https://mastra.ai

In today’s episode, we sit down with the co-founder of Mastra — the YC Winter 2025 startup that gave away 80,000 physical books to jump-start adoption and build community.

We dive deep into:
• Why they chose the bold move of distributing 80,000 books instead of just digital hype.
• How Mastra is re-architecting how AI agents are built and deployed — beyond simple chatbots.
• The founder’s origin story: from open-source and web frameworks to the bleeding edge of agent technology.
• The tools, the hacks and the orphaned prompts most devs don’t know but Mastra is solving.
• What comes next: how this startup plans to turn developer momentum into an AI infrastructure business.

In this episode, we sit down with the co-founder of Koyal, a groundbreaking YC F25 startup that’s reimagining how we create video. Koyal transforms ordinary audio into dynamic, lifelike video using cutting-edge AI. We talk about the moment that sparked the idea, the challenges of teaching machines to understand voice and emotion, and what it means for the future of content creation. Whether you’re a filmmaker, podcaster, or just curious about where AI storytelling is headed, this conversation offers a fascinating glimpse into the technology that could make traditional cameras a thing of the past.

https://koyal.ai

We talk with Kenneth Wu, co-founder of Mundo AI (YC W25) — the company on a mission to make AI truly multilingual.

Most large language models are trained mainly on English data, leaving billions of people underrepresented. Mundo AI is changing that by building high-quality multilingual datasets that help AI models understand and perform across the world’s languages — from Hindi and Arabic to Swahili and beyond.

Kenneth shares how his team works directly with native speakers, builds scalable data-collection pipelines, and ensures cultural and linguistic accuracy that machine-translation shortcuts can’t achieve. We also dive into what it takes to make AI global, why it matters for the next generation of models, and the opportunities ahead for multilingual AI.