Tag: artificial intelligence

August 24, 2026 – One Tiny Seedling…

August 24, 2026 / / 0 Comments

I probably shouldn’t admit this, but I am ‘working’ on the weekend. I have this many, hour training class that I am required to take and it is due at the end of this week. The only real time I have to do it is on the weekend. Hurray for me.

The subject of the training is Generative AI. We have all heard of AI, the acronym for Artificial Intelligence. But, what is Generative AI? The word generative is somewhat like equivalent to learning. This is a way of saying that AI evolves by learning. It is accepted that the current versions of Large Language Model (LLM) style AI systems are time and knowledge bound. This means that the data is loaded and it remains static. Generative AI is an LLM that connects to other systems like the internet to keep more current.

This will not be a post to regurgitate everything that I have learn or to educate on the nuances of AI. But I did want to take some time to discuss AI. The first thing is to look at AI in the context of other civilization changing technologies. Can you imagine a business that never adopted the telephone as a technology? Imagine if you could only communicate with the company by either mail or in person? I think we all realize that is just not a real scenario in the 2020s.

There are still some business that didn’t quite make the internet revolution. They tend to be small businesses that were established prior to the personal computer invading our lives in the 1990s. Even so, it is pretty difficult to find a person without a smart phone so while the business may not have a website or e-commerce capability, they still likely have access to the internet by some means and it is highly likely that the internet plays into things like 3rd party payroll companies.

The point of all of that is AI is here. When we hosted our Taiwanese exchange student a year or so ago, he didn’t use a browser to answer questions with the internet or lookup things, he went directly to ChatGPT. We cannot simply bury our heads and say, I will never get a telephone. Like it or not, we need to understand at some level that this has fundamental consequences.

I am getting to the age that I understand that this is going to impact my life but I can also feel that I have put in my time on the bleeding edge of technology. In my training, one of the instructors was saying that AI skills are becoming a requirement for technical people on their resume and job interviews. Companies are trying to sell this as a productivity accelerator and yet there are some fundamental flaws that I see.

On one screen I have the training session going on the other is a gothic church in Spain as the Windows background of the day. I was looking at these two diametrically opposed images and was struck by the paradox. Here I was getting ‘preached’ at with the evolve or die message while on the other screen was something that was built before electricity was “discovered” and has stood for over 800 years.

As someone that remembers spending $200 1990s dollars on graphics cards that simply did not work, I am glad that things have improved to the point that we don’t wonder if hardware is going to work anymore (within reason). I also don’t have the time to spend weeks to try and get them to work and I don’t have the patience or the desire to try anymore. This is partially why my Linux experiment has been stalled in the same place it was a year ago. At least certain areas I still have desires to learn new technology but I am very time limited.

I guess I wonder, is a technological life better? So much work in my space is simply busy work. It is considered a proper deliverable with the project but I can think of projects that were extensively documented where the knowledge is simply lost when there is a vendor or employment change. It makes me consider the value of the time spent writing manuals and how to’s that are simply lost and nobody knows that they even exist due to lack of continuity or even knowledge of the use.

My current project is modernizing a computer system that is about ten years old. That is it, the lifespan of millions of dollars spent is less than twenty years. There is no doubt that companies can collect an exponential amount of data on products and manufacturing processes compared to the 1950s. But to what end? I would argue that of the things that are really important, we are still getting significant and deadly foodborne illness without a definitive source of the problem.

A significant theme that I have observed in the last ten years is large service companies have grown into behemoths by making the assumption that one person paid in US wages can be replaced with a person making an eighth of the US salary. For insurance purposes, they can always add seven more if necessary. I am painting a broad brush but I often see one person getting replaced with four and taking three times as long to accomplish the same objectives. Of course AI is seen as the solution. Not only will it take less people to do the same job but they won’t even have to have the same fundamental hard skills as today. Think of the technological world the internet has opened up. It has fundamentally changed our access to knowledge whereas AI is going to fundamentally change our access to ability.

