Why would people buy software?
All software is a commodity.
Everyone could always build it themselves.
But they didn’t, because the intrinsic complexity and cost of building a solution were high. Much higher than what it costs to pay someone else to deliver a solution.
Generative AI reduces the cost of building solutions exponentially. It allows people to build exactly what they need to solve an exact problem at an exact time. It lets people take ownership of their own solutions.
This is great for many specific problems that require specific solutions that normally cannot be solved easily. This has been the evergreen promise of computers, programming and hacking. But there is a difference between solving your specific problem and owning a problem domain.
Paying for software is not paying for a solution. It’s paying for someone else to take ownership of a problem.
It’s paying for someone who has the taste and context to think about the details. For the activities and structures needed to scale it, maintain it, and solve even bigger things.1 For the relationships that connect, abstract, and extend the problem in ways that are not possible because they own the specific solution, not the entire problem.
Leaders determined to use generative AI to build everything in-house are asking their companies to own a broad portfolio of solutions, and with them a broad portfolio of problems. But when a solution fails, when new capabilities are needed, or when an edge case arises, constant judgment is needed. It requires problem ownership. And a company that maintains its own solutions focuses its attention on those problems instead of the problem it’s actually selling. It allows itself to become unfocused.
And besides, if every company builds all their own solutions instead of buying software, who can they sell to?
—
1 Thanks Charles Vestal.
#pay