Britain must not repeat the treatment of Tommy Flowers in the AI age: David Richards
I would bet that most of you have never heard of Tommy Flowers. So let me tell you that he built the world’s first programmable electronic computer, that his machine helped confirm Hitler had swallowed the D-Day deception, and that he paid for much of it himself because his superiors thought the idea fanciful.
You would expect him to be a household name in the country that produced him. He is not. The government that employed him saw to that.
Flowers was a bricklayer’s son from Poplar who served his apprenticeship at Woolwich Arsenal, took a London degree at night school and joined the Post Office in 1926, and by 1930 he was at Dollis Hill, working on telephone exchanges and replacing mechanical switches with valves. It was unglamorous trade work of a sort the professions have never much respected, and the one piece of expertise the country would shortly need.
At Bletchley Park he was the Post Office engineer among the university mathematicians, and the scepticism he met there was as much about station as about physics.
Told that a machine of a 1,000 valves would be too fragile to run, he knew otherwise, having learned on the exchanges that valves fail when you switch them on and off, not when you leave them alone.
The project was shelved anyway, so he built the thing himself, in ten months and partly out of his own pocket. Colossus attacked its first Lorenz message in February 1944, and by June a faster version was confirming that the German armour was waiting in the wrong place.
Britain then broke up eight of the ten machines and ordered Flowers to burn his drawings, which he did in a boiler at Dollis Hill, and bound him and hundreds of colleagues to silence for 30 years. The reasoning was narrow rather than stupid. Whitehall had commissioned a codebreaker, the codebreaking was finished, and a secret weapon is something you lock away. Nobody in authority appears to have asked what else the thing might be.
In Philadelphia they asked. On Valentine’s Day 1946 the Army and the University of Pennsylvania unveiled ENIAC to the press and let the newspapers call it an electronic brain, and that summer the Moore School ran a course teaching anyone who could get a seat how to build one.
Eckert and Mauchly left to found a company, IBM turned its research towards electronics, and within five years there was a commercial computer industry, American and no accident. Washington had understood what Whitehall missed, which is that the machine was never the asset. The knowledge sat in the engineers' heads, and it compounds only when allowed to spread.
Among those who sailed to Philadelphia was Maurice Wilkes of Cambridge, who sketched EDSAC on the voyage home, a British scientist crossing the Atlantic to be taught principles a British engineer had proved two years earlier in north London and was forbidden to mention.
Flowers paid for that narrowness himself. He went to the Bank of England for a loan to build another machine and was refused, because the bank did not believe such a machine could work. He had built ten of them and could not say so.
He received £1,000 for his war work, less than he had spent, shared most of it with his team, and was appointed MBE. His family knew only that he had done something secret, and knew only that until the 1970s.
Britain’s problem has never been a shortage of ideas. It is that we treat the invention as the achievement and the industry as somebody else's affair, which is how DeepMind and ARM came to fly other people’s flags.
Britain says it will lead the world in artificial intelligence. The question is who will own it when we do, and who we are quietly asking to burn the drawings this time.
David Richards MBE is managing partner of Yorkshire AI Labs