ai-native-product-refounding

Installation
SKILL.md

AI-Native Product Refounding

In the AI era, product market fit must be constantly "refounded." This framework moves teams away from "blunt instruments" (long roadmaps, rigid PRDs) toward a high-velocity, hands-on approach where the product is shaped by the unique capabilities of evolving models.

Core Principles

  • Vibes before Evals: During the divergent "discovery" phase of an AI feature, prioritize "vibe-checking" (open-ended testing) over rigid evaluation benchmarks. Converge on formal evals only once the core "Aha!" moment is found.
  • The Hybrid Prototyper: PMs, Engineers, and Designers must collapse silos. A PM must be "technical enough to be dangerous" and a designer must understand LLM tool-calling limits to build realistic UX.
  • Greedy Inference: Be "intentionally wasteful" with compute for strategic insights. Spend hundreds of dollars on LLM calls to analyze sales transcripts or user data if it yields one "astute" product insight.

The Refounding Workflow

1. Conduct the "Clean Slate" Audit

Before adding AI to an existing feature, ask: "If I were founding this company/feature from scratch today with current AI capabilities, what would the native experience be?"

  • Identify if your current product is a "Lego kit" (useful primitives) or "Legacy weight."
  • Determine if the AI should be an assistant (sidebar) or the primary agent (the default interface).

2. Bifurcate into Fast and Slow Thinking

Restructure the team into two distinct modes to prevent infrastructure from slowing down innovation:

Related skills

More from samarv/shanon

Installs
4
Repository
samarv/shanon
GitHub Stars
23
First Seen
Feb 9, 2026