Quoting Andrej Karpathy

2026-06-09 · Source: Simon Willison's Weblog · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Fundamental Awareness, quick

Summary

On June 9, 2026, Andrej Karpathy observed a significant shift in software development, noting that "working software increasingly comes out on a tap," likely referencing capabilities like Claude Fable 5. This ease of generation is triggering a "Jevon's paradox," where the accessibility of software substantially increases demand. Karpathy highlights the expansive possibilities, including the rapid creation of explainers, visualizers, dashboards, and bespoke single-use applications, such as hyper-specific Weights & Biases instances for individual projects. He also points to the potential for 10X improvements in test suites, automated code optimization, and the generation of custom HTML for research project results, suggesting a future where almost any software need can be met on demand.

Key takeaway

For AI Engineers and Software Engineers evaluating development workflows, recognize that AI-driven software generation, exemplified by tools like Claude Fable 5, will fundamentally alter your project scope. You should anticipate a surge in demand for bespoke tools and capabilities, enabling you to rapidly create custom explainers, optimize code, or expand test suites by 10X. Embrace this shift to on-demand software creation to significantly accelerate your development cycles and explore previously unfeasible project dimensions.

Key insights

The ease of AI-driven software generation will dramatically increase demand and expand development possibilities.

Principles

• AI-generated software enables Jevon's paradox in demand.
• Bespoke applications become readily available.
• Software creation shifts to on-demand generation.

In practice

• Generate hyper-specific Weights & Biases dashboards.
• Auto-optimize code and 10X test suites.
• Create custom HTML for research results.

Topics

• AI Software Generation
• Jevon's Paradox
• Claude Fable 5
• Code Optimization
• Test Automation
• Bespoke Applications

Best for: Machine Learning Engineer, Research Scientist, AI Engineer, Software Engineer, AI Scientist

Related on AIssential

• Former GitHub CEO raises record $60M dev tool seed round at $300M valuation — Entire's platform unifies AI-generated code and context, enabling better human-AI collaboration in software development.
• If AI was actually killing software engineering, why is there more code than ever? — AI shifts software engineering bottlenecks from code generation to validation, architecture, and complex system judgment.
• llm 0.32a3 — Claude Fable 5 largely authored the "llm 0.32a3" tool, showcasing AI's role in software development.
• Beyond Code Review — Specification-Driven Development shifts focus from code review to continuous verification against evolving specifications.
• Why demand for code is infinite: How AI creates more developer jobs — AI is a platform shift that will create explosive demand for human-developed code and new developer roles.
• How Braintrust turns customer requests into code with Codex — High-speed AI code generation fundamentally transforms development workflows, enabling real-time customer iteration and autonomous problem-solving.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Simon Willison's Weblog.