Working Up

by Dwayne Phillips

We have come a long way with AI aiding in decisions. The machine is powerful and new. Or is it?

AI is helping us decide what to do. It performs analysis in seconds that would have taken weeks or months. Wow! Look at us now—machine-aided decisions. Great new stuff.

Or is it? Seems that we have been using machines to aid decisions for … well, several thousand years. And these machines aiding decisions use probability to predict the future and aid our troubled minds just like today’s approach to AI predicts the next word based on the probability of all words gathered during training.

Hmm. What ancient machine-aided decisions? Flip a coin.

A coin can be used as a lever, so it is a machine. The two sides of a coin are different and that difference is easily distinguished on sight. Which side will appear is based on probability. The next side up depends to some degree on all the results of all the prior tosses into the air.

Flipping a coin is artificial intelligence. The decision the coin flip helps determines depends on the question we ask before flipping. Hmm. Sounds like prompt engineering.

Wait. Stop. Nope. There must be something wrong with this analogy. Good grief. A penny only costs a penny and I can use that penny again and again. An AI datacenter costs a few gazillion pennies. Could it be that we are … naw. Can’t be. We’re too smart for that.

Now, what’s the next word I should type in this little blog post. Hmm. Sentences start with either a noun or verb, so heads is a noun, tails is a verb, and then…wait a minute. A coin flip didn’t write this. Or did it?

by Dwayne Phillips

The one great advantage to working with a chatbot.

curt: adjective, using or expressed in few words, in a way perceived as rude.

Ah, that’s it. That’s what I like about working with a chattering bot: I can be curt and I won’t hurt anyone’s feelings.

I was attempting to create a family tree picture yesterday. I tried several different chattering bots. None of them did anything worthwhile. The answers were wrong. The pictures were junk. Good grief.

And you better believe that I laid into these chattering bots about what a lousy job they were doing! I told them again and again that they were making stupid mistakes and they needed to do a better job. I didn’t mince words. I WAS CURT!

It was okay. No one cried. No one pouted. No once stormed out of the room or whatever people do this days when someone is curt with them (does anyone use “curt” anymore?).

I knew there was something good about these chattering bots. Now I know.

by Dwayne Phillips

Software that writes software is quite helpful. Well, maybe sort of not.

MIT jumps into the fray and asks a few dozen programmers if AI that writes software is helpful to people who write software. First, the folks at MIT are smart enough to not base a survey piece on a survey of a few dozen people. The sample size, as is often repeated, is not big enough.

Anyways, is software that writes software helpful? It certainly is helpful to those folks who don’t like the folks who write software as a profession. Those programmers are difficult to talk to and get along with and all that stuff. The machine is much easier to disdain.

But how about those folks who write programs professionally. Do they like the machine helping them? Of course they do. Good grief. If there wasn’t a machine, they wouldn’t have any machine to program. And compilers, software that writes software, are the greatest invention since 1s and 0s. And editors, like vi and emacs and such, are great software that writes software. The “modern” systems have auto-complete so you don’t have to remember all the names of all the functions and variables that you have to use. These new IDEs predict the next character or ten for the programmer.

And speaking of predicting, these chatbots predict the next character over and over and over until you have 100 lines of code predicted. Programmers understand the loop concept. Do one good thing. Loop over that N times. What’s the difference between predicting one character and predicting N characters? Nothing.

So all this software writing software is great stuff. You just have to specify what you want correctly.

Wait a minute. Specify what you want correctly. That is what computer programmers have always done. Instruct the machine to do something. Do this in assembly language or Java. Well, maybe now we do it in … what do you call that specification language we use with these chattering bots? Oh, English. Well, its sort of English.

Wait a minute. Are these chattering bots are nothing but new compilers? No one told me about that.

Back to the drawing board (does anyone use a drawing board any longer?). We can do better.

by Dwayne Phillips

This current AI boom all comes down to the cost of providing bathrooms for people. Or so it seems.

I think we are over thinking the idea of what we think about these chattering bots. Folks, the chatbot is just software running on a computer. Yes, some of that software makes me think I am talking to someone. I frequently hear people describing their interactions with Claude as “talking to Him” and “He gave me these ideas” and “He put this data into a table for me” and so on. What a great name for a piece of software: Claude. That is so much nicer than ChatGPT. Oops, is software nice?

