Software is becoming abundant

Ideas for software have never been particularly scarce.

The ability to turn them into functioning products has.

For most of the software industry's history, building something meaningful required access to specialist technical capability. Developers were expensive. Good developers were difficult to find. Products took time to build. Experiments therefore had a meaningful cost.

That scarcity shaped the software industry. It influenced which products were created, who could create them, how companies were funded and how large they needed to become.

AI is beginning to change that.

And I think the consequences extend considerably beyond making developers more productive.

Software itself may be becoming abundant.

The cost of making is falling

There have been several stages in this process.

Cloud computing removed much of the need to own infrastructure. Open-source software made enormous amounts of functionality freely available. APIs allowed companies to build on capabilities created by others. No-code and low-code tools reduced the technical knowledge required to create applications.

AI takes another step.

It begins to separate the ability to create software from the ability to write software.

Increasingly, you can describe a product, generate an interface, create a database, connect services, debug problems and deploy working software without personally understanding every line of code involved.

The tools are imperfect. They will make mistakes. Complex engineering still requires expertise.

But focusing on their current limitations misses the direction of travel.

The distance between having an idea and having working software is shrinking.

When everyone can build

Imagine the cost of producing software continues to fall.

We don't produce less software. We produce much more of it.

Ideas that weren't commercially viable become viable. Businesses commission software for problems previously too small to justify development. Individuals build tools for themselves. Niche markets become addressable. Entrepreneurs test ten ideas instead of committing everything to one.

That creates enormous opportunity.

But it also changes where competitive advantage lives.

If your competitors can build roughly what you can build, the ability to build stops being sufficient.

The constraint moves.

Can you identify a better problem? Do you understand the customer more deeply? Can you reach them? Do they trust you? Can you create a better experience? Do you have access to data, relationships or knowledge others do not?

Software still matters.

But software alone matters less.

Value moves around the product

This is an example of the broader pattern I keep returning to: when production becomes abundant, value moves towards the things surrounding production that remain constrained.

For software, that could mean distribution. There may be a thousand products capable of solving a problem. Customers still need to discover one.

It could mean trust. Being able to generate software quickly doesn't establish whether somebody should depend on it.

It could mean domain knowledge. Understanding the problem may become more valuable than knowing how to implement the solution.

It could mean taste. If everyone can generate interfaces, recognising which experience is actually good becomes more important.

And it could mean customer relationships. Technology can scale. Genuine relationships are harder to.

Smaller companies become possible

Perhaps the most interesting consequence is organisational.

If producing software requires fewer people, software companies don't necessarily need to grow as quickly.

A company that previously required twenty people might require five. A product that required five people might require one.

This doesn't simply reduce costs. It changes which businesses can exist.

A niche producing £500,000 a year may be unattractive to a venture-backed company employing fifty people. It could be an exceptional business for two people.

As the cost of creation falls, the minimum market size required to support a software company falls with it.

That could produce an enormous expansion in small, specialised software businesses.

The question changes

For decades, one of the first questions confronting somebody with a software idea was:

Can we build it?

Increasingly, the answer will simply be yes.

That doesn't make entrepreneurship easier. It changes the difficult part.

The questions become: Should we build it? Who actually needs it? How will we reach them? Why will they choose us? What do we understand that others do not?

Software becoming abundant doesn't mean software stops being valuable.

It means the ability to produce it becomes less scarce.

And when that happens, value moves somewhere else.

The interesting question is where.