The Hidden Logic Behind Everyday Tech

Imagine a Sudoku grid. If I fill it in, you can easily check whether it is correct. But finding a correct solution can be much harder. Computers face the same difficulty: they can quickly check a proposed solution, but no fast general method (algorithm) is known for finding a correct solution that works efficiently in all cases. I will explain why finding such a method would be enough to solve one of the biggest open problems in computer science, worth one million dollars. This problem, known as P vs NP, asks whether every computational problem whose solutions can be checked quickly can also be solved quickly. I will also introduce some other hard problems alongside Sudoku.

It’s the 2000s. You’re preparing a mix CD for your friend. How did it become totally normal to literally burn microscopic patterns into a plastic disc at great speed and phenomenal accuracy using off-the-shelf hardware? What does this tell us about the power of industrial research and development, standardisation, and software? The answer to these questions takes us on a journey through the history of technology and comping, requires a healthy dose of media studies, and promises a nostalgic reminder of when computer users still concerned themselves with the “stuff” of physical media.

Artificial intelligence (AI) is changing healthcare by supporting earlier detection, prediction, and decision-making. This talk presents my research on AI in clinical practice, including risk stratification for cancer chemotherapy patients, natural language processing to audit discharge communication and hospital-to-community handover, and models to predict patient deterioration in A&E. It also covers skin cancer prediction. Key issues including data quality, clinical integration, and oversight will be discussed, with a focus on how AI can support clinicians and improve patient care.