Why does a company have software to “talk to me” instead of a person? The software runs 24 hours a day and doesn’t “get tired.” If the software has errors, it always has errors. It doesn’t have more errors after not taking a nap or not eating enough protein. And the software doesn’t need to take a break and go to the bathroom.

Ah, we’ve hit the bottom line: chattering bots don’t need bathrooms. Bathrooms are the most expensive room in any building. They have all that extra plumbing and hardware. Bathrooms need cleaning often and that costs money.

The AI boom? It is all due to the cost of the bathroom. Remove the need for the bathroom, and you really have something. Claude never goes to the bathroom. He has a job for life.

by Dwayne Phillips

Let us begin anew. Why not?

By some arbitrary counting of days, a.k.a., a calendar, today is the first day of the new year 2026. I suppose that is a good thing for us. Hence, some of us have the new year’s resolutions wherein we resolve to do something.

I resolve to breathe as long as my heart keeps pumping and I can breathe.

I resolve to think as long as my brain continues to function and I can think.

I resolve to write as I am the type of person that if the above two resolutions hold, I must write as I have no idea what else to do.

I thank God for my family—every person in my “family” (yet another arbitrary concept).

Today is called Thursday. It is in the latter half of this week. Hmm. A new year beginning in some random part of a week. Why can’t we align the new year with a new month with a new week? I wander off topic.

Anyways, I wish you and yours a happy new year and a happy day and a happy moment. I hope I wrote something that helps you have some happiness. I try.

by Dwayne Phillips

All this $$$ is killing AI.

I used to work in AI. That means that I researched things that might allow us to do things that we couldn’t do. I attempted to instruct a computer or, as we used to say, I attempted to write a computer program to do something that people could do and computers couldn’t do. I struggled with the fundamental that all a computer could do was tell the difference between a one and a zero.

It was research. That is what AI is: a field of research.

At one time a home thermostat was considered AI. Make it colder when it was too hot and make it warmer when it was too cold. Amazing stuff that only thinking people could do. After successes in that research, AI turned into control theory. There are many other examples of research in AI turning into other endeavors that we take for granted.

Now, however, AI is big BIG B I G money! I have nothing against folks making money and supporting their families. I have nothing against folks turning research into practice and creating jobs for other folks. Good stuff.

If you want to make money, shout AY EYE! or something like that. People with money will beat a path to your door. Who care$ if we are a bit loose with word$. We are talking money here.

Nevertheless, I am troubled by folks in technology not understanding these basic tenets. Come on folks, let’s do a little research into history. Journalists? Well, their profession is marked by getting things wrong far more often than getting them right, so I have low expectations there.

Come on, we can do better.

by Dwayne Phillips

Today is Christmas Day. Some thoughts.

By most arbitrary calendars, today is Christmas Day. Many folks celebrate this as the day that a boy was born to a carpenter and his young wife in a stable in Bethlehem some 2,000 years ago. Many folks celebrate this as a day to exchange gifts.

Different cultures have different traditions around this day such as putting a “tree” of one sort or another in a room and putting decorations on it. I loved the silver aluminum tree of the 1960s. If you had it right, you also had a lamp with a wheel of colors turning so that the silver tree would turn colors.

It seems that in the last 20 years, people do a White Elephant gift exchange exercise. It is some type of game or conversation starter or something. I guess I understand the point of all that, but it is not my favorite past time.

What I like about this day and this week is that I sit and talk with people that I otherwise wouldn’t see. I like to “catch up” with their lives and know what they are doing.

I have what I consider to be two “sets” of grandchildren. One set comprises two teenagers while the other set comprises twin toddlers. That is quite a spread in experience. I enjoy either, or, and, and both experiences. I wish I could have the toddlers sit in my lap while I talk Calculus with the teenagers. Alas, that won’t happen this day.

I am of the age where I have seen more Christmas Days in my past than I will see in my future. That is a great influence on my thoughts about this day.

I enjoy good health this Christmas Day as does my wife as we have our 42nd Christmas Day together. Our children and grandchildren also enjoy good health this day. Those statements summarize the best earthly gifts I have. I will take them, happily.

by Dwayne Phillips

I often hear people speak with great certainty. Further consideration, however, reveals almost no certainty.

