Simon Willison built a newsletter page using voice control with Codex
Simon Willison published a Newsletters page that he built through a voice conversation with Codex while working on a local Django project. After half an hour, it was almost finished; reviewing the code, fixing the import and making further changes took approximately half an hour at the keyboard.
Simon Willison published a Newsletters page on his blog listing weekly newsletters from Substack and monthly newsletters for supporters. He built the feature while cooking, through a voice conversation with Codex in the ChatGPT desktop app. He used the model GPT-6 Astra High with the local development environment for the Django project simonwillisonblog. According to his account, the feature was almost ready for deployment after approximately half an hour of voice interaction.
He first asked for the development server to be started and the site opened in a browser so he could check the results as he went. He then used voice input to refine the requirements for a new database model, a migration, page handlers, templates and data imports from external sources. Codex edited the code, responded and occasionally asked follow-up questions; Willison also commented on the page design based on the local preview.
Finishing the feature required work at the keyboard. Willison had Codex create a branch and a pull request, then reviewed the code in the GitHub interface. One import used Git via subprocess; he had it replaced with an API-based import because the data was in a private repository. He also needed to create an API key and adjust how the pages were displayed. Approximately another half an hour of keyboard input produced the version he deployed to production.
Why it matters
The case offers a concrete workflow for developing a simple web feature: combine voice input with a local preview and a subsequent review of the changes. It also demonstrates the limits of this approach: the nearly finished feature still required an import fix, access to private data and further manual review before deployment.
Relevant practical impact
What this means
For individuals
Voice input allowed Willison to make progress on development while cooking. When choosing how to work, however, his experience suggests that a keyboard makes more sense for precise edits and for pasting examples and error messages.
Check the original
Event sources
only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.