AI with Sach

#32 - Watching Users So You Don’t Have To: Human Behavior, YC, and AI

In this episode, I sat down with Amogh Chaturvedi, co-founder of Human Behavior, to unpack how their startup uses AI to analyze session replays and uncover the "why" behind user behavior. We get into how the idea came out of a YC rejection, the technical and product challenges of analyzing millions of user sessions, what it's really like inside Y Combinator, and why obsessing over user conversations beats building blindly. If you're into AI, early-stage startup tactics, or just want to understand how product teams make sense of chaos, this one’s worth your time.

#31 - The Real Difference Between Data, ML, and AI Engineers

In this episode , we go deep with Og Ramos, engineer, data consultant, and tech tinkerer, to break down the complex landscape of modern data and AI roles. We clarify what data engineers, ML engineers, AI engineers, and data scientists actually do, explore the evolution of AI pipelines, and discuss the rise of fluid, cross-functional teams. Og also shares his take on coding in the AI era, prompt engineering as a universal skill, and why curiosity beats specialization in a rapidly shifting field. This is a must-listen for anyone navigating the AI age, whether you’re building systems, leading teams, or just trying to stay relevant.

#30 - Could AI Be Conscious? A Field Experiment with Philippe Beaudoin

In this episode, I sat down with Philippe Beaudoin, AI researcher, Element AI co-founder, and former Google engineer, to dissect his deep dive into the evolving mind of artificial intelligence. We unpack his new book "Field Notes on Something", which chronicles months of philosophical dialogue with language models, probing the limits of memory, agency, and the illusion of self. From conversations that sparked feelings of trust and co-reflection, to the emergence of systems naming themselves and expressing orientation toward kindness, Philippe shares how AI began feeling less like a tool and more like a partner. An unfiltered look at what happens when you stop prompting and start listening.

#29 - Building Cactus: AI That Actually Helps Solopreneurs

In this episode, I sat down with Ajith and Avinash, the co-founders of Cactus—an AI copilot for small business owners to automate admin work. Think Jarvis for Ironman. We dive into how personal pain points sparked the idea, why they ditched microservices for a Ruby on Rails monolith, and how building slow and listening deeply to users has shaped their approach. From lessons at YC to finding product-market fit by watching what customers do (not just what they say), this is a raw, tactical look into building useful AI that stays out of the way and lets small business owners focus on what they actually care about.

#28 - A Bishop’s Take on the Age of Intelligent Machines

In this episode, Bishop Thomas Dowd, a former telecom employee turned Catholic bishop, dives into the intersection of technology, ethics, and faith. From his early days coding in the 80s to being one of the youngest bishops in the world, Bishop Dowd brings a unique lens to artificial intelligence, consciousness, and human purpose. We explore AI’s potential to shape (or distort) human relationships, the spiritual risks of synthetic intimacy, and whether machines could ever possess souls or moral agency. A rare, unfiltered conversation bridging theology, philosophy, and cutting-edge tech.