Britain has reason to try to stay in the game. “No one knows how [the AI race] plays out,” says Guy Podjarny, founder of Tessl, a startup that uses AI to help software developers write code.
The winners could be those with the smartest models—or those building applications on top of them, using the models as a utility. Either way, data and applications will matter, and infrastructure will be needed for Britain to carve out a niche.
Britain’s best chance lies in nurturing AI talent. It is a world leader in machine-learning research. DeepMind, Google’s AI arm, and Wayve, a promising self-driving-software firm, are already based in London’s King’s Cross. The question, says Matt Clifford, until recently the government’s AI czar, is “How do we make this the best place in the world to try radically ambitious ideas?” One answer is to build hubs like Isambard for “frontier AI” model training. Professor McIntosh-Smith says the machine is already drawing researchers from America, Switzerland and the EU.
Isambard is not big enough to train the largest language models. It will, however, enable other research breakthroughs. Dima Damen’s team at Bristol University helped build the world’s largest dataset captured by wearable cameras. Until Isambard, they lacked the processing power to use it. Professor Damen now hopes to train models that can recognise actions and predict intentions—videos which in future may be able to prompt people with dementia (and also help train home robots).
Read more | THE ECONOMIST