We are in the midst of a shift that is fundamentally redrawing the map for how we view software in our organizations. For most of my career, the truth has been simple: programming is expensive and engineering time is scarce. That's why we've bought SaaS solutions (Software as a Service) to spread development costs across thousands of users.
But the rules of the game have changed.
As AI now pushes the cost of writing code toward zero, we must rethink what we buy and what we build. We're moving toward an era of "Disposable Software" – and to navigate this correctly, we must understand that the concept is now more important than the code itself.
From Scarcity to Abundance
The traditional SaaS model is built on solving general problems for the masses. Salesforce, Slack, and Jira are built to work "good enough" for everyone. But we all have those specific, niche problems in our teams that SaaS tools never really fix – those gaps in the process that we usually fill with Excel spreadsheets or manual work.
This is where AI-generated applications, or what we call "AI building," come into play.
Since generating code is now cheap, we can create tailored micro-applications to solve a specific problem for a handful of people in the organization. Do three people on the marketing team need a tool to track social media sentiment this week? Ask an AI to build it. Use it. And here's the key: be prepared to throw it away when it's no longer needed.
We need to stop seeing software as a monument meant to stand for ten years, and start seeing it as "executable artifacts" – tools that live as long as the problem lives.
Concept and Taste Are the New Bottleneck
When AI takes over the heavy lifting of writing code (execution), the value shifts to the human. The bottleneck is no longer how we build something, but what we should build.
This means your most important skill is no longer technical syntax, but clarity and taste. It's about having the judgment to know what's worth building and the ability to steer the AI toward that vision. In a world of disposable code, it's the concept that drives; the code is just a temporary solution to realize the idea.
The Balance: Concrete Foundation vs. Disposable Tools
But – and this is a big but – we can't build our entire businesses on disposable code. There's a reason SaaS is still critical.
We need to distinguish between infrastructure and tools.
SaaS is your infrastructure: For critical data, security, and long-term stability, we still need robust SaaS solutions. It's the "concrete foundation" that we don't want to be fragile or require maintenance we can't handle.
AI apps are your temporary tools: For the specific, quick, and niche, we can use AI. But we must be aware of the maintenance risk. AI-generated code can be "fragile" and difficult to maintain over time if it breaks.
The strategy forward is about finding this balance. Let SaaS handle the heavy, boring, and stable ("bland chicken" functionality). Then use AI to quickly build the unique, concept-driven layers on top that give your team superpowers in daily work – even if the app is only used by five people and deleted in a month.
The future belongs not to those who write the most code, but to those who have the best concepts and the courage to build, test, and discard.