miguel

The industry is transitioning from Software 1.0 (deterministic code) to Software 2.0 (prompts, skills, context files executed by LLMs). Companies will maintain prompt/skill repositories instead of traditional codebases, and open-source value shifts from code to shared prompts, skills, and training data.

Anthropic (Claude Code) and likely OpenAI are pricing consumer coding products far below API costs (13x subsidy) to drive adoption and lock in developers. This vertically integrated model threatens middle-layer tools like Cursor that must buy tokens at wholesale rates.

AI coding agents have shifted the bottleneck from code generation to code review. Factorial sees 150+ PRs pending review (up from 30) because human review doesn't scale. They are investing in automated review agents with persistent memory to handle 50-70% of reviews within a month.

Factorial is deploying 'Auto Research' (inspired by Karpathy's repo) where agents run hundreds of parallel experiments overnight (performance, structure validation, etc.) and submit PRs with improvements. This turns compute into a continuous optimization loop while engineers sleep.