✏️ My blog posts
Amazon as experiment - on e-commerce, department stores and fish. Link
Computers that can see. Link
🗞 News
The EU is working on regulation for how platforms remove harmful content, following the UK government's 'Online harms' white paper earlier this year. This issue will spread, it will come from different ideas of what free speech means to the presumptions made in the USA (no-one outside America cares what's in the American constitution), and a different organisational model of regulation (mostly outcome-based rather than the US rules-based model), and there could be 'lowest common-demonimator-effects' - 'local' laws will have global consequences. By default, of course, all of this raises the cost-base of running a UCG network (which may for example mean you need to get to revenue quicker and grow slower), and entrenches the incumbents, which is a direct conflict with what other regulators in the same building might be working on. Link ($)
Meanwhile, Australia is also doing a competitive review of internet platform companies. Link (PDF)
Cruise (GM) is pushing back its autonomy timeline. Link
Softbank raised a second $100bn 'Vision' fund, this time apparently without Saudi money. (I don't normally do funding announcements in this newsletter, but a hundred billion here, a hundred billion there... pretty soon you're talking about real money). Link
The NY Times is involved in a project to use bitcoinish tech for provenance and verification to images and video/audio. The problem is real - it will soon be trivial for anyone to create convincing fake video of public figures, and news organisations are seriously worried about being tricked by one of these. But a fake could spread just as easily on social (or inside encrypted messaging), and would normal people care if it wasn't verified? Link
The US government has ordered the removal of (cheap, commodity) Chinese security cameras, but no-one knows where they are. Link
🔮 Reading
Jeff Jarvis: amidst all the moral panic over tech, can we please have some evidence before uncritically repeating stuff about 'filter bubbles' or 'screen time'? Link
WSJ story on the high-end feature phones ('smart feature phones') that fill the price gap below low-end Android (~$50), especially in India. Historically, Moore's Law and economies of scale tend to squash 'bridge' products like this, but there's often a time window before the 'right' product gets cheap enough. Link ($)
Good FT story on the spread of data labelling shops in cheap-labour economies. Machine learning gets rid of the call centres and replaces them with labelling centres. Link ($)
Apparently, 'Famous Birthdays' is the blue check/wiki for the people who are celebrities to the under-20s. Link
Growing pains in Amazon's TV business. Link ($)
It's worth looking at the slides in Sony's latest IR day relating to vision and sensing. Building image sensors for machines instead of for people. Link
The role of AI in fragrance. Everything will use 'AI', just as everything today uses 'databases' or 'software'. Link
😮 Interesting things of the week
Queuing for hours to take the 'perfect' photo of Antelope Canyon. Link
Nice NYT profile of Teenaged Engineering and the $60 synth. Link
📊 Statistics
Useful Ofcom data on where people get news in the UK. Link
WhatsApp now has 400m users in India. Link
|