In November 1984 I attended Middlesex University’s AI Weekend, which was dominated by discussion of the Reith lecture by John Searle, titled Beer Cans and Meat Machines, broadcast that week. In it, Searle questioned whether computers could ever truly think, imagining a machine made of beer cans that, if correctly programmed, might on the theory of “strong AI” possess a mind.
Though he used this to argue that syntax alone cannot create understanding, his challenge invigorated debate about strong AI. It prompted researchers to explore how meaning, embodiment and learning might emerge from computation itself. Four decades later, with neural networks and large language models achieving feats once thought impossible, his questions remain central – reminding us that today’s AI stands on deep philosophical foundations.
