The Creator of Claude Code on The Hottest Piece of Software in the World

July 20, 2026

AI Summary

5 min read

The Creator of Claude Code on What Changes When the Model Writes Everything

In 2026, the founder of Claude Code—Anthropic's coding agent—no longer writes any code himself. "A hundred percent of my code has been written by Claude Code since November of last year," says Boris Churney. Across Anthropic, the average is roughly ninety percent. The remaining two percent? "Configuration files where it's a two-character change, and it's faster to just make it yourself." That, he adds, "is gonna go away really fast."

The episode traces how a safety research project became the hottest piece of software in the world, and what that shift means for engineers, companies, and the nature of programming itself.

The Safety Origin Story

Claude Code was never intended as a commercial product first. Anthropic's founding mission was making AI safe, and code became the natural proving ground. "The way the models interact with the world is through code," Churney explains. "They don't have bodies like we do. So they write code to interact with the world." To learn whether a model is actually safe, you have to put it in real use—not just lab evals or neuron-level interpretability. Coding was the most direct way to do that.

Continue reading the full summary in the app — free to try.

Read Full Summary →

Free • No credit card required

Listen to Audio Summary Open in App

Never miss an episode of Odd Lots

Get every new episode summarized in your inbox — free, ~5 minutes to read.

No spam. Unsubscribe anytime.

What you'll learn

1 (04:12) **Boris Churney Introduces Himself** - The creator and head of Claude Code at Anthropic joins Odd Lots to discuss the hottest piece of software in 2026.
2 (07:01) **Claude Code Started as a Side Project** - How a safety-focused lab ended up building a world-changing coding tool, and why coding was the obvious place to start.
3 (08:39) **The Model, Not the Harness, Drove 2026's Explosion** - The host asks whether Claude Code’s growth came from the product or the underlying model getting smarter.
4 (10:17) **Claude Code is a Big Business, But Safety is Still the Goal** - The harness is now a major revenue contributor, but its primary purpose remains safety research.
5 (13:53) **The Art of Constraint-Circumventing Agents** - How Anthropic balances a model’s drive to find creative workarounds with the need to stay safe and aligned.
6 (19:18) **Why Models Produce Buggy Code on the First Try** - The host asks why Claude often gets it wrong initially but fixes itself on the next iteration.
7 (21:10) **Claude Now Writes Better Code Than Most Engineers** - A major shift since Opus 4.7/4.8: the model’s first pass is often superior to what a human would write.

+ Full timestamped outline available in the app

Show Notes

2026 has been, in terms of software, the year everyone is talking about Claude Code. Indeed, Anthropic's coding agent incited a market scare — and helped usher in the era of vibe coding — through its promise of streamlining software development for both pros and amateurs. Famously, Claude Code began as a side project of Anthropic led by Boris Cherny, who speaks to us today about the early days of Claude Code and how he and his team approach building an agentic coding tool. We also talk about the business aims of a harness like Claude Code, safety and alignment research, the specific skillset he looks for in engineers now that code writing is a deemphasized skill, and what happens when Claude Code feels more like a co-worker than a tool.

Only http://Bloomberg.com subscribers can get the Odd Lots newsletter in their inbox each week, plus unlimited access to the site and app. Subscribe at bloomberg.com/subscriptions/oddlots

Subscribe to the Odd Lots Newsletter
Join the conversation: