Coffee with a Founder Podcast: Andrew Dai, Elorian AI

22- Sep-2026

In this episode of Coffee With a Founder, host Beck Bamberger sits down with Andrew Dai, co-founder and CEO of Elorian AI, to explore the next frontier of artificial intelligence: visual reasoning.

Drawing on his experience developing language models at DeepMind, Andrew explains why today’s AI systems still struggle to understand and reason through visuals the way people do. He shares how board games helped expose this gap, what it takes to build a next-generation multimodal model, and why the accelerating pace of innovation demands more nuanced conversations about AI safety and regulation.

Andrew also discusses the rising standard for technical talent and what founders must do to attract exceptional people in a rapidly evolving industry. Whether you’re a founder, AI enthusiast, or future-focused leader, this conversation reveals why the next breakthrough in artificial intelligence may be something that must be seen—not written.

show notes:

00:00 - Introduction to Andrew Dai, Elorian AI

06:30 - Testing AI on board games

10:29 - Nuanced AI research discussion

12:33 - Discussing technology regulation

17:26 - AI's impact on hiring standards

Read More About The Episode:

00:00 — Introduction to Andrew Dai, Elorian AI

Andrew Dai, co-founder and CEO of Elorian AI, introduces the company’s mission to advance visual reasoning in artificial intelligence. While today’s models have made tremendous progress in language and mathematics, they still struggle with everyday visual tasks. Elorian AI believes vision is essential to helping AI understand, simulate, and eventually interact with the physical world.

06:30 — Testing AI on board games

Andrew explains how board games such as Dune: Imperium, Terraforming Mars, and Heat exposed limitations in current AI models. Although people can understand these games’ visual layouts and spatial relationships, AI often struggles with basic visual comprehension. This realization helped inspire Elorian AI’s multimodal approach.

10:29 — Nuanced AI research discussion

Andrew challenges the polarized conversation surrounding AI research and safety. He argues that public discussions are sometimes shaped by dramatic predictions or individual experiences presented as universal truths. Meaningful debate requires greater specificity about the risks, technologies, and safeguards involved.

12:33 — Discussing technology regulation

As AI advances faster than policymakers can respond, Andrew makes the case for focused regulations that address clearly defined risks instead of placing broad restrictions on the entire industry. He argues that regulation should target specific dangerous capabilities and applications without unnecessarily slowing productive research and innovation.

17:26 — AI’s impact on hiring standards

Andrew discusses how AI is changing what companies expect from technical talent. Elorian AI is looking for people across software engineering, inference infrastructure, and research whose judgment and abilities exceed what AI can already produce. As AI handles more baseline technical work, engineers will increasingly distinguish themselves through deeper expertise, original thinking, and the ability to solve problems existing models cannot.