A world after software

Let's run a thought experiment. It's 2031. Every tool your company works with — project management, CRM, calendar, analytics, docs, chat — is open source. Those who want to self-host do. Those who don't, rent it in the cloud at exactly what the infrastructure costs, with zero margin.

This isn't science fiction: it's the extrapolation of two trends that already happened. AI pushed the cost of writing software to nearly zero, and the cloud always had a public price. Put the pieces together and watch what happens to everything else.

The IT budget collapses (and that's fine)

Today, a 50-person company pays around $30 per user per month to move cards, $50 to store contacts, $12 to shorten links. Add it up and you easily get one extra salary, every year, for software whose marginal cost is zero.

In 2031 that budget line doesn't drop 10% or 30%. It collapses to its real cost: cloud, backups, bandwidth. Cents per user. What's left of the IT budget moves to where it should have been all along: people who understand the business, owned data, security, custom automation.

For the CFO, this isn't a tragedy. It's the first time the "software" number on the books tells the truth.

Software stops being an expense line

When something becomes too cheap to meter, it stops being discussed in board meetings. Nobody debates the water bill. Work software is heading there: a utility. It's there, it works, it costs what the electricity to run it costs.

Sound like a loss? On the contrary. Companies will run more software, not less: the tools they don't buy today because "the license doesn't justify itself" will exist anyway, self-hosted or at cost. The bar stops being "what does it cost?" and becomes "does it help us?".

Innovation moves somewhere else

When copying costs zero, value concentrates in what can't be copied:

• Data and trust. You can't fork a customer relationship.
• Hardware and the physical world. Chips, robots, sensors, energy.
• Models trained on what you own. AI on your data, on your infrastructure.
• Judgment. People who decide well with tools everyone has.

Companies selling software seats today have two paths: become honest utilities — charge for running what runs — or defend a margin that no longer holds up. We know how that second story ends: we watched it with music, with newspapers, with cable TV.

Software is dead. Long live software.

This closes the arc we opened with the manifesto. Software died as a scarce product rented per seat. And it was born again as what it should always have been: free knowledge running on at-cost infrastructure.

In 2031 nobody will say "we bought a CRM". They'll say "we use a CRM", the way we say "we use email" today. The verb changes because the owner changes.

We've already started: Link is available today, open source or in the cloud at cost. Track, CRM and Cal are on the way. A world after software isn't something you wait for: it's something you build. Link by link.