A Booming A.I. Except It Is Not Ready For Prime Time


I first encountered A.I. back in 1985 when I was working at the Massachusetts Institute of Technology in Cambridge, MA. Back then, A.I.’s language was named LISP. We played with it but at the time the most powerful computer we had was something called a DEC-2, DEC being an acronymn for the Digital Electronics Company. But it did not have either the computing power nor the disk space or memory space to properly handle the language. An AI company in Cambridge popped up with the idea that it could solve the AI dilemma. It was named Thinking Machines which went under after a few years.

Today, the explosion of A.I. is due to the progamming language named Python. What Python has done is it made a dramatic reduction in the amount of code needed for its A.I. product. But the problem still remains that even this breakthrough language still developes a code that is far too big for a single computer, or even a small number of computers. Hence, the rise of the data center. A top tier computing centers can use as many as 2.6 million servers, a server being a machine that is behind a particular program that you use, and it tells us how many computers needed to make the A.I. program work. That’s 2.6 million computers!!! The sheer number of computers tells us that A.I. is still far too large to use.

That people are protesting the building of these new computering centers should tell everyone that A.I. is still in its infancy and should be taken down in most cases. Those 2.6 million computers generate a lot of heat and need a heavy load of electricity to run. And if that was not enough, these centers are noisy, meaning there is considerable noise polution generated.

You are finding A.I. in your apps and in the Window’s 11 CoPilot. But unfortunately, A.I. is what we used to call in its Alpha mode. Alpha used to mean that the program was only ready for early testing to see how it worked. Then a program would move into its Beta mode where is was used by a select number of users to prove that it is ready for general distribution. Well, as I said above, at the Beta part, it should have been painfully obvious that, as I said, it is still not ready for “Prime Time.”

Worse, they have even tried A.I. on super computers. They are the most powerful computers available; one is the Japanese model named Fugaku, and the other is the U.S. model Summit. (This is as of June 2026, Trends in AI Supercomputers) These super computers are expected, by 2030, to use 2 million chips and 9GW of power, that’s gigawatts which translates in 9,000,000,000 watts, and it is wattage that creates all the heat. That further means they will need cooling jackets of water to keep them at a necessary temperature. As an aside, it was in the late 1970s and early 1980s that large computers of the day back then also used this sort of a cooling element. It seems like we have taken a step back.

The problems facing A.I. today:

  1. the public is unified in its resistance to computing centers
  2. the A.I. computer language of choice still leaves too big of a program to be easily used by a small group of computers
  3. the computer centers use too much water
  4. the computer centers use far too much electricity
  5. the computer centers create noise polution
  6. the companies using A.I. have over promised and under delivered
  7. all computer programs, regardless of the language used, are only as good as their worst programmers

Now, until this latest iteration of the new A.I. language, the language(s) of choice is called “C,” which is used in conjunction with a program called C++. The language C needs but a single computer and has been invisible to the public. C can be used on any sort of computer, meaning, it can be easily used on an Apple computer as it can on a Dell computer, two computing systems that are vastly different. It was so revolutionary that it got named as a “portable” language that did not care what computer it is used on.

I want to impress upon anyone who reads this to resist further spread of these computing centers no matter what country you live in. Ask lots of questions about how good the A.I. program actually is. One point of hard resistance has already arisen from the medical community over its mistrust over A.I. conclusions, reference my words on how good an A.I. program actually is. Most people like to trust the computer programs put in front of them. However, A.I. is a complete departure from any other existing program in that it creates, by itself, new “paths” of “answers” to any question posed. That problem should be obvious in that how reliable is that new path?

I wish I could get down into the nuts and bolts of exactly how A.I. works. But words like recursive, a common A.I. idea, takes too long and goes too deep for the ordinary user to understand. But it is exactly those words that must be questioned, and be questioned by experts in the field. Do we have enough experts, or we like to call them in the computing field, Gurus, to go around? I highly doubt it, at least at this point.