Author: J

Avoiding the Trap of Bad Metrics in Business

Goodhart’s Law is the principle that “when a measure becomes a target, it ceases to be a good measure.” It describes the unintended consequences that occur when people alter their behavior to manipulate a specific metric in order to achieve a reward or avoid punishment, distorting the metric’s original purpose.. [1, 2, 3]

This quote has been popping up in feed quite ubiqitously lately. Usually, it refers to some company that has decided to implement an AI usage dashboard in order to track their employees’ use of AI tools. It reminds of the burn rate metrics during the dotcom heyday.

Most people with a little experience would most likely scratch their noggins in wonder and ask themselves the question: “but why?”.

The funny thing is, this is not really a new trend, we just hear more about it because of the AI hype. I think many software delivery people have been presented with all kinds of odd metrics they have to report on. At least, I have seen my fair share.

The most common metric I have encountered is time. How much time has a team spent on a task. This is a pervasive metric that most organizations use, and more to do with budgeting constraints than actual delivery or value. It is also a metric that often becomes the whole point. It is more important that we meet the estimate than we deliver robust software or, dare I say, actual customer value.

I get it though. It’s all about money in and money out. Of course, at the end of the day a team still costs what a team costs in terms of wages.

There are numerous of examples like this that I think illustrates the same problem. We tend to measure the wrong things. And in a age of data I think there are some very real opportunities for some organizations to improve their deliveries and their business value provided they can keep the the on the ball and stay clear of the wrong metrics.

We live in an age where data is abundant, but insight is rare. The organizations that thrive won’t be the ones with the most dashboards — they’ll be the ones that understand which numbers matter, and which ones are just noise. If we can shift our focus from tracking activity to understanding value, we might finally start measuring what truly counts.

Clearing the Mechanism

The Slow Exit: Reclaiming My Digital Life

For most of my life, technology was something that energized me. The early web felt alive — personal blogs, RSS feeds, small communities, native apps that felt crafted rather than assembled. It was a time when the internet felt like mine, and my computer felt like a personal space rather than a corporate terminal.

But somewhere along the way, that world disappeared.

The Collapse of the Personal Web

The personal web didn’t die because people stopped caring. It died because the environment that supported it was dismantled. RSS vanished. Blogs faded into obscurity. Platforms replaced websites. Algorithms replaced communities. And the open, creative, weird internet I loved was swallowed by Facebook, LinkedIn, and other engagement‑driven giants.

The result is a web that feels less like a neighborhood and more like a shopping mall.

Windows Lost Its Soul Too

The same forces that killed the personal web hollowed out Windows.
Native apps vanished. Design coherence evaporated. Everything became cloud‑first, telemetry‑heavy, and enterprise‑focused. The PC stopped being personal. It became a delivery vehicle for services I never asked for.

The excitement I once felt for tech slowly turned into fatigue.

The Decision to Step Back

I realized I wasn’t “overreacting.” I was responding to an environment that no longer respected my attention or my identity. So I started pulling back.

  • I emptied my Facebook account.
  • I stopped posting.
  • I removed photos, posts, likes, and groups.
  • I turned my profile into a quiet, minimalist shell.
  • I’ll do the same with LinkedIn.

Not as a dramatic exit.
Not as a statement.
Just a quiet withdrawal — a way to reclaim my mental space.

Minimalism as Self‑Preservation

This isn’t about abandoning technology.
It’s about abandoning bad technology.

I’m keeping the parts that serve me.
I’m removing the parts that drain me.
I’m shrinking my digital footprint until it feels calm again.

A minimalist Facebook profile.
A quiet LinkedIn presence.
Less noise.
More ownership.
More peace.

What Comes Next

Maybe I’ll rebuild a small personal web presence — something simple, something mine. Maybe I’ll move more of my digital life to tools that respect me. Maybe I’ll switch platforms entirely.

But for now, the goal is simple:

Make technology small again.
Make it quiet.
Make it personal.

One platform at a time.

Not every AI is created equal

At least, it doesn’t seem that way.

For a while now, I have been playing around with AI generated mock-ups. Mock-ups are a great way to communicate functionality to a development team, and with AI, I can do it fast and easy.

