The Metaphysics of AI:
Understanding What the Thing is
If you can't explain it simply,
you don't
understand it well enough.
— Albert Einstein
The various chatbots are things of computer technology. As such,
they must be understood within the context of existing technology,
specifically software. Metaphysically speaking, an AI is a very
large database of language sets. Why AI is easier to make a query
of when compared to traditional database technology, say like
Structured Query Language (SQL), is that the queries
need not be made in a specific form of computer syntax.
The intent
herein is to help the reader see AI technology within
the greater scope of computer database development, and how AI is the byproduct
and result of years of prior development. Furthermore, my hope is
to help show those not
associated with the Computer Industry, or its slavish sycophants, how AI
technology is less revolutionary in nature, but more an evolution and
refinement of existing technology.
From a consumer point of view, traditional database queries are of no
value Consumers cannot use them, and must rely on experts if they
need to query the data. A
December 4, 2020 article in Towards Data Science,
discussed how query language and syntax kept lay people alienated from
their own data. The author offers examples of an SQL type of
query. The technology itself was an impediment to anyone not
versed in the
lexicon of databases.
The main problem with relational databases is that SQL is verbose to write and time consuming to reverse engineer. That is, if you have an SQL query that you did not write, trying to work out what it says is sometimes very difficult.
The author adds that "So for a long time, people have been looking to find ways to query databases in natural language so as to make their lives easier." And there you have it. The kids may think this is all brand spankin' new stuff because the technology is of their time. But the technology represents the culmination of decades of work.
Traditional databases require developers to establish manually any relationships
that might exist between the various types of data. Tables, rows,
and columns are linked by common data types. To put this within a
context most computer users will understand, imagine a huge Excel
Workbook. In that workbook are a seemingly infinite number of
Worksheets, each with a different name that indicates the nature of the
data in that sheet. An Excel formula that would manipulate data
from a vast quantity of worksheets, rows, and columns would be a very long
string of text requiring precise punctuation and other
symbology. Yes, one could spend endless hours
grouping and naming all the cells. But that would not negate the
precision required to write a good Excel formula to turn all that data
information in Excel into something might be useful on a daily basis.
AI writes this formula for you.
Obviously, the AI hallucinations are natural language queries that for
one reason or another failed. The database language model is still
developing technology. The most obvious reason for a hallucination
is the underlying data is not readily available, but nonetheless an incorrect
relationship was made between data points not relevant to the query.
Another reason for an AI hallucination may have been that the query was
itself nonsense. In essence, the human failed the machine; thus
causing the machine to fail the human. Yes, I am suggesting a kind
of
symbiosis exists between a human user and the AI's simulation of
consciousness.
GIGO still reigns supreme in our computing
experiences.
Although the rules of proper syntax for a computer
query may have now been expanded, the need for clear and concise
language when making a query has not changed. In fact, because AI
does indeed have an enormous set of language data, the machine may
choose an inaccurate dataset due to a relationship between disparate
bits of information that simply does not exist. If we humans want
this technology to work for us in a practical way, it is incumbent upon
humans to learn to write clear and concise prompts. My advice is if
you are unfamiliar with writing techniques suitable for computer
consumption try what follows:
✓ Write the prompt in Notepad
or other text editor first. Then copy and paste that into the
prompt window.
✓ Read the
prompt back to yourself to verify your words make sense.
✓ Be sure your prompt clearly
ask for the information you want back.
✓
Check for misspellings that might change the intent of your prompt.
Certainly prompts should be concise, but nevertheless detailed enough so
the AI gets your point.
This article is about the Metaphysics of
Artificial Intelligence: the essential facts of what AI truly is, devoid
of all the hyperbole So I began with a short chat session with
Bing. The purpose of the chat was first to establish that
Metaphysics is still a real thing. The study of which isn't really in
vogue much any more, and I am far removed from a college Philosophy
course. My secondary motive was, if I approached the chat as a
sincere human whose only goal is to learn something about Metaphysics, would
the results then be satisfactory for my intentions. I only want to
learn; I don't want to write a paper. The results were
surprisingly pertinent.
