Skip to main content

Posts

Showing posts from 2026

The Saas-pocalypse. Opportunity or doom and gloom for software as a service?

An ex manager of mine brought up the term Saas-pocalypse a few weeks ago and it was the first I heard of it. Nowadays, instead of paying for Software as a Service, you can log into an AI agent and have it build a website for you that does exactly what you (think you) want in mere minutes. As a software developer, some of it does look like magic and to be honest it does worry me to an extent. However, I don't think the role of software developer is going away, though it might be changing and I will explain why. There are some companies that are going to be affected. If your Saas has very few touch points or deep interaction with your users then you probably will not survive. But here's the thing, AI probably was not what killed your Saas to begin with. Web sites and web apps have been easy and cheap to build and prototype for quite a while now, AI, has just made it a little bit easier and cheaper (ok, a lot easier and cheaper). Just having a website is not what builds a business...

AI has killed the DSL. Long live the DSL!

(This article is pure opinion and is not based on any facts what so ever...) At present (in April 2026) not only are most programmers not writing code directly, we are not writing Domain Specific Languages directly either. A domain specific language, unlike a general programming language, is a language written to solve a specific problem. Often a DSL is translated into a general purpose language (either by interpreter or compiler) and exists to help a programmer read code and reason about a problem in a more readable way. One example that comes to mind is a Arel which is a DSL for writing SQL queries. While SQL is already quite user friendly on the surface, queries can become complex quite easily after a few joins. Also, there are slight differences in SQL depending on which database you use (e.g. Postgres vs MySQL vs Oracle) so having a DSL smooths over these changes as well. However, sometimes achieving what you want to do in Arel can be harder than doing it in raw SQL and also Arel ...