When I started I loaded my AI with the thoughts I had about the design; just your good old fashion prompting. I wanted the apps to look like they belong on Windows 11, and since Fluent design, much like Material You, is very well defined, I fed the design principles into the AI.

I want a suite of apps, much like Windows Live Essentials of old, so I did most of the design in the the chat. I figured it would keep the design principles consistent across the various apps. It did, which is something, but lately it seems like my AI is mixing things up more and more. For instance, when it rendered the blogging app, it add the Gmail icon, for some odd reason.

Now, I just used the standard Copilot app that we happy Microsoft365 subscribers get to enjoy. Because, why not. However, I must say that it takes quite a bit of work to make it behave the way I want, and I sometimes wonder if I am too lazy when prompting it.

Github Copilot does not seem to have this issue, however, it is a very different use case, and I usually use the Claude models in VS Code anyway. Could this have something to do with the quality of the model? Could a better model have better hallucination detection? The Internet says yes. Not surprising at all. In a way, it seems to resemble the same issues we see in search engines. Google is just better than Bing, due to better superior search accuracy, larger index of the web, and vastly higher data volume, which powers more precise, personalized, and relevant results. And the same seems to be true with AI models.

Does this mean we should not use them? Hardly. But I think it is important that we look at AI as any other system. That means we need to define what problem it solves, how it solves that problem, and which problems it doesn’t solve.

At the end of the day, AI is a system, not like any other system, but a system nonetheless.

Cleaning up the blog

I finally got around to cleaning up the blog. Turns out it was pretty straight forward.

I started this many years ago during the early Web 2.0 days, so most of the posts were uncritical regurgitation of already known stuff. So, in one fell swoop I deleted all that stuff leaving the blog woefully small, which is the point.

Going forward I won’t be writing a lot, I will use it more as a place to store the ideas I get when I mess around with the various AI solutions.

Citadel Writer. A spiritual successor to Windows Live Writer

I have been thinking about making a desktop blog editor like Windows Live Writer, and now that AI can do it for me, I think, I have been tinkering with some mock ups.

The idea is that it should be a Windows 11 native app, that looks like something belonging on that OS. There’s no doubt that Microsoft has been ignoring Windows for years, and that is the reason I wil end up switching to mac eventually, but until that day, I can play around with the idea of making cool Windows apps.

My design is a bit over the top at the moment, but I am going to publish it anyway. Chances are I will never get around to acutally doing anything.

If I go forward, I will probably tone down the effects to make the UI a little more subdued.

This is getting out of hand

I have been using AI to code for about two weeks now, and I cannot seem to stop. Perhaps it’s more it’s like I could keep going forever, or as long as I have tokens.

So far I have been adding to my Citadel project. I can now get VIX data and I have made upgrades to the UI. The more I use Material You, the more I like it.

The next module I will be puting into dashboard will be a simple RSS reader that shows articles from various financial news sources. I am reviewing the OPML file at the moment, but I may allow the user to add feeds instead of me providing them.

For the time being my old laptop can handle the workloads, but pretty soon I will have to upgrade. As usual I cannot decide in which direction to Mac or Windows, but I guess time will tell. I do want a 4K monitor though. Maybe even a curved one. I use one at work and the curved design is surprisingly good

AI based dev project

Oh my. I never thought I would revive this blog in a million years. Essentially the posts were crappy as hell, and did not reflect any real knowledge based value on my part, so why even leave it up?

Anyway, with the arrival of AI, and inspired by a brilliant co-worker, I have decided to start my own development project.

It is nothing fancy at all. Just a dashboard that can be used for personal finances. The logic is fragmented at best and the use case could stand to be polished substantially, but that’s not the real point here. I want to build something by myself, and for myself.

So far I have a working frontend and a working backend. Right now I am trying to pull data into the Market Metrics module, but I seem to have issues with my environments.

For now I just want to share the first iteration of my dashboard.

It’s a React frontend using Material 3, which I really like. I have serious Mac envy, but since I am still on a Windows/Android product cycle, I decided to use Google’s design language. People tend to go for the Apple look, which makes sense, but I wanted to go in a different direct. Plus, it looks really awesome in my humble opinion.