Ep2 - Why most factories stay 'grey': scaling vision-AI automation (Jonathan Berte, Robovision)

What happens when you give machines eyes? In Episode 2 of the Azumuta podcast, Robovision founder Jonathan Berte takes us from the research lab to the real factory floor to explain how computer vision and AI are reshaping the future of manufacturing.Machine vision is what lets robots adapt to product variation, handle edge cases, and move beyond high-volume, low-mix production. Jonathan has spent almost twenty years building it, so he knows exactly where the hype ends and reality begins.You'll learn how deep learning let shop-floor experts build their own AI vision without code, why data capture has to come before the AI works (and what the Tesla bet teaches manufacturers), and the manufacturing "data catch-22." Jonathan also explains why 2026 could be the breakthrough year for humanoid robots, driven by cost rather than capability, why the cobot revolution stalled, and how human-in-the-loop supervision really works in production. Plus Nvidia vs. China, a possible 2026 compute crisis, and the "Ralph Wiggum Loop" that made one AI task 500x cheaper.The throughline: machines automate, but people stay in control. Less programming, more prompting.Built for manufacturing leaders who want signal, not another humanoid demo video.🔗 Learn more: https://www.azumuta.com/?utm_source=buzzsprout&utm_medium=podcast_description&utm_campaign=dark_factories_ep3&utm_content=homepage📅 Book a demo: https://www.azumuta.com/demo?utm_source=buzzsprout&utm_medium=podcast_description&utm_campaign=dark_factories_ep3&utm_content=book_demoIf you enjoyed this episode, follow the show and leave a review.