Friday, January 9, 2026

Lofton Struggles to Understand the Essence of AI


ChatGPT Says the Mark of the Beast Is Coming Soon (NOT Clickbait)
Reason & Theology | 13 Aug. 2025
https://www.youtube.com/watch?v=eklCZZOhIoo




13:32 Have you ever asked how much mathematics an abacus* understands?

An abacus can do mathematics, or better, you can do mathematics on it, but it doesn't understand mathematics.

ChatGPT can understand neither language nor logic.

Interesting test case.

Trent Horn claimed (I think twice even) that Cardinal Baronius (a very holy man, disciple of St. Philip Neri, also pretty crucial in Apologetics contra Lutherans on the terrain of history) had said "the Bible tells us how to go to Heaven, not how the heavens go" ....

The immediate search on google for the phrase was in Galileo quoting it to Duchess Christina. However, Galileo didn't specify either in what context or what Church man he was quoting.

On Quora I posed the question "is there any early source that it was Cardinal Baronius who said etc."

Now, I have got so far no answers. But, there is a Quora bot, pretty similar to ChatGPT. It kept circling back to the Galileo letter which I already knew didn't mentioned Baronius by name. It wouldn't have done so if it understood the terms of the question. However, what such bots are programmed to do is copy text and especially copy probable continuations of any phrase already given (including questions, obviously).

So, ChatGPT is giving you the answers people are giving on the internet.

15:24 The answer pretty obviously means, from someone's side, "given the hypothesis of Christianity, x is likely" ...

And when I say "someone" I don't mean ChatGPT but a programmer.

16:45 The machine is neither confused nor lying.

It's simply not conscious at all.

That you presume to do theology apparently without knowing that is however either confused or facetious in your case, since you are conscious.

21:26 Popularity leads to frequency of an answer being written online.

Frequency of an answer leads to higher likelihood of ChatGPT sharing it.

Or do you believe it's 23:16 simply a sign that AI isn't yet fully 23:18 consistent and reliable and is open to 23:21 manipulation?


23:18 Did you say "isn't yet"?

Will never be, more like it.

* For the illustration, I modified the illustration on wikipedia:

https://en.wikipedia.org/wiki/Abacus#/media/File:Abacus_(PSF).png

By Pearson Scott Foresman - This image has been extracted from another file, Public Domain, Link

No comments: