Governments have long used radio to change minds beyond their borders. Take World War II, when Germany made broadcasts to Britain: William Joyce - nicknamed Lord Haw-Haw by the listening public - provided the German account of news, 'predicting' which towns would be the target of an air raid that night. Joyce was not German but, as a former member of the British Union of Fascists, he sympathised with them, leaving London for Berlin a week before war in 1939. He was on the air there within weeks.

It was all guesswork and propaganda, but the programmes resonated with the British public, exploiting a desire for information from the frontlines. Many listened - almost as many as listened to the BBC News. Some believed him, and found him endearing. Many wanted another perspective of wartime engagements, while others saw it as light entertainment - a welcome alternative to the dull and serious programmes broadcast domestically. Joyce's familiar accent must have helped here, even if he sounded condescending to some, getting German narratives into homes that would have switched off a German voice.

The Open Graph Protocol, or OGP, is a series of <meta> tags designed to provide metadata about content hosted online, including articles, photos, and videos. It's created by Facebook, which would normally be a big red flag for me - but since it's an open standard, other sites have embraced it too, including Microsoft's LinkedIn and the decentralised Mastodon.

Platforms use OGP data to display rich information alongside webpages their users link to, including photos, bespoke headlines that differ from the page's actual title, and descriptions to elaborate on both of them.

AI is big business, and it's largely fuelled by the vast troves of data on the Internet. Much of it was published in an era when harvesting authors' work to train models that may eventually supercede them, if not unimaginable, certainly seemed like a distant future. That future has now arrived.

Recently, I came across a proposal by the independent search company Mojeek for new ways for website owners to control how AI companies crawl data to train their models with. It's not of much use to me - all my posts are available on GitHub, free of any restrictions I place on them here - but I was intrigued to explore how I could implement it given I don't write any of the HTML for this site myself.

Over the next few days, Signal will stop supporting SMS messages on Android. After almost ten years, Signal's greatest asset for introducing the masses to secure messaging will be removed. In this post, I'll discuss why I think SMS support was so important, and why I think removing it is the wrong decision.

Zola is a Static Site Generator (SSG) that converts text you write in Markdown documents into a website that can be served on the Internet. Being a static site, it's simple and doesn't require a database behind it to serve files from - and, in using Markdown, text is formatted in the same way as note-taking apps such as Notion and Obsidian, meaning you don't need to learn HTML syntax to start writing.

In theory that's simple, but figuring out how to make Zola build a nice-looking, themed site from those Markdown files takes a little more understanding. You might also want to use Git to track the changes you make, and publish the site automatically using a platform like GitHub or Cloudflare Pages.

In this post I'll share how I made this blog, from installing Zola, to adding a theme, and building a site you can see in your browser - as well as learning the basics of Git in the process.