Brains Behind AI

Brains Behind AI is an interview show about the companies being built on artificial intelligence and the people building them. Ari Yacobi hosts, and in a number of episodes Natalie sits in alongside him. Guests include founders and CEOs, along with CTOs, a chief scientist and a designer. Read as a run, the episode titles cover unusual ground. There are conversations about deep learning for drug discovery, stroke detection and radiology. Others concern yoga pose estimation, aerial imagery for property insurers, sports data, online disinformation, fabless semiconductors and a company trying to digitize the sense of smell. Public policy for an AI and robotics driven world gets an episode too. Episodes can start cold, with a clip of the guest mid-explanation, before Ari's standard welcome. His questions tend to move from a plain overview of the technology to the hardest problems and finally to advice for would-be founders. That structure keeps the technical talk tied to a business story. Abraham Heifets of Atomwise recalls people looking at the same data and reaching opposite conclusions about whether the company was fundable. He frames drug design as a yes-or-no prediction of whether a molecule sticks to a protein. He also argues that pharma is the last major industry still building and testing every prototype physically. Lawrence Ngo, a practicing radiologist and co-founder of CoRead AI, describes labeling much of his training data himself. While reviewing cases his algorithm had flagged, he found that the original radiologist, not the model, had sometimes made the error. He also contrasts healthcare with app development: an engineer who builds a ride-sharing app can use it to get home, while healthcare adds clinical depth and economics he admits he barely understands. Jared Ficklin of Argo Design leans hardest on metaphor. Narrow models are trained beagles that can drift into sniffing popcorn instead of gunpowder, and a design career starts like a jazz player mastering an instrument before ever playing a gig. Personal data, he suggests, should be handled more like currency than like a right. Ari's enthusiasm is audible throughout, and he is open about his passion for AI in drug discovery. The sign-off points to his monthly AI Startup Tracker at brainsbehind.ai.