AI Could Break The Link Between Jobs And Income

As AI grows more capable and takes on more tasks, and increasingly more jobs, the question is no longer what some people will do. It is what everyone will do.

Some people might be needed, but what about the vast majority of those employed today? How can they all possibly be needed when it is plain as day that much of what they do now can be automated? Denying the evidence or expecting a modern Luddite movement to reverse technology adoption is impractical and dangerous. Preparing for this transition and developing a national-level plan to navigate it while preventing social unrest should be the key consideration.

But before we speak to the macro impact, let us examine the details. Today, software licenses generate about one trillion dollars in revenue, with another four trillion coming from related services and maintenance. This is nearly five trillion dollars of revenue caught up in the production of a thing that is increasingly being automatically churned out by today’s AI. How then can we imagine that this sector will survive as it stands today?

The world’s most popular operating system, Linux, is already free. Databases are free. App servers and programming languages are free. If an AI model with the capabilities of Opus 5.5 becomes free, in say two years, then the entire app stack on top of these foundational components can also be assembled or built for free using such a model. Microsoft Office has been reverse engineered by AI already, as has the entire suite of Adobe applications. I say this as a multi-time CEO: if a CIO or VP of Sales came to me today asking to buy a Salesforce license, a quaint forms-based application that has made billions merely by being entrenched, I would refuse the request immediately. I think most smart CEOs today would do the same. They would direct their VP to find an open source alternative and arm one person in sales operations with Opus 5.5 to build out whatever custom features and reports are needed.