AI can't replace your editors, but it can do 80% of their most tedious, repetitive work in a fraction of the time. Welcome to The Media Odyssey Podcast!

In this episode, Evan and Marion open with the news that OpenAI is shutting down Sora's B2C offering, unpacking what it signals for the AI and media landscape. They welcome Olivier Reynaud, Co-founder and CEO at Aive, a platform Aive built around the central challenge facing every creator and media company today: how do you produce enough high-quality, platform-tailored video content to keep pace with the demands of social video without burning out your team or blowing your budget? Olivier draws on his background co-founding Teads, where the team broadcast billions of videos daily, to explain how the bottleneck was never video creation itself, but large-scale personalization. The conversation explores how Aive is solving that problem through proprietary meta-learning technology, and what that means for the future of creative work.

Key Takeaways:

1. OpenAI Shutting Down Sora Signals a B2C Dead End

OpenAI announced it is closing down Sora and stepping back from its deal with Disney to refocus on enterprise. The hosts argue that selling AI tools directly to consumers was never a sound business model. As Evan puts it, AI is best understood as "an arrow in your quiver, not the bow."

2. The Real Problem Isn't Making Video, It's Personalizing It at Scale

Olivier, who co-founded Teads and has spent 20 years in video, argues the hard problem isn't producing video content; it's tailoring that content for every platform and audience at meaningful scale. Aive is built specifically to solve this: taking a long-form master and generating hundreds of format-adapted clips in days rather than months.

Aive Eliminates ~80% of Repetitive Production Tasks

3. Using Match Group's Meetic as a case study

Olivier explains that Aive helped produce nearly 300 campaign variants across a full quarter, cutting production costs by roughly 80%, reducing time-to-market from two months to days, and delivering a 50% performance uplift on paid Facebook and Instagram campaigns. The savings were reinvested into more content, bigger media buys, and team training.

4. The Technology Is Proprietary and Built for Enterprise Security

Aive runs on in-house meta-learning and is not trained on OpenAI, Google, or Amazon models. For clients sending unreleased films or large campaign assets, data stays within the platform and never trains outside systems. The platform is SOC 2 certified and currently designed for enterprise and mid-size agencies, not individual creators at a consumer price point.

5. Netflix's 1.5 Million Trailer Versions Prove Human Editors Can't Keep Up Alone

Evan opens with a striking data point: for the final season of Stranger Things, Netflix created 1.5 million different versions of their trailer for YouTube and social — a volume impossible to achieve through human editing alone. This frames the episode's central question: how do media companies and creators scale social video output to the level the market now demands, without AI doing the heavy lifting?

Book a meeting with Aive: https://0icjqan647l.typeform.com/AivexTMOpodcast

Use Aive’s exclusive code (NS015504) for a FREE Show Floor Pass and join them at NAB, the event redefining the future of media and entertainment: https://invt.io/1exbuo45cd4

Thank you to Olivier Reynaud for joining the pod!

Olivier Reynaud - https://www.linkedin.com/in/olivierreynaud/ 

Aive - https://www.linkedin.com/company/aive/ 

Interested in sponsorship? https://forms.gle/2LCWfX2HBNT8mtpx8

Connect with us on Linkedin:

Evan Shapiro - https://www.linkedin.com/in/eshap-media-cartographer/

Marion Ranchet - https://www.linkedin.com/in/marionranchet/

The Media Odyssey Podcast - https://www.linkedin.com/company/the-media-odyssey-podcast

• (00:00) - Netflix Trailer Explosion