Working Up

by Dwayne Phillips

AI enabled completing a 12-page report in five minutes. No, that took ten years of effort.

Somehow, using a new piece of software became cheating or something bad. A request arrived at work. Click. Boom. Pow. Out goes a 12-page report. Five minutes. Next?

Cheating!

Consider, however, that the five-minute task was backed by ten years of work. The company uses SharePoint (I receive no compensation from any of the purveyors of products.) It was quite an expense and ordeal to install that ten years ago. The company has piled layers upon layers of documents into SharePoint over the ten years. That was a lot of copying, pasting, moving, etc. over a long period of time. It cost blood, sweat, tears, and salaries.

Then the company pays one of these chattering bot companies to use the new tool. Then the company pays people who have decades of experience in logical thinking, structured thinking, and using tools like FORTRAN, C, and all those things nobody uses any longer. All those by-gone tools instilled knowledge into these people. And all that cost blood, sweat, tears, and salaries. Apply all that experience to markup languages and “skills” and such. More salaries.

And all of a sudden, out of no where, Click, Boom, Pow. Five minutes and the report is done. Funny how that is so simple and inexpensive (N O T !).

Sure, it all seems so simple and quick and must be cheating to accomplish so much in so little time. Funny how that is.

by Dwayne Phillips

These blog posts are longer, i.e., don’t have many words. Sorry, I can’t help myself.

My blog posts aren’t very long (see, I could have shortened that sentence by removing the word “very” and also the word “blog.”). I recently wrote about the Paramedic Method of Editing. That method helps editors shorten writing, eliminate words, bring clarity, have brevity. I have been using that method without knowing about it. Brevity and clarity were my goals.

When writing, I edit before my fingers touch the keyboard. That is a great skill; that is a horrible curse. I edit all the time. I edit my thoughts. I edit your thoughts. I edit your words. I edit your speech while you speak. (You and your refer to anyone else in the room.) I cut, cut, and cut more in the moment. Sometimes, too often, I am not nice. That is the curse portion of this.

I read many blog posts. I read many things written by many people in many forums. I marvel at how blog posts are so long. I wonder why I’m not writing such long posts with so many words. And then I remember that I edit in a harsh and unreasonable manner.

I am stretching this blog post. I am trying to leave the words and not cut, cut, and cut more. I am not doing well at that.

And that is why these posts are not longer.

by Dwayne Phillips

If it is important to me to tell others what I think, I write it myself. If it is just stuff that a machine could write, let the machine write it.

I write this blog. I sit in the coffee shop on Saturday mornings and press every key on the keyboard to put every word in the blog. I don’t have to do that. I could prompt AI, copy, paste, and be done with it faster with much less strain on my joints. At my age, anything that lessens strain on the joints is worth it. Well, almost anything.

Why don’t I have AI write this? Because this is important to me. I want to tell you what I think. These thoughts have been bouncing around in my head for days, weeks, months, years. Yes, some of these things have been bothering me for years. Something finally turned on the light and here it is. Don’t blame AI for this; it is me.

And then, why do I have AI write something? Believe me, I have AI pump out many words on many topics on many days. Because those somethings are just somethings. Words are required. Bam. Push the button. There are words. Pump them out. Don’t waste time typing. Just do it and move on.

If it is important, it deserves the effort.

If it is just something the machine can do, let the machine do it.

Simple. Now for a refill of coffee.

by Dwayne Phillips

Perhaps those big, bad datacenters aren’t so bad.

How do you think that works?

• You take a dozen photos of your new-born nephew.
• You listen to a new song or an old song.
• You find a Chinese restaurant that is open at this hour.
• You find a gas station on an empty stretch of highway when the needle is bouncing off empty.

or how about

• That automated external defibrillator that brought your grandmother back to life
• The integrated continuous glucose monitors (CGMs) paired with wearable insulin pumps that keeps your next door neighbor alive
• The imagery of the inside of your colon that locates cancer while it is still treatable
• A dozen things in an Intensive Care Unit that keeps several dozen people alive

They work because of datacenters.

Datacenters are big. Still think datacenters are bad?

by Dwayne Phillips

AI writes more code than people can review. Now AI can review code. So why waste time by having people review code? There are benefits.

AI is writing code, i.e., software is writing software. AI can write code 24 hours a day without a break. Wow! And AI writes code fast. Double wow!! And AI can review the code AI writes. Triple wow!!!

Where are the people? What do the people do? Should people waste time reviewing code? No, but yes. No, people should not waste time. Yes, people should review code.

Many years ago, the late consultant and writer Jerry Weinberg told me about organizations where programmers reviewed the code of other programmers. After a while, errors dropped significantly in these organizations. It was not from the code reviews catching errors early. The improvement came from the learning. People reading the work of their colleagues learned from their colleagues. The learning improved the quality of work.

Reading led to Learning, Learning led to higher Quality.

Hmm. Maybe people don’t have the time to read everything AI writes. I believe people have some time for some reading and some learning and some improving. Let’s do better.

by Dwayne Phillips

We return to the era of the main computer center and waiting for computer time.

Way back in the previous millennium, I was leaving graduate school. While in school, I had access to a new 32-bit minicomputer (I think it was made by Perkin-Elmer if anyone else remembers that time). One problem was I had to be sitting in the computer room of our lab to use the minicomputer. I didn’t have to sign up for time on the machine as I was a chosen student and given access anytime I was awake. Still, I had to make the drive to the building.

