Working Up

by Dwayne Phillips

AI is a tool to be used by those who provide ice cream cones and honey buns. Lose sight of that and advance at your peril.

Who is going to be the winner in all this AI competition? Who is going to be the loser in this? After all, avoiding losing is almost as good as winning.

Here are a few folks who won’t lose: those who make ice cream cones and those who make honey buns. Please, running AI here, there, and the other place is nice. I’ll settle for an ice cream cone on a hot afternoon and a honey bun with a hot cup of coffee on a cold morning.

Forget about those tech folks saying, “Every company is a software company,” and now, “every company is an AI company.”

Companies that make ice cream cones and honey buns, and thank goodness there are many of those companies, should focus on ice cream cones and honey buns. They can and will use AI in “the back office” just as they use word processors, spreadsheets, the internet, and databases in the back office.

It’s not the tools; it’s the final product.

My plea to the ice cream cone and honey bun companies is to remember this and focus on their products. Those products are SO MUCH BETTER than AI.

by Dwayne Phillips

Which way is AI going? The big datacenter or on my little laptop computer?

We seem to be at some fork in the road regarding artificial intelligence. The big (colossal) datacenters are rising from the farmlands of America. Big, bigger, and even bigger. If the price doesn’t run into the hundreds of billions of dollars, it isn’t worth mentioning.

The little artificial intelligence is also bearing fruit. I’ve recently read of ways to run AI locally on my electronic typewriter (an Apple MacBook Air with an old-time M1 processor). Hmm, I like the sound of that.

Will running AI locally become practical and practiced before those datacenters are completed?

I guess that is one way of asking the question of questions.

I have always cheered for the local AI. For years, I’ve wanted to type, “What did my dad think of such-and-such?” while I point AI to the folders that contain what my dad wrote years ago. Search, summarize, and display the answer. Please. Here are some private, handwritten notes. Scan them, convert them to digits so I can file them and ask questions about them as well. The whole world of those chattering bot companies don’t need to know these things.

The big datacenter, at least as of today, is winning. More money and people are going that way. The tide is turning against the datacenter as it becomes the new villain in a country where we love to have villains, especially if they are rich and appear villainous.

I would hate to see datacenter construction stop halfway. There are advantages to have four walls, a roof, electric power, air conditioning, and access to high-speed data lines in a community. I have written of that before. I still believe it would be good. What would be bad is farmland covered with cement, a few steel girders standing, and then the site abandoned. I guess those things could become the world’s largest outdoor pickle ball arenas, but, who wants that?

Perhaps the better is some of both—not a fork in the road but two finished and well-traveled roads. I would like that. Please. Let’s do this well.

by Dwayne Phillips

Sometimes, some things are important to me. It is usually a mistake to extend that importance to another person.

The idea hit me. This is IMPORTANT! I will dig into this topic. I will expand it in many ways.

Me: Hey, look at the results of my analysis. I have looked at all the details in all these dimensions. Look what I discovered.

Other Person: With a blank expression looking at the numbers on the screen. Silence.

Me: These findings are important. See how they affect the outcomes?

Other Person: More silence.

Me: Given this, you should understand how important my work is to you.

Other Person: Even more silence. Finally, “Can you start at the beginning and tell me why I want to know any of this.”

Me: Because it is IMPORTANT.

Other Person: Can you start at the beginning and tell me why I want to know any of this.

Me: Silence.

The meeting ends in mutual silence.

Sometimes it simply wasn’t important to the other person. I should ask. I should do better.

by Dwayne Phillips

One of the greater challenges is to continue what was started on the other side of an event.

We have the best intentions. Let’s work on topic A. Topic A is important. Topic A deserves all our attention. Here we go! Topic A is it.

Boom! Something happens. Sometimes it is merely going home, eating, sleeping, eating, and coming back to the office. A small canyon or event. Sometimes it is bigger. The weekend! Lots of activities and fun and all that. Monday morning comes: what were we doing? Something about an A or B or something? A bigger canyon or event.

Sometimes it is an event. “Hey, look at the news. See what happened? This changes our industry completely. Let’s react. NOW!” A even bigger canyon. A week or month later, “What was that Topic that was so important?”

Whatever it is and however large it is, there is a canyon, a gap, time or event or anything. We are on the other side. What was it that we were doing back on the first side?

Something derailed our best intentions and best efforts. How can we return to topic A? Should we return to topic A? What was topic A? Sigh.

Sometimes we just lose track of what we were doing. We can do better.

by Dwayne Phillips

Vibe coding shows us that coding is not all there is to developing software. No duh!

For many years I applied to jobs for software engineers and later what was called “developers.”

“How many recent years have you written Java?” was a common question (substitute any other language).

I tried to explain that coding or programming was about 10% of software engineering or developing software. The reply was blank stares. (See ya’ old man).

Fast forward a few years and we have vibe coding or whatever we call it this week. The current slate of chattering bots are spitting out code. The code works for the most part.

Computer programmers are obsolete! No one is majoring in Computer Science! It is all over! Those were the headlines for a day or three.

But wait. Recent discussions indicate that there is more to this than coding. There are the ideas, the requirements, the designs, maybe even specifications, and Heavens to Betsy (does anyone say that anymore?), there is testing. THERE IS MORE TO DEVELOPING SOFTWARE THAN CODING!

