I tracked every trending AI repo's stars daily for 3 weeks. The growth is not where I expected

Short summary

Author analyzed 3 weeks of star-count data from 611 trending AI repos (2,671 snapshots). Findings: growth has shifted from agent frameworks to specialized domain tools and skill packs, and old single-purpose tools are seeing renewed adoption as agents learn to invoke them. The competitive moat has moved from orchestration code to encoded judgment.

• •Market growth concentrated in domain-specific tools and skill packs, not frameworks
• •Decade-old single-purpose tools seeing explosive adoption as agent companions
• •Competitive advantage now lies in encoded expertise, not orchestration code

Generated with AI, which can make mistakes.