The current hype in academia and industry is to use AI to augment, or even replace, traditional software engineers. But we should ask ourselves: is this putting the cart before the horse? The raison d’être for many software engineers is to create enterprise software that makes end-users more productive. What if we cut out the middleman and used AI to empower end-users to create their own automations directly?

In this talk, we will explore how to build a novel programming platform for end-user programming by treating LLMs as neural computers (Automind), with a Prolog-like reasoning language called Universalis as its Mentalese. If we are going to let end-users run AI-generated code autonomously, we must design safety and correctness into the language. That’s why Universalis is kept simple by design, enabling the creation of provably correct software—seen by many as the only path to controllable AGI.

Erik Meijer has been trying to bridge the ridge between theory and practice for most of his career. He is perhaps best known for his work on, amongst others, Haskell, C#, Visual Basic, and Dart programming languages, as well as for his contributions to LINQ and the Reactive Framework (Rx). Most recently he is on a quest to make uncertainty a first-class citizen in mainstream programming languages.