Build Real Things With AI
The launch of Claude Opus 4.6 on 5 February 2026 made building software tools much easier. AI models were suddenly able to write code as well as human beings. I first realised things were changing when I worked with Anti Gravity in December 2025. It was my first interaction with agents which could write code for complex tasks that could work with no editing required. I was genuinely impressed but it did not prepare me for the release of Opus. That is how big the jump in ability was. I honestly view computers differently after that release.
Computers are wonderful machines. I like them. Working with them has in the past been gated through computer programs. A program is a set of instructions enabling a computer to complete a task. Very few people could write these instructions. They were our translators and out of necessity focused on tools which covered the majority of needs or were most profitable. It was not possible to cater for all our individual needs.
AI has made it possible to instruct your computer directly in natural language. No translators needed. This changes how we work with computers. The change has naturally started with the software industry since they created the tool but it is coming to every industry. We are in the very early days of this new reality, literally the days of typing commands in the terminal like the early 90s but it will change fast and soon you will wonder how we ever lived without AI built tools. You can really improve your productivity with these new tools.
Most of the people who can benefit from this increase in productivity are not yet aware the tools exist or have not yet learnt how to use them. The hype and jargon surrounding the communication about this change is unhelpful and keeps people away from testing the waters. No, AI is not going to eat us all (I hope), and learning about context windows, tokens and harnesses is not as hard as you think. If you are reading this, you can probably learn how to build a useful AI agent for yourself. You just require some curiosity, and stick-to-it-ness. I believe the people with those qualities exist all over the world and just need the information served to them in a straightforward format they can apply.
Mwangaza exists to provide that information. Articles, guides and case studies to help people use AI to talk to their computers and build the tools they need. Everything you see here will be curated from the best learning resources on the internet and from our experience building our own tools and solutions. We will just serve it in accessible format and cut out all the hype whenever possible. If you are a school teacher who wishes existing software could fit your particular teaching style, an engineer looking for automating tedious tasks at work, or you just always wanted to create a small computer game, Mwangaza is the place to learn how to use AI for that and meet like-minded builders. Let’s build.
How this was written: Written by Kye. AI used only for grammar checking.
Have feedback or questions? Reply on LinkedIn or email kye@mwangaza.ai.