Practicing Multimodal AI

Slater Victoroff, CTO and founder of Indico, the enterprise AI startup, interviewed by host Trond Arne Undheim, futurist, investor, and author.

In this conversation, we talk about how Slater was picking trash off of the Wellesley dump for school engineering projects, loves Chinese fantasy fiction Xianxia- with its immortal heroes, his experience at Techstars, founding a startup, and how startups beat juggernauts like IBM spending billions of dollars. We discuss how his company Indico practices multimodal AI, a set of blended techniques that make data sing. We muse about the future one where citizen data scientists contribute to better problem framing driven by subject matter experts. 

Having listened to this episode, check out Indico’s well as Slater Victoroff’s online profile:

Trond’s takeaway: The secret to making money with today’s AI techniques seems to lie in blending various approaches, being able to handle a myriad of data sources, and meshing it together without losing the context and stumbling along making predictions that make sense even though the underlying dimensions are seldom fully understood, using transfer learning approaches. I would personally hope we could get a few steps further soon, so the explainability also increased. We will get there soon enough, I guess. Let’s see if the technology is weatherproof and whether we can get there without another AI winter. I find it refreshing to talk with smart people who are also humble. That’s why my bet will be on folks like Slater to build these systems for the future. 

Thanks for listening. If you liked the show, subscribe at or in your preferred podcast player, and rate us with five stars. Check out @futurized2 on Instagram, @futurized2 on Twitter, or our YouTube channel at

If you like this topic, you may enjoy other episodes of Futurized, such as episode 74, AI Talent Diversity, episode 79 Futuristic AI, or episode 48, The Future of AI in Government.

Futurized—preparing YOU to deal with disruption.