Notion’s Token Town: 5 Rebuilds, 100+ Tools, MCP vs CLIs and the Software Factory Future — Simon Last & Sarah Sachs of Notion

Simon Last and Sarah Sachs from Notion discuss their extensive experimentation with AI-native product development, including rebuilding core Notion features five times as capabilities evolved and integrating over 100 AI tools into their workflows. The conversation covers the architectural tradeoffs between Model Context Protocol (MCP) servers and CLI-based tooling for AI agent integration, drawing on Notion's hands-on experience deploying these systems at scale. They explore the broader vision of a "software factory" future where AI agents continuously generate, test, and iterate on software, and what that means for product development teams and tooling infrastructure.

Build an MCP server with Nuxt
tldr: everyone is converging on the same product shape: a general harness that takes a goal, uses tools, and does knowledge work. once every…
great question! knowledge (system prompt, skills) and tools (apis, clis, whatnot) will always be needed, even as the best way to build agent…
There's a speed and focus tradeoff when building a core product and just using a frontier model/harness, but this is a pretty compelling arg…
Learn how to use coding agents in 30 minutes! This course teaches you how to build software with agents: plan new features, fix bugs, review…

Source: Latent Space | 2026-04-15

Related