How to use AI to get ahead, from someone who actually does it every day.
Matt Medved sits down with Zeneca to break down how he uses these tools and how anyone else can too. Zeneca introduces Yoshi, the AI agent he treats as a co-founder, and the gamified Stardew Valley-style dashboard he built to beat his own ADHD and stay consistent.
They get into his Dota 2 prediction model built in Claude Code, why domain knowledge is the only durable edge in prediction markets, and the two-Mac stack he runs across GPT-5.5, Hermes and Opus. Zeneca explains the Anthropic crackdown that forced everyone to rethink their setups, his thesis on AI and privacy coins, and why less is more beats ten open terminal windows.
His advice for getting started is simple. Pick one boring task. Hand it to AI. Learn by doing.
In this conversation:
– Going all in on AI and what actually changed
– Building Yoshi and the case for an AI co-founder
– The gamified dashboard that fixed his consistency problem
– A Dota 2 prediction model and the hunt for a real edge
– His current agent stack and the Anthropic OAuth crackdown
– The AI and privacy coin thesis
– How to monetize what you already know in prediction markets
– Dashboards and external memory as workflow hacks
Subscribe to Now Media: https://nowmedia.co/
Michael Blau: https://x.com/Zeneca
Matt Medved: https://x.com/mattmedved
Subscribe and rate the show wherever you listen.
","author_name":"Now Media"}