The Machine Codes Better Than Us, and That Is Not the Problem

After 30 years in IT support, networking, training, and a variety of other areas, I have come to realize that technology gets better over time. Wow! I know what a revelation.

There is still an argument about who is better, and to borrow a great quote, “From a certain point of view…” honestly, both viewpoints are right. That is the focus of this blog. AI is better at coding. AI can generate code faster and with fewer typos than a human. But can it generate better code? Yes and no. Yes, if it's guided properly, and no if it is not.

Vibecoding, according to Webster’s, is a semi-recently coined term for programming by prompting an AI, which then writes the code. In vibecoding, the coder does not need to understand how or why the code works (which, in my opinion, is wrong, but we will get to that soon enough), and often must accept that a certain number of bugs and glitches will be present. I disagree with that last statement. In fact, I think the opposite is true. AI can help remove bugs and glitches. Look at Mythos by Anthropic. Yes, great marketing and hype, but also proving to be as good as the marketing and hype.

One assumption I make when it comes to vibecoding is based on how I view and understand AI. On a recent episode of Security Now!, a statement was made in passing that caught my attention and that I truly believe in. To paraphrase: “AI has vast amounts of knowledge and absolutely zero understanding.”

Think of a triangle: references are your past chats or linked sources. That is just information or data. Knowledge is the sum of those references. Understanding is grasping the underlying meaning of that combination. I would like to say I came up with this, but honestly, it comes from the Institute for Faith and Culture. Not exactly a common source for a computer blog.

The AI can reference something and knows where or when to use it, but does it know why to use it? That is understanding. The line between the two is thin, and I want to point that out.

I am starting this blog because I am retired. With all my years in IT, I was not a coder. I wrote scripts and small things, but nothing major. I wrote things when I needed them and had no other way to get it done. A good example was Windows Deployment Services. I had several scripts I kept using over the years to help automate Windows deployments. I took examples posted online and, through trial and error, made them work.

I also took several programming classes over the years in the hope that something would stick. One of my bucket-list items was to write an Apple app.

I retired only a few years ago, but like most IT professionals, retirement does not mean I gave up my love of technology. I noticed that AI was taking off quickly, and I decided to jump in.

One area I tried as an experiment was coding. It made sense that AI would do coding well, and I was curious how well. I saw the YouTube videos saying you can code an app in a weekend. I knew this was not true. Well, maybe if you were a coder already, but I was curious. It started with a simple brew.sh script to automate updates on my MacBook. What surprised me was how fast and accurate it was. What would have taken me maybe an hour or two took me five minutes.

Since then, by accident and curiosity, I have written two Safari extensions, approved and published on the App Store. One Chrome extension and one Firefox extension. These are also approved and on the various stores.

It did not start with extensions. It started with much larger, grander ideas. I am still working on them. It has been a learning process. There have been some amazing revelations and some pitfalls.

I have always had the philosophy in computers of keeping it simple. It is more like the pirate code. I use it as a guideline, so not absolute. Still, it has worked. For example, if you buy an iMac, then do things the way Apple wants you to. If you use Windows, then do things the way Microsoft intends. Things just work better overall. When you start to tweak things you can cause more problmes. Again, it is a guideline, but it mostly works.

This blog is going to be about that journey of learning to vibecode, along with the benefits and pitfalls that come with it.

Don