ABSTRACT
International meetings are the principal perk of academic life. Free travel across the world; chat; chat-ups, and of course, gossip – though sometimes of the Sour Grape vintage. Presumably to avoid too many distracting temptations, conferences are often held in isolated places – a gamble between close-knit intense discussion and claustrophobic boredom. Worst of all, when the organizers of such meetings are demons for work, they allow no spare time to let what has been said sink in; so one ends up confused and unable to remember a thing except perhaps the odd joke at the bar. Then discussion hardly develops, as there has been no time for thinking. Some meetings, though, are memorable, and just a few vitally important for launching a new idea or even initiating a whole new subject. No one who attended the ‘Mechanisation of thought processes’ conferences, at the National Physical Laboratory at Teddington – though as long ago as 1958 – could possibly have forgotten it. This meeting, with a few others in America around that distant time, launched a new way of thinking about brain function and mind, in terms of neural nets and computing. Marvin Minsky spoke on ‘Artificial intelligence and heuristic programming’; John McCarthy on ‘Programs with common sense’; Albert Uttley on ‘Conditional probability computing in the nervous system’; Rosenblatt on Perceptron seeing-machines; Selfridge on ‘Pandemonium: a paradigm for learning’. Most memorable of all was Warren McCulloch on ‘Agatha Tyche: of nervous nets – the lucky reckoners’. This was the first time I had met Warren, and he 24remained a hero and I am happy to say a friend. Then there were papers on machine translation from Russia, and another hero-friend, John Young, chairing ‘Implications for biology’. In my experience this was entirely exceptional as so many of these ideas were new and proved to be important, and so many of the stars one saw for the first time remained navigation beacons guiding a generation of research into the future of the brain sciences and computers.
