From Coding to Autonomous Proposals

Running a team over the past year, I noticed a clear pattern: people who thrive in AI-augmented work share one quality — genuine curiosity that spills outside their job description. Those who struggle often can’t work without clear instructions. That gap, I’ve come to believe, is exactly how future talent will sort itself out.

---

Boris Cherny, the creator of Claude Code at Anthropic, recently gave interviews on Lenny’s Podcast and Y Combinator’s Lightcone that I can’t stop thinking about. His take on talent in the AI era is genuinely radical.

https://www.youtube.com/watch?v=Mr2eVO052bQ

1. Look for People with Side Quests

Boris pays close attention to what candidates do outside work. His favorite example: an engineer who brews kombucha at home. The brewing itself doesn’t matter — what it signals is self-directed curiosity and the drive to solve problems nobody assigned you.

In an era where technical barriers are dropping fast, that intrinsic drive is the real differentiator.

2. From “Engineer” to “Builder”

Boris predicts the title “software engineer” may disappear as early as 2026, replaced by “Builder” — someone who synthesizes product, design, and development fluidly. At Anthropic, PMs, designers, and even finance people write code. Job boundaries are becoming fluid. The 50%+ overlap between roles isn’t a problem — it’s the goal.

3. Generalist > Narrow Specialist

When AI handles 100% of code writing, the value shifts to:

• Defining the right problem
• Validating outputs (Is this correct? Is it safe?)
• Maximizing the AI — one great generalist with unlimited tokens outproduces a traditional team

4. Build for the Model Six Months From Now

Don’t learn skills based on today’s AI capabilities. Ask: will this be automated in six months? If yes, its value is already zero. Design your trajectory around where the models are heading, not where they are.

5. Protect Your Human Moat

Boris is direct: almost everything done at a computer will be disrupted. The lasting advantages are things AI can’t replicate — trust-building, ethical judgment, asking the right questions. AI is excellent at answering. Humans need to stay excellent at asking.

💬 Read more: 2025 Year in Review (English)