Boris Cherny on Building Claude Code and Career Growth

How Boris Cherny Thinks About Career and Product Building

This conversation with Boris Cherny (creator of Claude Code, former Meta Principal Engineer) is a masterclass in intentional career building. What stands out isn't just the trajectory from Meta E4 to Anthropic - it's the pattern underneath.

The latent demand principle: Boris's biggest wins at Meta came from identifying what people were already trying to do, then building products to make it easier. Not inventing new behaviors, but removing friction from existing ones. This same principle shows up in his TypeScript work, his state management frameworks, and now Claude Code itself.

The side quest philosophy is particularly relevant for AI engineers right now. Boris broke his arms in a motorcycle accident and literally couldn't write JavaScript anymore - so he learned Haskell and functional programming because they required fewer keystrokes. That constraint led to a complete shift in how he thinks about code (type signatures over implementation). This kind of lateral learning compounds in ways you can't predict.

The Claude Code insights are striking: Anthropic has tripled in headcount, but productivity per engineer has grown 70% because of Claude Code. That's not automation replacing work - that's augmentation creating leverage. Boris's advice: "Don't build for the model of today. Build for the model 6 months from now."

When asked how many engineers it would take to rebuild Facebook Groups migration today (originally 20-30 engineers for 2 years), his estimate: 5 engineers for 6 months. Then he pauses: "If you ask me this question in 3 months or 6 months, my answer will be totally different. In 6 months the answer might be this is actually one engineer."

The generalist theme runs through everything. At Anthropic, product managers code. Data scientists code. User researchers code. This isn't about job titles blurring - it's about reducing the coordination costs of building. When everyone can contribute to the full stack, velocity increases exponentially.

His recommendation of "Functional Programming in Scala" as the most impactful technical book is telling. Not because you'll use Scala, but because it rewires how you think about problems. The specifics matter less than the mental model shift.

8 Insights From Boris Cherny on AI-Era Engineering

• Latent demand is everything - Don't try to create new behaviors; find what people are already doing and make it easier (40% of Facebook Groups posts were commerce, so they built Marketplace)
• Side quests compound - TypeScript book, state management frameworks, meetups - the "cherries on top" build relationships and skills that become core later
• Come in underleveled - Lower expectations give you space to explore and build momentum through exceeding them
• Think in types, not code - Type signatures matter more than implementation; this mental model shift is more valuable than any specific language
• 70% productivity boost from Claude Code - Anthropic tripled headcount while productivity per engineer grew 70% through AI tooling
• Build for 6 months from now - Models are moving so fast that building for today's capabilities is already obsolete
• Generalists win in AI orgs - At Anthropic, PMs code, data scientists code, researchers code - reducing coordination costs is the new leverage
• Facebook Groups migration estimate: 5 engineers, 6 months - Down from 20-30 engineers for 2 years, and will be even less in 6 months

What This Means for Software Engineering Teams

The creator of Claude Code says a project that took 20-30 engineers two years could now be done by 5 engineers in 6 months - and in another 6 months, maybe just one. We're watching the meaning of "engineering team" get redefined in real-time.