What has value when anyone can build anything?

Everyone is building “all the products”.

If an agent can build it in a couple of hours then anyone can build it in a couple of hours. No edge. No moat. And it’s only getting simpler, faster, cheaper.

Someone figures out some cool software to build. They ship it. People see it. Instead of buying it they get an agent to build it for them.

So the real question isn’t “what should we build”.

The real question is what is valuable when everyone can build anything.

Some say it’s the model layer. I don’t think so. Models are becoming a commodity. You can even run fantastic open source models locally.

This means the only cost for creating whatever software you need is hardware and electricity.

So what has value in a couple of years?

It’s wherever there are scarce stuff. Like:

Distribution Placement, defaults, channels, partnerships. Being in the place people already are.

Trust Security, privacy, reliability, audits, compliance, liability. Software is easy to copy, trust isn’t.

Data Unique data that keeps updating. Data that comes from real usage, real edge cases, real feedback. Data that makes the system better tomorrow than it was today.

Networks Marketplaces, reputation, standards, coordination. Code can be cloned. Liquidity and reputation can’t.

Operations Owning outcomes. Not “here’s the product,” but “here’s the result.”

Nobody will pay for an invoice generator. Plenty will pay for the system that already holds their vendor relationships, compliance records, and payment history.

The winners won’t be the people who can ship features the fastest.

The winners will be those who can build strength in one of the areas above.

In other words: the stuff that can’t be prompted into existence.