Automating the boring

Tonnes about AI taking jobs and robbing humankind of creativity. I agree. But I’ve said – as far back as 2022, on stage, in front of my industry peers at the (then) Radio Academy Festival – that I like making the stuff and hate the admin around uploading/distributing it.

I’ve been building a system where Whisper transcribes audio locally, per-show voice models get built from what’s already been uploaded to the RSS feed, and one drafting pass produces the title, description, on-air intro, newsletter line and socials – plus attaches the audio to an email, ready to send to anyone who needs to hear it before it goes live.

Automating the typing is fine.

Automating the content is not.