TP The business of what comes next Saturday 21 November 2026 · updated 00:59 Get the weekly briefing
Tomorrow’sPublisher.
AI & Technology AI Technology

Sinclair’s publishing tool revives cross-posting and drafts image descriptions

Noah Wire · 2 min read

Editorial cartoon accompanying: Sinclair’s publishing tool revives cross-posting and drafts image descriptions

Cartoon · AI-generated

A custom editor sends excerpts between Sinclair’s blogs and suggests alt text for images, but he checks the wording to ensure it does not claim more than a picture shows.

A custom publishing tool is now doing two jobs for David Sinclair: sending excerpts from two of his blogs to a third, and drafting text descriptions for images. The changes, described in a post on Sinclair Trails, revive a cross-posting routine he had handled by hand for years before abandoning it because it took too much effort.

The tool sends material from Sinclair Trails and Dejal to Dejus, with an opening extract, a link to the original post and, where available, a sample image. Dejus is Sinclair’s personal blog, which he describes as a place for informal writing and links to his other sites. Its archive shows the 20 November entry alongside a varied mix of posts, including travel writing, photographs and software updates.

The editor also uses Apple Intelligence to suggest alternative text when an image is added to a post. Sinclair says the descriptions are generally useful, and can be revised or generated again if needed. In one example, the system described a grey cat looking over a laptop as being above a screen displaying code. Sinclair changed the wording to say the screen showed blurred content, removing a detail the image did not establish.

Advertisement · Noah

That correction illustrates why automated image descriptions are best treated as drafts. The W3C’s Web Accessibility Initiative advises that text alternatives should reflect an image’s role on the page: an informative image needs a concise account of its essential message, while a decorative image may not need a description at all. Naming visible objects is not always enough; context matters, and a description should not claim more than the image supports.

Apple’s guidance for developers describes how images can be analysed alongside text prompts using its Foundation Models framework, although it does not identify the technology behind Sinclair’s particular editor. Apple also documents image-recognition features for VoiceOver, which can help users explore visual and screen content. In the blog workflow, the practical benefit is different but related: machine-generated descriptions can help prepare posts for accessibility, while the option to review them leaves the final judgement with the publisher.

Inspired by headline at: [1]

Sources by paragraph: - Paragraph 1: [2] - Paragraph 2: [3], [4] - Paragraph 3: [2] - Paragraph 4: [5] - Paragraph 5: [6], [7]

Source: Noah Wire Services · Original

Make sense of what's next.

A weekly briefing for people shaping the future of publishing.

Free. Email requests are handled by our team.