bitter-lesson-scorecard

What problem does it solve?

Evaluates agent architectures against the Bitter Lesson to reveal how much procedural "how" knowledge is encoded relative to the desired "what" outcomes, flagging hard-coding, rigid orchestration, and domain hacks, and it generates a practical simplification roadmap.

Core Features & Use Cases

• Alignment scoring: computes an Overall Alignment score from outcome specs, tool interfaces, and hard constraints.
• Lock-In risk assessment: measures reliance on procedural orchestration, model compensation, and domain hacks to highlight areas that resist model improvements.
• Roadmap recommendations: outputs concrete steps to replace procedural constructs with outcome-driven specs and scalable capabilities.
• Use Case: designed for evaluating new agent designs, architecture reviews, or system prompts to plan simplifications ahead of deployment.

Quick Start

Provide a system description or architecture doc to generate a Bitter Lesson scorecard and a targeted simplification plan.