What if the smartest way to use AI right now is knowing exactly where it stops being enough?

In this episode of Lumos Talks, Brogan Neagle — AI Technical Product Manager at Dell and co-founder of identity verification startup Ethos Social — has built his career around a simple but urgent idea: AI can get you eighty percent of the way there, but the last twenty percent still needs a human. That instinct traces back further than you’d think — to filing his first patent at fourteen after solving a problem on the lacrosse field, and to building an early chip designed to detect deepfakes in court years before most people were paying attention to AI-generated fraud.

He gets into why so many AI ”wrapper” startups disappear the moment a bigger model catches up, why trust — not the model itself — is the only real moat left, and why the world is racing toward a future where agentic AI could outnumber human interactions a thousand to one.

He also breaks down exactly where people over-trust AI today, why judgment and human skill still matter more than most people admit, and why a LinkedIn reply or a real conversation at an event still beats anything automated.

Grounded, clear-eyed, and a genuinely useful way to think about where AI actually earns your trust — and where it doesn’t.