Boris Cherny

Software engineer at Anthropic and creator of Claude Code

Claims

As speaker

Mentioned in

Key positions and views

The act of manually writing code is obsolete; the future of software engineering is orchestrating thousands of AI agents to perform development tasks.

A company's true competitive advantage in the AI era stems from its AI-native organizational structure and processes, not just its proprietary technology.

AI will erode traditional business moats like switching costs and process power, while leaving moats like network effects and economies of scale largely intact.

The future of product teams involves cross-disciplinary generalists where every member, regardless of title, uses AI to contribute to the codebase.

The skill of building software will be universally democratized by AI, becoming as common as basic literacy skills like sending a text message.

Podcast consensus on Cherny

Points of consensus

▶Boris Cherny is a central figure in the creation and development of Anthropic's Claude Code, having joined the incubator team in late 2024 and shaping its strategy.May 2026

▶Cherny advocates for and practices a software development methodology where AI agents write 100% of the code, rendering manual coding obsolete. He personally manages thousands of agents to perform complex development tasks.May 2026

▶The success and exponential user growth of Claude Code are directly and repeatedly attributed to the performance improvements in successive AI model releases, particularly the Opus series starting with Opus 4.May 2026

▶Anthropic's internal culture, particularly on the Claude Code team, is 'AI-native,' where all members, regardless of their formal role (e.g., finance, design), are expected to write code using AI tools.May 2026

Points of debate

▶There is a tension between the claim that Anthropic's core technology is 'publicly available' and the simultaneous mention of internal-only models like 'Mythos' for testing, suggesting a potential gap between internal capabilities and public offerings.

▶Cherny's assertion that improving multi-agent systems is 'primarily achieved through prompt engineering' contrasts with his prediction that models will soon take over the orchestration of complex workflows themselves, questioning the long-term importance of user-led prompt engineering.May 2026

▶The initial lack of success for Claude Code is presented both as a failure (Cherny only used it for 10% of his own tasks) and as a deliberate strategic choice to build for a future, more capable model, creating ambiguity about whether it was a misstep or a calculated risk.May 2026

▶A conflict exists between the current necessity for product-level safety mechanisms (like human-in-the-loop approvals) and Cherny's prediction that model alignment will improve so significantly within a year that such features will become less necessary.May 2026

Key themes

▶The Obsolescence of Traditional CodingMay 2026

Cherny argues that manual software development is being replaced by the orchestration of AI agents. He exemplifies this by stating he hasn't written code in 2026 and that Anthropic as a company no longer writes code manually, instead relying on AI models to generate 100% of it.

This signals a fundamental shift in the software development labor market, where value will accrue to those who can effectively manage and prompt AI systems at scale, rather than those who write code from scratch.

▶AI-Native Organizational Structure as a MoatMay 2026

Cherny posits that Anthropic's primary competitive advantage is not its technology, which he claims is publicly available, but its advanced organizational structure and processes built around AI. This is demonstrated by the Claude Code team, where every member, from finance to design, writes code using AI.

Investors should recognize that true defensibility in the AI era may come less from proprietary models and more from a company's ability to fundamentally restructure its internal workflows and culture to be AI-first.

▶Building for the Future ModelMay 2026

The development of Claude Code followed a counterintuitive strategy of building a product for the capabilities of a future, unreleased AI model. This involved knowingly accepting a lack of product-market fit for six months until the release of the more powerful Opus 4 model, which then catalyzed exponential growth.

This 'build-ahead' approach represents a new, high-risk/high-reward paradigm for product development in the AI space, requiring deep conviction in a company's model roadmap.

▶The Shifting Landscape of Business MoatsMay 2026

Cherny provides a nuanced view on how AI will impact business strategy, predicting the erosion of moats like 'switching costs' and 'process power' as AI can easily port data and optimize workflows. However, he believes foundational moats like network effects and scale economies will remain durable.

Companies relying on proprietary processes or high customer lock-in are vulnerable to disruption, whereas those with strong network effects are better positioned to withstand the competitive pressures introduced by AI.

Source episodes

Sentiment over time

Not enough data for timeline

Changes over time

Late 2024

Cherny joins an incubator team within Anthropic Labs to create Claude Code, at a time when state-of-the-art AI coding was single-line 'type ahead' completion.

Late 2024 - Mid 2025

The Claude Code team deliberately builds the product to target the capabilities of a future AI model, accepting that it would lack product-market fit for approximately six months.

May 2025

The release of the Opus 4 model serves as a major inflection point, triggering the start of exponential user growth for Claude Code.

October/November 2025

The Claude Code team reaches a milestone where the AI model writes 100% of the product's code, marking a shift in their development process.

2026 (to date of interview)

Cherny reports that he has not written a single line of code in the year, fully transitioning his role to orchestrating AI agents.

Post-launch

The original innovation team that built Claude Code is disbanded and later reassembled for a 'round two' led by Anthropic's Chief Product Officer, Mike Krieger.

Suggested prompts

How sustainable is Anthropic's 'organizational process' moat if competitors can observe and replicate their AI-native workflows? &nearr;What are the second-order effects of democratizing software development to the level of basic literacy, particularly on the economic value of specialized engineering talent? &nearr;If AI alignment improves as rapidly as Cherny predicts, what new product categories and agentic behaviors become feasible when safety guardrails can be relaxed? &nearr;Analyze the risks and rewards of the 'build for the next model' product strategy. Under what conditions is this approach most likely to succeed or fail? &nearr;

Key concepts

Claude Code 22 ep Anthropic's Models (Opus/Sonnet/Mythos) 12 ep AI Agents & Automation 11 ep Anthropic's Strategy & Culture 8 ep Future of Software Development 6 ep Business Strategy & Moats 5 ep Product Development Strategy 3 ep AI Safety & Alignment 1 ep

Notable quotes

“I think actually the place that we're ahead is not the technology, because the same technology available to us is available to everyone here... But I think there's actually a far bigger lead in kind of the organizational structure and organizational process.”

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

“we got to the point where the model just wrote 100% of the code. And for us, this happened sometime in October, November last year.”

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

“everyone on our team code. So like our engineering manager, our product manager, our designers, our data scientists, our finance guy, our user researcher, every single person on our team writes code.”

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

“we were trying to build this thing that was like pre-PMF. And we knew that it wouldn't have PMF for six months because we were building for the next model.”

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

Report last updated: May 10, 2026

Get started free

Back to Entities Intelligence Report