and a little with him in person, and my not-so-tentative conclusion is that he’s a genius who has lost touch with the need to validate his ideas. And have a look at Melanie Mitchell’s notably restrained review of A New Kind of Science, available here.

*sb

[Harry Binswanger]

Re: Boris Reitman’s post 104479 of 12/29/24

 he proposes a deterministic theory of Free Will, which says that it can’t be predicted what a person will do, because 

He’s making a rookie mistake. Determinism vs. free will is a metaphysical position. Prediction is epistemological, and irrelevant.

That no one, including you, knows your fate matters not at all if it is fated.  

/sb

[David Hayes]

Evidence that AI tools merely make connections (rather than offer properly-analyzed results) pops up often. A particular instance may prove eloquent to readers of HBL, as the misuse of a source is clearly apparent in the instance.

On a Facebook group within the past few weeks, one participant posed this question:

Does anyone know of any recordings where Ayn Rand referred to Richard Halley?
That name can be pronounced either Hay-Lee or Hal-E.
I am trying to figure out which she intended it as!

One respondent began with “I have no idea if this is accurate, from ChatGPT:”, and then let the remainder of his comment be the results of his ChatGPT inquiry. It read, in part:

The pronunciation of “Richard Halley” as “Hal-lee” (rhyming with “valley”) comes from several sources tied to Ayn Rand’s works and those who were close to her:

[…]

The Ayn Rand Lexicon: While primarily a philosophical reference, it includes commentary on Rand’s works. Some references to Halley align with the standard pronunciation endorsed by Objectivist scholars.

Do you see something wrong here? ChatGPT implies that it consulted an audio version of “The Ayn Rand Lexicon,” or text that reports on the pronunciation of this character’s name! I’ve had the book since it was published in the 1980s, and afterward received mailings and announcements from ARI and from booksellers catering to Objectivist audiences, and never was I informed of an audiobook version. I have periodically searched “Ayn Rand” at audiobook seller sites, and never saw this title offered. What’s more, as everyone reading at HBL must know, “The Ayn Rand Lexicon” is designed as a reference work, and thus does not lend itself to being offered as a set of lengthy audio files.

In reporting this here, I am fully aware that this may come to the attention of the one person in all the world who knows with the greatest certainty all of the licensing agreements that The Ayn Rand Lexicon has had, and thus can settle definitively whether ChatGPT could have accessed a recording or any text within the book that answers how Miss Rand intended her character’s name to be pronounced.

(Sarcasm is so rare on HBL that I realize that it can’t come across as becoming here, but in situations such as this it is hard to retract myself from it.)

/sb

[David Hayes]

Adding to my previous comment (David Hayes’s post 152140 of 2/4/25):

Just so that there is no doubt that Miss Rand pronounced Richard Halley’s surname as “hal-ee,” you can hear her yourself pronouncing the surname in her “Fiction Writing” lecture series, in lecture 5 at 31 minutes and 33 minutes. (Times are based on the editions prepared for public distribution, as offered in MP3 format on the Ayn Rand Institute’s eStore and as streaming audio on YouTube.)

/sb

[Author]

Posts