Runtime

AI is making code cheaper to produce. That is useful. It is also easy to misunderstand. The hard part of enterprise software was never only the act of typing code. Companies do not succeed because a screen exists. They succeed when that screen can execute a real workflow, respect permissions, connect to data, survive change, be monitored, be published, be audited, and evolve without becoming a liability. That is why the next phase of AI software will not be defined by who generates the most code. ...

There is a question every company should be asking in 2026: if AI can already help us build software, why are we still choosing between heavy ERPs and generic SaaS? For years, that was the default tradeoff. On one side, the traditional ERP: powerful, expensive, slow to implement, hard to adapt, and often far away from how the business actually works. On the other side, modern SaaS: fast to start, polished, accessible, but often rigid. It solves the average problem well. The problem is that no serious company wants to become only the average of its market. ...