Creating AI That’s Actually Intelligent: Machines That Learn in the Real World, in Real Time

What if machines could learn like humans and other animals, continuously, in the real world, in real-time? No pre-training. No simulations. No backpropagation. Just direct learning from interactions and raw experience.

In this keynote, Viktor Luthman, CEO of IntuiCell, unveils a fundamentally new software architecture inspired by biology. One that enables any machine (physical or digital) to continuously learn and adapt directly in the environment. This is a ground-up rethinking of how neural networks can learn from data streams of real-world interactions, freed from the constraints of static data and brittle models. Drawing from breakthroughs in neurophysiology, IntuiCell’s architecture unlocks generalisation without retraining and enables machines to adapt on the fly in dynamic, unpredictable conditions.

This talk challenges conventional AI assumptions and explores how real-time learning can unlock a new era of embodied AI, transforming robotics, edge systems, and beyond. Not just a new advanced tool. A new path toward embodied intelligence.