The Man Who Says AI Code Will Break Everything - EP 59 Will Wilson
Will Wilson spent five and a half years in stealth building Antithesis, a company with a radical premise: the way the software industry has been testing code for the past 75 years is fundamentally broken, and almost nobody is talking about it. Wilson, a Yale-trained mathematician who previously co-founded FoundationDB — the database Apple acquired and still relies on — has built a fully deterministic simulation environment that lets engineers do something that should be impossible: rewind time on a software crash, replay it from any angle, and fix it the way a detective would if they had a time machine at a crime scene. It sounds like science fiction. Wilson is convinced it's the only approach that actually works. The timing turned out to be lucky in ways he didn't plan for. As AI coding tools exploded into the mainstream, Wilson found himself with a new sales pitch practically writing itself: if you don't really know what the AI wrote, how do you know if it works? In this conversation, Wilson gives one of the more clear-eyed takes on the state of AI-generated code — where it genuinely helps, where it quietly fails, and why he thinks the software reliability crisis is more dangerous than most people in the industry want to admit. He's not a hater. He's just someone who has spent a long time thinking about what happens when the planes stop flying and nobody knows who wrote the code. Our host Ashlee Vance is a journalist, author, and filmmaker best known for his New York Times bestselling books, including the biography of Elon Musk and When The Heavens Went on Sale. He spent years at Bloomberg Businessweek covering Silicon Valley, space, and the people building the future, where he also created and hosted the Emmy-nominated series Hello World. His documentary work includes Wild Wild Space for HBO and Don't Die for Netflix. Now he runs Core Memory, a media company telling stories about scientists, inventors, and startups changing the world—from biotech labs to factory floors to the edge of space. Subscribe to watch more Core Memory videos / @corememoryvideos Check out our podcasts / @corememorypodcast https://www.corememory.com/podcast Find Ashlee Vance at www.corememory.com / ashlee.vance https://x.com/ashleevance Find Kylie Robison at https://x.com/kyliebytes / kylie.robison Chapters 00:00 Intro Debugging Software vs. Crime Scene Investigation 01:31 Introducing Will Wilson, CEO of Antithesis 06:51 Why Big Tech Companies Still Use Primitive Testing 10:50 Determinism: Why Software Behaves Differently on Different Hardware 14:28 How to Build a Fully Deterministic Simulation 20:25 Debugging with a Time Machine 24:09 Why Antithesis Left Silicon Valley for Virginia 31:18 AI Coding: Superhuman Speed vs. Poor Taste 35:07 The Danger of Technical Debt and AI Slop 40:09 Training AI to Write Bugs Humans Can't Spot 47:44 Making a Bet Against the Extreme AI Future 55:54 Why Software Testing is a Great Opportunity

Is the AI Boom About to COLLAPSE?

Why America Is Losing The Robot War To China

Zig 2026: No-AI Policy, $670K Foundation, Left GitHub & Why Zig Isn’t 1.0 - Andrew Kelley Explains

Inside Anthropic, the $965 Billion AI Juggernaut | The Circuit

Testing Distributed Systems the right way ft. Will Wilson

Andrej Karpathy: Software Is Changing (Again)

Unfortunately, I Was Right

Andrej Karpathy: From Vibe Coding to Agentic Engineering w/ Stephanie Zhan

Anthopic, OpenAI Should Not Be Allowed to IPO, Says Ed Zitron

Building OpenCode with Dax Raad

The AI Breakthrough That Will Change Everything (Google DeepMind CEO Interview)

Yann LeCun Says LLMs Have 2 Years Left…

The Factory Of Youth - EP 60 Nabiha Saklayen

AI Billionaires Want to Control EVERY Aspect of Your Life | Aaron Bastani Meets Karen Hao

Ex-Google Officer: You Only Have 3 Years Left Before It Hits! - Mo Gawdat

Kubernetes and retiring at the top with Kelsey Hightower

The Uncomfortable Truth About AI “Reasoning” | World Science Festival

Anthropic's Boris Cherny: Why Coding Is Solved, and What Comes Next

Emerging Situation: Anthropic's Global Pause, Recursive Self-Improvement, and AI Personhood Arrives

