Dr Laura James's writings

Starting with random technology things: The launch of the open source  Autoware Foundation  (thanks to  Trucks newsletter ) Robotic tiles for virtual reality experiences  — solving a problem I’ve wondered about :) Occasionally there’s something interesting in  Stratechery , although it sometimes feels like the (paid subscriber) newsletter needs careful parsing to remember what particular world view it represents. Just before Christmas it covered  Discord , which I’ve used for game chat a handful of times, but which is far more: The rise of Discord — which was valued at $1.65 billion  last April  — reflects something very fundamental about gaming that I suspect is missed by many: for young people in particular gaming is a social experience. The social nature of much gaming certainly isn’t a revelation for me, although that valuation is :) The real value of 150 million users, though, is as a distribution channel, which is where the D...

Everyone following machine learning / AI stuff will probably have already read these and the inevitable twitter threads, but I thought there was an interesting common thread between  Gary Marcus’s piece about the problem with deep learning  and  Beth Singler’s about AI, storytelling and social intelligence . Marcus questions “the notion that deep learning is without demonstrable limits and might, all by itself, get us to general intelligence, if we just give it a little more time and a little more data” and Singler takes a different angle on what today’s focus on data-driven AI might be lacking. The potential power of symbolic AI, unlocked by the level of processing we have access to today, seems to have been forgotten. In the “tech world” there’s so much hype around machine learning, and the incredible benefits it is supposed to bring, that it’s easy to forget that you need good data to feed such systems, there need to be useful insights in that data, and then you h...

Mostly things I read the week before last. I’m not a big fan of the idea that leaving Facebook is great advice, or a way to change the world if many of us did it, but it seems popular these days. It always feels like a very privileged position, to be able to step out of a social infrastructure which supports so many useful interactions for so many people. Local communities, volunteer groups, friends and family dispersed around the world — these are not things everyone can just give up without sacrifice. And so I was pleased to see  Joanna Bryson writing about these important connections  and the idea of perhaps nationalising it, setting out to fix it together, instead. A  paper from Farrell and Schneier about Common Knowledge attacks on democracy , proposing we take a cyber security approach when thinking about how information can be manipulated on a large scale. Seems a useful analogy, if only to encourage to a new mindset the technical folks who tend to jump to t...