Apple is boosting encryption of data in iCloud, a response to ongoing data breaches.

The Pentagon has awarded its Joint Warfighting Cloud Capability contract to four companies: Microsoft, Oracle, AWS and Google. That’s good, inasmuch as it lessens vendor capture and holds out the prospect of benefiting from the tech capabilities of each. The challenge will be to exploit those tech capabilities to best advantage, manage the overhead associated with multiple vendors, maintain competitive tension and avoid the prospect of collusion between vendors.

MIT Tech Review article on Saudi Arabia’s Line project, based on newly available imagery–and a good illustration of the commercialisation of space, satellites and imagery, and how quickly imagery can be made available.

Like many others, I’ve been impressed and rather alarmed by ChatGPT. It’s remarkably capable generated the sort of material that is the ‘meat and potatoes’ of the public service for example. But: ChatGPT makes stuff up–though it would argue it cannot, as below. Nonetheless, as illustrated by a rather alarming thread, it can seemingly generate entirely new, nonexistent fields in science and with confidence and erudition.

Wired: what the demise of Twitter will mean. Twitter has been at times infuriating, inane, ocassionally a wateland, always a distraction. But at other times, magic. It’s where you could often find news not covered elsewhere–certainly by the mainstream, and particularly so in Australia. I’ve found it a useful research tool, allowing the uncovering of insights and perpsectives not otherwise or readily available: it’s a series of signposts, indicators, rather than a treasure of deep insight in and of itself. It made connections, and let you see there are truly brilliant, funny, insightful and caring people in the world. And, as the Wired piece argues, it has served an important role in society that may well be lost.

Computing upgrade delays on the F-35 are a reminder that software is hard. Lockheed builds planes–hardware, kit. Those are tough engineering problems themselves. Software is different again. Integrating the two tougher still. But software, digital systems, done well look like magic and can encourage us to under-estimate the complexity, effort, and path dependency, of its development.