End Your Programming Routine: To be clear, I am not advocating for or against AI. It is here and we might as well learn the nuances because it is here to stay. What I am saying is that compared to human history and what has stood the test of time, nothing really replaces fundamental skills applied to durable materials. We do what we have to do to get by, let us consider perspective of it all. AI is probably necessary and it will change our world but it will be something else in 800 years.

December 19, 2024 – Things That Makes You Go, Huh

December 19, 2024 /

I am kind of behind the eight ball here. It has been party after party, late night after late night and working like a banshee. Add to that some emotional struggles not just what I wrote yesterday but also some additional drama. So, a normal Christmas season. I need to get my Festivus podcast put together so I am putting out a cheap one here.

I had planned to review a book that I am reading. I thought that I had two weeks to go to finish it and I haven’t read a page since. That is getting pushed to next year since I am taking a break for Christmas. The other day, I checked the mailbox and it was crammed full of envelopes. It took me four handfuls to empty it.

This picture is 41 benefits statements from the insurance company. I couldn’t help but laugh and laugh as I was sorting through the pile. I surmised that the hospital finally got around to submitting to the insurance company for my wife’s treatment over the summer. I only read a couple to make sure they were for the same event. Sure enough, it was.

We are definitely not there yet, but I imagine that this is the kind of practice that is going disappear as Artificial Intelligence gets a real footing. The reason being is that this logic is pretty complicated. A thinking person would say, one claim and one document. Instead of generating a document for every line item, it would be one document with every transaction (think bank statement).

But, this gets into the complication of the logic. You can imagine that each vendor/provider/service ends up with a different task code or billing line. With that each of those equal a transaction. The logic then ties each transaction a letter. This way the system does not have to know if these transactions are related or not. It is the simplest way of laying things out. And simple is always the most robust.

If there is any doubt, this is what I do. Take a scenario, break it down into to the steps, that is critical to getting it right. A warning though, the output is always important. The reason it is critical is that the output changes all of the details within the steps. As an example, I have often had to deliver reports (letters in this case) with the project. And often those are an after thought, with the assumption that all of the data is captured. But when it is not, then sometimes you have to even re-program the steps and step logic to capture.

Because the software process is complicated and technical, what goes on behind the scenes is obscure. For that reason, the industry has become reliant on sloppy practices. Think about electric cars. I hear that a new model (like the Mustang Mach-E) consistently has improved over time. Without substantial infrastructure changes, a 2020 model is essentially the same as a 2024 model. This is because they are running the same operating system if you will.

Because nobody knows what is going on at the consumer level, the industry uses this as cover to release substandard code and iterate to a final or better state. Realize that some of this is a necessary evil, it is not just a conspiracy theory. The reality is that there are deadlines and that a manufacturer would not want to release a 2020 vehicle finally in 2024. That would also be bad for business.

The other day, I had a Windows 10 Blue Screen (of Death). This is the first time I have seen this in years. It used to be fairly common twenty years ago, in fact I didn’t even know that they still existed it has been so long. A Porche 911, is fifty years of evolution. Sure, if you compare today’s to an original, they are not the same car but the result is constant evolution. It didn’t make a leap from one generation to another. That is what software is.

I know that all of the contrarians like to highlight the simplicity of Linux and the advantages of crowd sourcing. I also agree that it is good and I appreciate the benefits. But what cannot be denied is that Microsoft has built a stable and widely accepted operating system. Good luck getting your very specialized TV tuner card to work in Linux unless you are going to program it yourself.

End Your Programming Routine: Choice is good. Sometimes what is seen as ludicrousness, waste or incompetence actually has a reason. Despite the $20 of postage to mail all of those statements out, they have to doing something to make billions in profits last year. Make no mistake, the only way insurance companies make money is by charging the patients so we paid for this but there is a method to the madness. In my book forty-one explanation of benefit statements in one day is mad.

June 17, 2024 – AI and You

June 17, 2024 /

This is an episode more in line with what I hope to deliver. I have had to recently take training in artificial intelligence. As a result, it got me thinking about why I should care. I don’t really, but I suddenly realized what the future was going to look like after AI becomes widely adopted and implemented like so many companies in my work space are pushing for.