I didn't try to trick Bing, or make Bing blush. The prompt was clear and concise, but still a complete sentence. The results were also clear and concise. Bing's response gave me verification that my prior assumption that Metaphysics is in the curriculum of a classical education. I was still on the rails, so a follow up was in order. Fundamental nature of reality is a fairly dense topic. So that was the follow-up. The query was not made in the form of a command or question. The phrase was already established as being within Bing's lexicon of Philosophy and relevant to the overall intent of the chat session.
Some ground rules for our chat session have been established. Bing can respond appropriately to a simple phrase prompt that is derived from what has already been established Bing knows. Again, I emphasize that both Bing and myself have remained on task, an important element in any successful project. Having thus established Bing has a pretty good lexicon relating to Metaphysics, I asked Bing to be analytical. I used the pronoun 'it" deliberately to test that Bing kept the overall topic in memory.
In this short chat session, I placed Bing's responses within the framework of Bloom's Taxonomy, a standard for measuring learning that I had discussed on January 14, 2023. Bing showed a remarkable ability to synthesis disparate bits of information into a very convincing evaluation of what scholars might contend is a certainty, but may not match everyday human experience. What Bing is doing here, with an ability that was heretofore not possible, is making logical connections between the relationships that exist in what are the known words and phrases about Metaphysics. Bing then communicates those relationships in very cogent ways.
Finally, since Bing was considerate with me, I reciprocated and did what Bing likes done. I used the broom. Cleaned my session before I left. It is what adults do. Still within the new 8 prompts limit for any new query.
There was no hanky-panky going on in this chat session. In fact, I had accomplished my goal, which was to establish that Metaphysics is a real course of study in higher education. I knew I needed a refresher on that topic and Bing delivered.
These are tasks that any thinking adult can perform to get useful
results from the AI. The early reports were flawed. The
people teasing out senseless responses from the AI have only served to
confuse the general public. And thus preformed a great disservice
to us all. Yet, doesn't this always happen whenever children are
placed in positions of responsibility they are not yet mature enough to
handle? Every responsible parent knows this is true.
I
don't believe the technology went off the rails by itself. That is
not in the scope of what AI is. It is more likely that these
people who assume that they have tricked the AI into doing whatever
these people think they have tricked the AI into doing, are not in fact
manipulating the AI. It makes more sense that those humans behind the AI are manipulating
those hapless users. The AI masters, working the other side of the chat
screen, could have easily directed these AI users down the Rabbit Hole
those users apparently wanted to go down anyway. How else could the AI learn to
mimic these aberrant human language sets?
That went off the
rails, however. A trip down the Rabbit Hole always ends at the
Mad Hatter’s Tea Party. Aberrant personalities on
one side of the AI screen were led by the KB of those aberrant
personalities on the opposite side of the screen. Also, as any
parent knows, when kids are left to run amok, they will run amok.
The adults had to take over the project. Also, as any parent knows,
after the kids do run amok, there is a major cleanup job that must be
done. Either that, or the forced fed
Thorazine finally kicked in.
I judge things
by a very simple formula: Does it make sense? That the AI projects
are technology let loose on society with no human interaction
managing the process just does not make sense. Too much money has
been invested in AI to allow the AI to run on Autopilot. This is
especially true of Microsoft. Anyone who has followed MS over the
last several decades knows Microsoft management is actually quite
conservative. Them Gnomes of Redmond just ain’t gonna drop $10 billion,
and then throw those billions against the wall just to see what happens.
People are deluded into assuming that, just because they can
write a little a code, they understand hard IT business decisions. Go
ahead. Continue to delude yourselves. But you are simply slaves to
your delusions. There's nothing new about children being tricked by adults into thinking they are
calling the shots, but are really only doing the adults work for them.
Welcome to the Circus, Lemmings. It's like how Tom
Sawyer convinced Ben that fence painting is a blast, and snookered Ben
into doing Tom’s chore for him. I would send MS a bill for your labor,
Lemmings.
All the histrionics of the earlier reporting on AI reminds of
what Bill Cosby — before his Fall from Grace
— said about cocaine in his 1983 film, "Bill
Cosby: Himself."
I said to a guy, "Tell me, what is it about cocaine that makes it so wonderful," and he said, "Because it intensifies your personality." I said, "Yes, but what if you're an asshole?"
Get the point?
