AI Training With Gematria

 

This has been building up for a long time, and the dreaded gematria bubble is about to explode.

Somewhere around 1805 Big Gematria laid their foundations for takeover via predictive programming.  Trying, and ultimately failing for the short term, to insert language into what would have then been the 11th amendment:

“The right to fuck around and create meaningless tiny numbers on a gematria calculator shall not be infringed.”

Some political historians and scholars have speculated the failure resulted from how nobody knew what a calculator was.  It’s tough when your decode predicts something like artificial intelligence correctly hundreds of years before it actually is created.  The majority of these scholars pointed to the use of the word “fuck” being naughty and unlikely to escape spell check and replacement with “duck”.  It’s amazing how two opposing parties actually arguing the same essential point could be so divisive and waste so many resources.  Big Gematria grifters continue the prediction of AI vs. prediction of spell check to this day.  Rumor has it Marjorie Taylor Greene, openly in the spell check camp, is secretly a closet predictive AI person.  But I digress.

The idea of trying to use AI in its current form to do gematria decodes is bonkers multiplied by bonkers.  One of the main purposes of gematria is to do fuzzy math.  Hey, I got this 46 here, but dammit I really could have used a 47.  Oh look!  Add an s to the end of that!  Cha Ching!  And the people asking questions of AI that involve even a basic math problem.  Yeah, currently it’s struggling.  We’re at an infancy stage of AI where you can test it with a question knowing full well what the answer is, tell it that it’s wrong and it will try again.  Produce another wrong answer.  Try again.  Then another wrong answer, tell it it’s wrong and try again.

Just like misspelling counts in gematria, bad math counts.  Being guided by gematria so called teachers who have a job of deliberately providing bad (but popular) information coming up with the wrong number is absolutely no big deal.  46 can be the evil number of the moment just as easily as 47.  And since we’ve already made up a push for a constitutional amendment let’s make up a psychological study or two or three.

In 2019 a group of  2,000 Yale students who were given a one hour training course on gematria basics.  After that, 100% of those students made decodes that were just as bad as decoders that had been doing it for years.

In 2020 a group of 450 randomly selected gematria loving people were given a gematria calculator app that included a random number generator.  When asked to decode that year’s Super Bowl their answers were just as good as the control group that had been given the “correct” gematria calculators.

In July 2024 a group of 17 researchers from the prestigious Omphaloskepsis Institute set about having a neural net self train itself to do gematria decodes.  5 of the researchers had to be institutionalized trying to make sense of the output.  The rival GIGO institute was more successful.  Those 14 researchers are all happy, contented married with decent jobs lots of kids and puppies.  Their AI, after 47 seconds produced a single three letter and one punctuation mark output:

WTF?