Hmm. What are the chances that I could revisit some of those recruiters who only cared about my recent Java experience. I could tell them, “See? I told you so.” Perhaps zilch. Never mind.

Let’s do better.

by Dwayne Phillips

This is interesting to me. Is it interesting to you?

interesting, adjective, something that attracts your attention, arouses your curiosity, or holds your interest because it is unusual, exciting, or engaging. It is highly subjective, meaning what is fascinating to one person may be dull to another.—Google search

The answer to my question is, “Probably not.”

See the definition above, the second part of the definition, the part I don’t like. This interesting thing is dull to you.

Also notice in the definition is the absence of “important.” Just because it is interesting does not mean it is important. And if it isn’t important to me, the chances of it being important to you are … pretty much zero.

Nevertheless, I will tell you about it. SEE THIS? WOW! Pardon my enthusiasm. Sorry to bother you. I’ll just casually turn and walk away. Maybe we will talk again tomorrow.

by Dwayne Phillips

All this computing power, datacenters, rising prices for hardware, land, water, electric power, etc. for what? To plan a party? This is utter waste. Come on, stop it!

Wired recently had an article on the full story of OpenClaw (is that the name they eventually decided to use?). That is the agent orchestrating thing that, according to Wired, revolutionizes something related to computing that needed revolutionizing.

Wired interviewed Peter Steinberger on how all this happened. One statement about Mr. Steinberger, “He figured that a digital assistant with access to your apps and data could automate a huge range of tasks. To plan a party, for example…”

To plan a party? To plan a party? TO PLAN A PARTY? Are you kidding?

Per the summary above, we are devoting all these resources so that software can run software can run software so someone can plan a party. Really? Utter waste.

There are much better things to do with all those resources than connect emails, calendars, recipes, etc. to plan a party. How about expeditions to some other planet or to the bottom of our oceans or simply to the middle of our great forests. And then there is medical research. Oh, that. Perhaps those problems are too difficult or too something-or-other so we plan a party.

Per the article, Mr. Steinberger strikes me as a smart and well-meaning persons. Perhaps he could have cited some other problem-solving examples. Some problems are interesting and pressing. That doesn’t make them worth the consumption of resources.

Let’s do better.

by Dwayne Phillips

The world of computing has turned upside down.

In 2001, Marc Andreesen noticed and noted, “Software is eating the world.”

Here I am to notice and note, “Hardware is the new software.”

Everyone wants to buy computer hardware. Prices of computer hardware are shooting upwards as demand far exceeds supply. If you make computer hardware, you own the world. Cancel your contract for $1. Open the bidding and get $10 for the same hardware.

Write it down. I think I said it first.

by Dwayne Phillips

Can’t use one of those AI tools in your workplace. Solve that problem.

Some wise person once said something about a hammer and a nail. It goes something like, “if the only tool you have is a hammer, the world looks like a nail.” This has something to do with another wise-person saying about, “There is no free lunch.”

Enough old, wise sayings. I face problems. I have resources. The resources are limited. What do I do?

Some of these current AI tools are useful. If the problem has to do with text, which many of my current problems do, these tools are quite useful. I have this AI tool and that AI tool. I work in this setting and that setting. Sometimes I am in a setting where my problem would be easy to solve if I had a tool that I don’t have. Rats.

Go back to the other setting, ask the AI tool how to solve the problem in the first setting with the tools available there.

For example, I have a problem in a setting, a tightly controlled office, that is solvable by writing a Java program. All I have, however, is Power Shell. How do I solve the problem with that? Another example: I have a problem I can solve with Excel. I don’t have Excel in this setting. How can I solve it with bash scripts?

And then I can extend this with: I have a problem that a group of degree-holding engineers can solve. I don’t have those people. What I have are 18-year-old high school seniors. How can I solve this problem?

Hmm. Use a tool to suggest tools. There may be something to this. All I have is a hammer. How do I make my problem look like a nail?

by Dwayne Phillips

These chattering bots merely predict the next N word(s). Software has been doing this for many years—nothing new here.

In the 1990s (yes, 30 years ago, I am that old), word processors started guessing what word we were trying to type. Some called that “auto complete.” It was nice. Not perfect, but nice. Sometime between then and now, we saw IDEs (integrated development environments). If any term ever needed an acronym instead, that one did.

Regardless, these IDEs did auto complete. Start typing the name of a function or variable, the software would look through of list, and show guesses at what it was I was typing. Well, that was convenient. It wasn’t perfect, but it helped me remember things that I new but were a bit murky.

Nowadays (does anyone still use that word?), we have these new chattering bots. They write essays and entire computer programs. WOW! New stuff! Great stuff!

New? Let’s see, in the 1990s, the word processors would predict the next one word. Today, the chattering bots predict the next one thousand words. What’s the difference? Just the difference between one and one thousand. It is N. N is bigger or smaller depending on various things that affect the outcome, but N is just N. Programming? We have gone from N is one to N is a larger number, but it is still N.

Why all the fuss? Why are school teachers at all levels in an uproar over students not learning this or that? It is just N words where is goes from zero to some arbitrary larger number. Same stuff. Nothing new here.

Let’s move on to something of more substance to discuss.