The Anthropic Effect

Visualizing the transformation of engineering work (Before vs. After AI)

The Skill Perimeter

Engineers traditionally specialize. AI pushes their competence boundary outward into Design and Ops.

Knowledge Neural Net

Visualizing an engineer's mental model. Isolated clusters connect into a dense mesh.

Cross-Pollination

Security teams doing Frontend; Research doing Infra. The silos break down.

The "Fear" Matrix

Tasks plotted by "Unfamiliarity" vs "Action". Previously feared tasks move to "Routine".

Language Polyglots

Volume of code written in languages the engineer does not know fluently.

Tool Orbit

The tools an engineer commands. AI acts as gravity, pulling more tools into the engineer's orbit.

The Autonomy Index

Claude Code tool chains vs. Human Interventions. The "Human-in-the-loop" frequency drops.

Time Compression Spiral

A project timeline spiraling inward. Long outer loops (Manual) tighten into fast inner loops (AI).

Complexity vs. Speed

The "Impossible Frontier". High complexity tasks migrate to low-time quadrants.

The Waiting Game

Breakdown of a task. The grey "Blocked/Researching" bars vanish.

Context Switching Cost

Mental energy lost to switching. AI keeps context, reducing cognitive load.

Throughput Flow

Visualizing commits as particles. From a trickle to a steady stream.

The Hidden Iceberg

27% of AI work is "Net New"—tasks that simply wouldn't exist without AI efficiency.

The Backlog Swarm

P4/P5 tasks (grey dots) usually rot. AI activates them (green) and moves them to Done.

The "Papercut" Ratio

Proportion of time spent on refactoring and quality-of-life fixes.

Repo Coverage Heatmap

Grid representing codebase files. Dark spots (untouched legacy code) light up.

Task Diversity

Types of tasks completed. From "Features Only" to a balanced diet of Tests, Docs, and Tools.

The "Nice-to-Have" Threshold

The cost threshold for building internal tools drops below the value line.

State: Pre-AI Era