Software is Eating the World ... NOW!
How AI lowers the threshold for creating software and opens a new frontier
When Marc Andreessen wrote “software is eating the world” back in 2011, it was a powerful metaphor. It captured a truth that businesses in every sector were being reshaped by software. Since then, this prediction has played out almost completely.
But the story doesn’t end there. In fact, the second act is just beginning because software is now eating itself.
Generative AI is fundamentally changing what it means to “build software.” It automates parts of design, implementation, and deployment to such a degree that the barrier to creation is collapsing. Suddenly, building small, temporary, and personal applications becomes economically and technically feasible.
That’s why the real “software revolution” might be happening right now.
The collapsing cost of creation
For decades, software development came with high fixed costs. You needed specialized skills, coordination, infrastructure — and patience. Building something small rarely made economic sense.
Generative AI has changed that equation.
Tools like Cursor, Claude (Sonnet), or GitHub Copilot now transform natural-language prompts into working applications. What once required a team can be done by one person in a morning.
The marginal cost of software creation is approaching zero.
That doesn’t just make software cheaper, it makes it different. Entire classes of applications that were previously unthinkable or uneconomical are suddenly viable now.
The rise of micro-apps
We are seeing the emergence of a new category: micro-apps.
These are not smaller versions of traditional software. They are purpose-built, short-lived, and often made for a single user or a single situation. They are designed to solve a problem, not to scale.
Maintenance and longevity, the traditional measures of good engineering, lose importance when it’s cheaper to rebuild than to maintain.
In this world, software becomes disposable. It exists as long as it’s useful. When the need changes, it’s replaced, not refactored.
This marks a paradigm shift: software stops being a product and becomes a momentary expression of intent.
A new layer of digital reality
The first wave of software ate industries. The next one eats behavior, the daily patterns, workflows, and micro-problems that define human activity.
Software becomes ambient, contextual, and hyper-personalized. Each task can generate its own tool. Each problem, its own temporary app.
But this isn’t about smaller bites, it’s about a new digital stratum. A layer of transient, context-aware logic woven into everything we do.
Credits to Oliver Wehrens, Stefan Schubert-Peters, Daniel Schmeiß and Sebastian Heide-Meyer zu Erpen. Without a Sunday morning text-chat with them this article wouldn’t exist.