People like to speak with great certainty. I write those words because I often hear people speak with great certainty. If people don’t like to do that, why do they do it almost all the time? Great puzzles for me.

Despite the certainty in the speech, further consideration often reveals almost no certainty. 1 + 1 = 2. Certainty. Consistency. Aha. We are getting somewhere.

“So-and-so believes such-and-such,” said with great certainty. But really? How does the speaker know what old Mr. So-and-so believes? Has the speaker confirmed all this with Mr. So-and-so? Can we talk to Mr. So-and-so on the phone right now and confirm this? Is there any hesitancy from the speaker to call Mr. So-and-so right now? Hmm.

“I spoke with So-and-so last week and I believe he is leaning towards such-and-such,” said with great hesitancy. Aha. Now we are making progress. This is candor and honesty.

People, however, understand that the certain statement is just a short way of saying the hesitant statement. Don’t they? Need I say all those extra words every time?

I guess I am not smart enough to understand that the shorter version means the longer version. I like to hear the words. Perhaps we can take the time and energy to say the extra words and show hesitancy instead of certainty. We can do better.

by Dwayne Phillips

“Just as soon as I build my writer’s cottage, I will write and write and write,” said no one who was actually writing and writing and writing.

I am going to write and write and write. But first, I need to:

• build a writer’s cottage where I can write
• buy a comfortable chair that I can write in
• buy a better computer for writing
• learn to paint illustrations to go with my writing
• stop making lists of things I need before I write

I put the last bullet in the list just for fun. he he he.

I learned many years ago from some older people who had been through all these things. People who say these things are never going to write. Most will never build that writer’s cottage or do anything on the list. They will think and dwell and analyze and plan and dream and never do anything. That is too bad. They have the desire, but just can’t force themselves to do anything.

I think the horror writer Stephen King wrote something about needing only a room with a door that can close. There is a point to that. I’ve written essays and stories about people who went to the public library and wrote novels using only Google Docs on the library’s Internet computers. No cost for anything. Just sit and write. I seem to be able to write at home and at Starbucks. I seem to be able to write using any keyboard. I seem to be able to write regardless of noise, music, chatter, or absolute quiet. I know those things don’t fit everyone.

Write.

I don’t care if you have a cottage or tent or cave.

Write.

I once wrote a 300-page book that had 200 illustrations. I scribbled the illustrations on the backs of pieces of paper that had book drafts. The illustrations were awful. The book sold to publishers with those awful scribbles. I guess I was just lucky.

Write.

Frustrated? Write about frustration. Cheated? Write about being cheated. Angry? Write about anger. You are uniquely qualified to write about those things.

Write.

Have a great desire to build a writer’s cottage before you write? Go ahead and (you probably guessed what comes next) write about the desire to build a writer’s cottage so you can write.

Write.

Please, write. The rest of us need to read what you have written.

by Dwayne Phillips

AI has eaten the world. It is the latest and greatest thing…or is it?

I hang around America’s Federal government. There are many new programs in our Federal government to push and pull and take advantage of AI. AI will do wonders. Isn’t it wonderfully wonderful? Sigh.

Perhaps I am just too old for all this.

Generative AI generates things that were not there, hence the adjective “generative.” Well, actually it doesn’t generate, it predicts based on things that were already present. Oh, so it’s Predictive AI. Wait, we’ve had predictive AI for many decades.

Wait, before that we had mathematics that predicted things. Remember y=mx+b? Put a new value as x, use the slope of the line and the Y-axis intercept point and predict the next value of y. But that wasn’t Predictive AI, that was just the equation of a straight line. Yes, and the equation of a straight line was a mathematical model of the line. And the AI model (like GPT-whatever) predicts the next word. The AI model is based on all the text that was fed into some software to build a model. The equation of the line is based on all the numbers fed into … wait, there must be something wrong here.

Could it be that all this generative AI is just a clever packaging of mathematics and software and … ? Naw, couldn’t be that. Must be something else, huh?

‘Fraid not. Predicting the next thing is as old as the equation of a straight line. Except now, we have super duper computers that consume lots of electricity and are clever enough to be cleverly deceptive.

Perhaps I am just too old for all this.