Smart-ish: Why AI Still Needs a Pint of Common Sense

Causality is sometimes an intuitive concept. In one pragmatic sense, having an understanding of cause-effect allows us to answer “what if” questions: what if I start exercising now, will this prevent me from having heart problems within a decade?  Being able to answer such questions, however, is easier said than done. In this lecture, we will explore why this is so, whether you are an AI with a colossal computing budget or a scientist working at the cutting-edge of their field.

The raw power of AI is skyrocketing, but real-world impact still lags behind. Bridging that gap is not just about bigger models, but a shift toward vertical, domain-specialised AI. We are moving beyond chat into cognitive amplifiers: digital scaffolding that orchestrates agents, memory, and safe professional tools. This is the bicycle for the mind in action - AI as support for human potential, not a substitute.

I’ll explore the socio-technical and engineering choices that will determine how much of AI’s progress translates into real-world impact.