Build for Future Model Capability
Summary
When building AI-native products, optimize for the model capabilities that will exist in six months — not the ones that exist today. The implication: accept poor product-market fit early as the cost of being ready when that model ships. This inverts standard product intuition, which says to find PMF as fast as possible. In fast-moving AI, the product that fits today’s model may be obsolete when the next model arrives; the product that’s slightly ahead of today’s capabilities will hit the ground running.
Boris Cherny runs this principle explicitly at Claude Code: design decisions are made with the assumption of capabilities that aren’t quite available yet. When those capabilities land, the product is already positioned to leverage them fully. The alternative — optimizing for current model performance — means constant catch-up as models improve.
How to Apply
- At product design time: Ask “if the model were twice as capable six months from now, what would this feature look like? Design for that, not for what the model can do today.”
- For investor/stakeholder communication: Frame early PMF struggles as intentional positioning, not product failure. “We’re building for the model that ships in Q3” is a different narrative than “we’re struggling to find fit.”
- For roadmap prioritization: Deprioritize work that optimizes around current model limitations (workarounds, fallbacks). Prioritize work that unlocks value when limitations are removed.
- For team calibration: Set a 6-month “target model” assumption and use it as a shared design constraint across the team. Avoids building for the wrong horizon.
The discomfort to accept: your product will feel ahead of its time early. That’s the point. The timing risk is real — “six months out” requires judgment — but the alternative (designing for today) guarantees you’re always one model release behind.
Sources
From: 2026-02-19 Head of Claude Code: What Happens After Coding Is Solved
Key quote: “It’s going to be uncomfortable because your product-market fit won’t be very good for the first six months. But when that model comes out, you’ll hit the ground running.” Attribution: Boris Cherny What this source adds: First-person validation from the creator of one of the fastest-growing AI developer tools. Cherny runs this principle explicitly at Claude Code — it’s not a theoretical suggestion. Links: Original | Archive
Related
- Intentional Understaffing for AI-First Teams — Complementary: understaffing forces AI adoption now; future-model design frames what you’re building toward
- AI Disrupts Strategic PM Skills Most — The strategic PM skills that matter most are exactly those needed to anticipate where AI capabilities are heading