Digital60: Kilburn Lecture

Apologies for this post being some 12 hours late, however fatigue and exhaustion were evident in your correspondent by 9pm last night.
I felt a real shift in focus and energy leaving bTWEEN and entering Crawford House for the Kilburn lecture night. The collection of research posters dotted around the foyer indicated this was a much more academic and research-focused affair.
The reason why I was present has been discussed in the paper already – on the 21st June 1948, a machine sat at Manchester University officially called “The Small Scale Experimental Machine” (SSEM for short), and known informally as “The Baby” did something no machine had ever done before.
It stored in its memory a program. It executed that program correctly. It stored the results in memory.

It sounds simple, but the reality of creating the technology to make that happen in the late 1940s is quite incredible. The entire machine had been built as a testbed for a radical idea: no longer did programmers need to stack up paper punch cards, or hardwire their program. It was possible (the theory said), to produce a universal, programmable machine. And so it was.
That first program was written by Tom Kilburn, and it is in his name that the University dedicated a lecture to celebrate the 60th anniversary of the occasion.
Before the lecture itself, medals were presented to the surviving four members of the team that worked on the original machine and Ferranti Mark I (the Baby’s successor), as well as an additional medal to the man who built the replica of the Baby ten years ago which still stands at the Museum of Science and Industry.
Professor Steve Furber CBE who gave the lecture has himself played an interesting part in the computer industry in the UK. Back in the 1980s he worked on the design of the BBC Micro, a machine that inspired a whole new industry to form. Few people know that in terms of units shipped, the British firm built off the back of the BBC Micro’s success (ARM Limited) beats all other companies combined even today.
His talk developed a theme through the research activities of the University over the past sixty years and looked towards the future.
The very simple way to summarise the progress made over the last sixty years in research efforts is by comparing the size of the machines built then and now, their performance, and their energy efficiency. With each decade that has passed the technology has become faster, lower power and smaller. He offered up a comparison between The Baby and the ARM968 processor (a modern British piece of technology) to help make things clear:

In other words, power efficiency has improved by 50,000,000,000 times since 1948. Furber pointed out that if car engines had improved in the same manner over the same time the entire UK’s car population would be able to move around for a year on just one litre of oil. He also accepted there are good scientific reasons why that analogy was a little weak.
The lecture moved on to discuss the future, and specifically the research being done at the University today. His currenct project named ‘Brain Box’ (by the University’s marketing department apparently), aims to connect hundreds or thousands of “nodes” and emulate neurons within the human brain. There is a great deal of research within this area right now, and Furber is confident that if he gets all the funding he needs he will be able to emulate a brain equivalent to that of several rats brains.
The eventual goal is make progress in the ‘Grand Challenge 5’, a challenge from a funding body to promote the research in the UK which requires the production of a robot able to emulate a range of behaviours and subset of the capabilities of a child aged between 2 and 5. You might have realised already that there is some considerable debate about this age range – it is much easier to produce a robot that acts like a 2-year old than it is a machine mimicking a 5-year old.
Much of the work being done today and over the next few years appears to be positioned to be as important to the next six decades of computing as the Baby was to the last, and I watch with interest. Whilst the thought of producing artificial robotic brains might worry some, I for one welcome our new silicon masters.