The Technical Explanation as to Why AI Technology Will Never Work

For my upcoming Crush The Street article, I will focus on a topic that I spoke about in my interview with the company's founder, Kenneth Ameduri.

In the above hyperlinked interview, I stated that AI technology is a scam. I received some pushback, including from Ameduri, which is the normal reaction. In fact, I'd be surprised if I didn't receive any pushback.

Previously, I explained in broader terms why I don't believe that AI technology will ever work (that AI platforms are limited to the programming that was originally put in). Below, I take tidbits from my Crush The Street article to explain in further detail why I think AI as it's presented to us is a scam:

At the heart of AI technology is a derivation of the Gauss-Markov Theorem. Without getting bogged down with the details, every decision in a real world environment is theorized to be based on a prior decision or event. AI technology is built upon a sequence of yes/no logic chains, as follows in the below diagram:

As you can see, the trajectory of a single “Gauss-Markov chain” is essentially only limited by the programmer’s imagination. The problem with deep learning technologies is that while a decision-making protocol can be established (such as the probability of winning, as in a chess game’s AI technology), the events to trigger this protocol must be programmed.

For example, in a chess game, none of the computer’s deep learning takes place until you activate the program, and initiate a sequence. But we know from our own personal experience that human intelligence requires no such stimulus. We can choose to be active learners, passive learners, or engage completely in introspective musings.

To overcome this limitation, AI technology engineers incorporate multi-faceted Gauss-Markov chains. Here, we have multiple trigger events leading to multiple chains that react to an ever-complex matrix of variable situations:

Undoubtedly, this is an impressive leap in artificial intelligence; unfortunately, it’s still not “deep learning” as all these matrices have to be programmed by the programmer. At a certain point, it becomes untenable to keep programming these multiple chains. It also detracts from the fact that AI technology has a central origination point.

As much as it's "sexy" and "cool" to believe in an AI, deep learning, self-aware technology, it ain't ever gonna happen! There are simply too many variables, too many sub-variables within mainlining variables, to program.

And even then, new situations and unforeseen circumstances can occur that can throw the whole experiment off!