AI vs. Gematria

Gematria is like a multiple choice test in school where all the answers are wrong, and before receiving the failing grade everyone gets to argue that, "At least I wasn't dumb enough to pick 'C'!!" 

 When I was in high school computers were still infants compared to today’s field where everyone has a phone with vastly more computing power than what NASA lunar landing vehicles had.  It was a glorious time when people did what computing was made for instead of serious analysis about astrophysics - GAMES.

I literally started off with punched tape loading of the program.  A huge pain in the ass since the machine that punched the tape screwed up often enough that one tiny dot punched incorrectly made an entire yards long tape unusable since you ended up with LOOP being read as POOP while inputting, etc….  But we did have a 3D tic tax toe game.  Which was far enough advanced to learn from mistakes.  The first few games would be a cakewalk for the humans, and play long enough it would get to the point where human victory was virtually impossible.

Here’s the Wikipedia article:

https://en.m.wikipedia.org/wiki/3D_tic-tac-toe

And pay attention to the part about how mathematical analysis proves that the first player always wins, but the method is too complicated for most people to memorize.  That’s kinda important to the story.

There’s a very silly game you might have heard of called chess.  I’ve been witness to the chess computers rise over decades.  To summarize as quickly as possible, based on human games over many centuries it is suspected that perfectly played chess would either be a win for white, or a draw.  With no chance for black to win.  The first move is that important, even with a game as complex as chess.

The first computers including my Fidelity Electronics were expensive, and not that good compared to humans with knowledge of chess things like pawn structures, taking the initiative at the cost of material and long term positional compensation.  Eventually top grandmasters would still win, but arguably having had the machine put up stiff resistance.  Strategy of playing against a computer program revolved around how they were extremely materialistic, always grabbing that pawn or rook versus minor piece exchange regardless of the consequences.

Fast forward to AlphaZero and Leela.  These are based on a neural network.  And like our almost 50 year old 3D tic tac toe game, learned.  Instead of playing against humans they learned by playing a butt ton of games against themselves.  For a pop culture reference to make an analogy- the computer in Wargames playing tic tac toe against itself to learn that nuclear war was a terrible idea.

The practical applications of neural net learning are something like feeding a large amount of complex data and the net has a huge reference library to better and more quickly make an evaluation than traditional more human reliant methods.  “You dumbass human, that’s colon cancer, not gassy flatulence from lactose intolerance.”

But every great new toy has the downside of meeting up with the crowd that is materialistic about grabbing that pawn that’s available now without regards to the long term consequences.  So we get deep fakes of bogus Putin speeches, ChatGPT creating fictional legal references and social media algorithms that do a terrible job at moderating content.

The early chess computers had different levels based on how long the machine was “thinking” about the position.  And at the higher levels there really wasn’t much difference relative to humans as they would waste a lot of time deciding between multiple wrong answers.  Now the neural nets are so advanced it’s literally impossible for human victory.  They have a reference library of so many wrong answers they know what doesn’t work.  Moves that in a human game would still work since humans, with their clock running can’t weave through the consequences of how their position will collapse five moves down the road.

Somebody like Alex Jones and the big names in the gematria community rely heavily on dumbing it down to appeal to the basic instinctive reactions.  They operate in a world that wants you to play the chess computer in the 1970’s at level one where you win the game by grabbing that pawn.  They want you to think that choosing among multiple wrong answers is OK, or even good.  Take Ivermectin for you flatulence, it will do wonders for your colon cancer you didn’t know you have.

This is the way the crank/grift magnetism operates.  Multiple wrong answers are constantly presented with an enormous amount of resources wasted on them.  Some players have the program set on a higher level; some arguments are more lucid than others.  If you want to impress me, search for the answer you would get from the neural net instead of downgrading the level of your computer until you get the answer you want to hear.

The reason Douglas Adams’ 42 gag is so memorable is because it’s the gematria response to an exceptionally difficult question.  Now people think that if they get laughed at for the wrong answer of 42, if they upgrade it to a three digit number they are now magically super smart, when they really are just lost in the rabbit hole of multiple wrong answers.