Then I bought my on personal computer. It only had 16-bit words. It only had 640K of RAM. It only ran MS-DOS. It was only this and only that and only lots of things. The difference was that I owned it and it sat in my home.

I could run the software I was developing anytime I wanted. The software was limited by all the “only this and only that” of the computer, but it ran now.

When was the last time you had to sign up for computer time and had to wait your turn? I had to do so last week. My credits, or whatever they call them, on my chattering bot account ran out. I pay $20 a month, but there is a limit to how much I can chatter and agent and loop. When I hit the limit, I have to wait until Thursday morning to do anything. (Why Thursday? A mystery of the universe that remains a mystery.)

Once again, there is the big computer that runs everything fast — when it is my turn. And I guess I can run a chattering bot locally, but slow, but now, but slow.

We have come a long way and it took a long time to return to where we started.

by Dwayne Phillips

Writers seem to struggle to express the lot of the average person.

Let’s talk about the average savings of an American in their 50s, i.e., born in the 1970s. We want to understand their lot in life and all that. The figures are somewhere.

Hmm, average means to sum everyone’s savings and divide by the number of people. Well, the figures state the average savings is between $1 million and $1.5 million. Below the average? Probably. Yikes. Well, look at the definition. Elon Musk is in his 50s. His wealth (a trillion dollars or something like that?) raises the average quite a bit.

The median savings is a better number (as I judge good, better, and best). It is about $300,000 or a third to a fifth the average. You see, median is calculated by…

Stop. I don’t want to get into the definition of median. Take a statistics course or Google it or something. Median is a better measure than average.

Except I have to explain “median” before using it. Aargh.

The word we want is “typical.” That doesn’t have a good definition in a statistics textbook. We all, however, understand what “typical” typically means. It is typical. Got it? Good. Let’s move on. We can do better.

by Dwayne Phillips

There are occasions when we spend the resources to have everyone in the same room at the same time.

We can do this review and do this meeting asynchronously. The documents are online. Everyone review them when they have time today and leave your comments —said an eternal optimist.

We have some wonderful computer technology today. Networks connect people who are flung far and wide across states and countries and time zones and all that. The documents are online and several people can read and edit at the same time and the technology keeps folks from bumping into or stepping on each other. Wow!

Let’s use all this technology and save us the expense of all being in the same room at the same time. See the statement of the eternal optimist above. Sometimes it just doesn’t work.

Okay, asynchronous reviews of documents can work. You read now, I’ll read this afternoon, someone else will read a little now and then all day, and someone else will read tonight. Our comments are all in the document, so the person who corrects the document has all the information they need. Right?

Wrong. Experience has shown me that this only works if all the readers write their comments well. Of course all our employees write well. We only hire people who can write well. Do they write as well as William Faulkner? If not, forget it.

Commonly heard is something like, “I wonder what he meant when he wrote that?”

If the document or topic or whatever is important, spend the resources to have everyone in the same room at the same time. Get it out. Say what you want to say. Ask questions. Ask more questions. Gain some clarifications. Get messy. Facilitate the meeting. Do the work. The results will be better. We will save resources in the days following the gathering.

Let’s do better.

by Dwayne Phillips

If “big” is in the description, it is in all the descriptions. That hurts.

Here is a process for writing a big document:

• The BIG draft
• The BIG review
• The BIG rewrite

Uh, this usually means a BIG waste of resources. The BIG draft means the writer works and works and works. Finally, finished. The BIG review means half-a-dozen people read and comment on the draft. The BIG rewrite is the writer trying to understand that BIG list of comments made during the BIG review.

Another BIG thing is the BIG collapse in the energy and excitement of the writer. That BIG draft took energy and heart and all those good things that good people pour into work. Someone called this “the tyranny of the first draft.” (I wish I could recall who coined that phrase.) All those comments from the BIG review crush the writer.

How about this process:

• small write
• small review
• small lessons
• repeat

Someone once called this “in-process review” or something like that.

The writer says, “Here is what I’m thinking. What do you think?”

The reviewers say, “I see the point. How about adjusting it this way?”

Work a little. Learn a little. Repeat. I think that is what agile development, DevOps, etc. attempt. We can do better.

by Dwayne Phillips

One way of accomplishing work is to employ fewer better people. Of course this works. That depends on the definition of “better people.”

One way of accomplishing work is to employ fewer better people. It works. Of course it works. Let’s use this definition:

Better people are those who can use the tools to accomplish the task.

Consider farming. When a better plow arrived, fewer hands were needed on the farm. A better farmer learned how to use the newer and better tool. This better farmer could farm more land in the same time and produce more food. Same production; fewer people.

Consider the American office. Less-expensive computing arrived in the 1980s in the form of personal computers. A better person learned how to use the newer and better tool. Word processing, just one application, meant fewer and eventually zero typists in the office. This better person could produce more text in the same time. Same production; less time and fewer people.

The current slate of AI tools arrives. Same story. New tool. Better people learn how to use the new tool wisely and produce more in the same time. Nothing new.

Fewer people bring other concepts. There is less overhead (don’t have to rent a big office building, one office area will do or everyone works from home). Less overhead means less expense. Fewer people bring fewer lines of communication. Decisions can be quicker. And on it goes.

Of course fewer better people works to accomplish work. That is part of the definition.