Artificial intelligence will discover a cure for cancer, which no man would have been able to find during his lifetime. Rene Haas, CEO of Arm Holdings, Cambridge-based chip design company, in an interview with podcast «Big Boss Interview» BBC.

Modelling the way a DNA indicator is affected by cancer remains today too complex a problem, even for computers running AI. But Haas predicts that this will change: «As we feed more and more models on these computers and computers become increasingly sophisticated to run them, they will solve it». He, who left the AstraZeneca Pharmaceutical Board in April, believes that within our lives AI will contribute to treatment of disease.

Different perspective testified Professor Chris Bakal, from the Institute for Research on Cancer of London and CEO of Sentinal4D. The crucial question, he argued, is no longer whether AI will be used, but with what it will be powered. In laboratories such as his, AI is trained in data derived from patient samples rather than from the internet, without need for a giant data center. «The future of AI medicine does not belong to anyone who builds the most powerful computer. It belongs to whoever has the right measurements.», he said, appreciating that this approach can shorten over years the development of new treatments.

Beyond medicine, Haas predicts widespread dissemination of anthropomorphic robots within the next five years, with Arm's chips feeding within a decade of robots learning autonomously in processing, cleanliness, security and construction. Thanks to AI, these machines can see, learn and reprogramme themselves: a robot made to make a bed can equally learn to clean towels or clean the bins.

This rapid development is currently halted by the lack of chips for the construction of data centres. Arm's low consumption technology is already being used in half AI data centers worldwide, while the company recently also passed on to its own chips: Meta asked her to develop the chip Arm AGI and, since its release in March, orders exceeded $2 billion. The restriction on supply delays both plans for multi gigawatt power centres in France and the USA and ideas for their installation in space.

Production today focuses on Taiwanese TSMC and Haas does not believe that it could be transferred to the United Kingdom, despite the relevant government discussions with the industry. It considers that the country does not need its own fabs (chips production plants), as the cost is huge, the needs of specialized personnel and natural resources large, and there is already a wide ecosystem.

On the issue of jobs, he admitted that there would be changes for workers, but he described the assessments of mass disappearance and complete replacement by machines as being offset by new opportunities. It also considers that the concerns about the adjustment to the stock market value of AI companies are equally exaggerated due to long-term demand.

Arm, which this year emerged for a time in a more valuable British-born company in history based on capitalization, designs processors (CPUs) used in hundreds of billions of mobiles, cars, smartwatches and electronic devices. Despite its sale to Japanese investors in 2016 and its partial introduction to Nasdaq instead of the London Stock Exchange, half of its employees remain in Britain, with Haas featuring Arm «the longest employer in Cambridge».



Source

EnglishenEnglishEnglish

Connection

Registration

Restore Password

Enter your alias or email address and you will be sent a link to create a new password.