A.
My lifelong practice when encountering any claimed font of wisdom is to first see what they have to say about a matter in which I’m proficient.
For compendiums regarding history or the globe today, language or religion the first thing I do is check the index for “Israel”, “Hebrew” or “Judaism”.
Needless to say, most compendiums are garbage.
Anyone with an IQ over 120 can pass as an expert an nearly any subject after a few days of cramming.
His or her analysis and conclusions will be WAY OFF because all of the tools they used to evaluate the subject came from beyond any true involvement with it. And the fact that they are mercenarial enough to pretend to expertise they lack in exchange for clout and cash doesn't help either.
But the only people who can tell from their statements of facts that they are actually ignoramii are people who actually know the area from a greater involvement than a week’s research.
B.
YouTube’s analytics tool for video creators called Studio is primarily focused on helping these bullshit artists engage in such “research”.
Noting my proclivity of late to discuss matters biblical, Studio keeps suggesting videos and searches to “help me create more content”.
I don't need to point you to the videos because they are all of the professional videos with stark thumbnails and titles that have hundreds of thousands or millions of views.
Not one of which is simply a “dude who knows his stuff” speaking unscripted directly to you.
Which brings us to AI.
My experiences with AI are too interesting and lengthy to delineate here but my overall conclusion is that it is not trustworthy. In fact, it's even less trustworthy than some garbage vontent provider because AI has no “tells”.
AI is the perfect bullshitter.
C.
I can't tell you how many AI programs I asked (and in how many different formulations) “What is the 11th word of the Hebrew Bible”?
It's okay that AI got it wrong (always).
What's not okay is how confident AI was, so much so that after telling me - inaccurately - that it was השמים, or אור, or את it goes straight into a scholarly explanation of the word or whatever salesman trick it chooses to breeze past the fact that it has no idea.
Some programs allow follow up questions and I used the option to ascertain how the error could have taken place.
AI apologized in accordance with the Indian Customer Support script it inherited and then provided me with the correct answer (which in every last case was again wrong) and explained to me the parameters that led to the error.
I then checked the accuracy of that claim (such that the wrong answer it provided was actually the 11th most common word in the Bible) and found it to have lied with a slickness I have yet to find on a used car salesman.
I’ve had fun using AI’s creativity when I was among the first “testers” for ChatGPT.
After fiddling with the options and phraseology I literally got it to compose a 2 minute comedy piece by Adolf Hitler to pipe into the gas chambers along with the Zyklon B.
D.
It was fuckin unbelievable. Needless to say, chatGPT did indeed learn a lot from me so good luck getting such goodness out of it now.
I feel like I should find and share it here but that takes time If you really care to see it sign up for Club Yedidya and tell me that you wanna read that (and presumably related stuff I saved from that period) and I’ll track it down for you.
A friend recommended a new AI app (perplexity). Its faults are similar to the others but I’m an adventurer so I asked it about myself.
The response remained too vague for anyone looking for expert information but accurate enough to provide me with a new ABOUT page for this site.
What do you think?
Any quick n’ easy improvements you can think of? Anywhere likely to be beneficial I should share it?