Emerging developer patterns for the AI era

Yoko Li at a16z discusses how AI is changing software development, moving beyond tooling to a new foundation. This involves rethinking core concepts.

Nine emerging patterns are highlighted:

AI-native Git: Version control shifts focus to prompt+test bundles over line changes.
Dashboards -> Synthesis: Interfaces become dynamic and conversational via LLMs.
Docs as interactive knowledge bases: Documentation serves humans and agents for context.
Templates to generation: Custom project scaffolding based on descriptions replaces static templates.
Beyond .env: Secrets management moves towards scoped tokens and local brokers for agents.
Accessibility as the universal interface: Accessibility APIs may become the primary layer for agent interaction.
Asynchronous agent work: Developers orchestrate agents working in the background across the development lifecycle.
MCP becoming a universal standard: Model Context Protocol enables tools to expose standard interfaces for agents.
Abstracted primitives: Agents use robust, composable services (auth, billing, storage) optimised for agent consumption.

The article states:

It’s not just AI layered onto old workflows, but is a redefinition of how software gets built with agents, context, and intent at the core.

Read the full article for details.