A Geek’s Guide to the hottest place in 20 parsecs

This week, The Register published a feature by me on the Culham Science Centre in southern Oxfordshire. This includes what staffer Guy Burroughes describes as “the hottest place in 20 parsecs”: the heart of the Joint European Torus, Culham’s biggest project and the world’s largest nuclear fusion tokamak, which at its hottest has reached 300 million degrees Celsius, many times hotter than the centre of the sun. (It’s rare to find a star this hot. 20 parsecs is about 65 light years, although it’s just possible that Mr Burroughes was getting in a Star Wars reference.) Continue reading “A Geek’s Guide to the hottest place in 20 parsecs”

The many problems with automated decision making

I have written about automated decision making or machine learning for Computer Weekly, in particularly the numerous problems with using it. The biggest set of issues is summed up nicely by Joanna Bryson of Bath and Princeton universities: “The reason machine learning is working so well is it is leveraging human culture. It’s getting the bad with the good.”

Continue reading “The many problems with automated decision making”

Keeping paper voting: right policy, wrong reason

In six days, Britons will use stubby little pencils to put crosses next to people’s names on pieces of paper. In each of 650 areas, the person with the most crosses becomes a member of parliament. If more than half of those MPs come from one party, that party forms a government and its leader is prime minister. It’s easy to understand and trust.

If someone hacks your bank account, you will notice and will probably be able to get recompense. If someone hacks an election, you are unlikely to know unless your votes was published – which would rather undermine the concept of a secret ballot. Also, we can all understand people counting pieces of paper. Very few of us, including apparently many NHS organisations, can say likewise for computer security. Continue reading “Keeping paper voting: right policy, wrong reason”

The paranoid style in IT journalism

There’s a type of technology journalism which is the non-fiction equivalent of a horror story, the opposite of uncritical technophilia. Scaring readers is a pretty good way to hold their attention, but it doesn’t necessarily get to the heart of the story.

Earlier this month, The Register published my piece on Cambridge Analytica and its targeted political advertising based on online psychometric profiling. The company does itself no favours by not responding to questions, but I did speak to academic Michal Kosinski, who according to a widely-read article in Swiss publication Das Magazin (republished by Vice’s Motherboard section) ended up racked with guilt at having developed ideas in this area which Cambridge Analytica has since exploited. Continue reading “The paranoid style in IT journalism”