In today’s Digest, we cover Netflix expanding its short-form video offering through publisher content deals, YouTube becoming the UK's leading podcast platform, and OpenAI launching the GPT-Live-1 voice model for ChatGPT.

Netflix expands short-form video

Netflix is expanding its short-form video offering through new licensing agreements with major media publishers, including Penske Media, BuzzFeed Studios, Condé Nast, Hearst Magazines and People Inc.. Beginning 3rd August, subscribers in the US, Canada, the UK, Ireland, Australia and New Zealand will gain access to news, lifestyle, celebrity and how-to videos ranging from around two minutes to more than 20 minutes. The content lineup will include programming from well-known brands such as Vanity Fair, Vogue, Rolling Stone, Bon Appétit, People and Variety.

YouTube tops UK Podcast market

YouTube has overtaken Spotify as the most-used podcast platform among UK listeners for the first time on record, according to new data from Edison Research. Among weekly podcast listeners aged 15 and over, 29% now identify YouTube as their primary podcast service, narrowly ahead of Spotify at 28%.

The shift reflects YouTube’s steady growth in podcast consumption, rising from a 19% share in 2023 to 29% in the first quarter of 2026, while Spotify’s share fell from 33% to 28% over the same period. BBC Sounds ranked third with 15%, followed by Apple Podcasts at 10%. Edison attributes YouTube’s momentum to the growing popularity of video podcasts, mirroring a trend that saw the platform become the leading podcast service in the US two years earlier.

OpenAI launches GPT-Live-1 

OpenAI has introduced GPT-Live-1, a new voice model that is rolling out globally across ChatGPT’s web and mobile apps for both free and paid users. The model replaces Advanced Voice Mode as the default voice experience, with the company saying more than 150 million people already interact with ChatGPT using voice each week.

GPT-Live-1 is OpenAI’s first full-duplex voice model, allowing it to listen and speak simultaneously for more natural, phone call-like conversations. The model can provide verbal cues to show it is listening, interrupt when appropriate to correct misunderstandings, and recognise pauses without immediately responding, creating a more human-like interaction. The launch comes as rivals, including Nvidia, ByteDance and Thinking Machines Lab, also develop full-duplex AI voice systems.