Notion's most significant AI breakthrough stemmed from a surprising source: simplification. Initially, software engineers at Notion AI experimented with advanced code generation, complex schemas, and extensive instruction sets when exploring large language models (LLMs) and agentic AI. However, through trial and error, the team discovered that eliminating complicated data modeling dramatically improved model performance.
Ryan Nystrom, Notion AI's engineering lead, and his team pivoted to simple prompts, human-readable representations, minimal abstraction, and familiar markdown formats. This re-wired approach led to the release of V3 of Notion's productivity software in September, featuring customizable AI agents. These agents have quickly become Notion's most successful AI tool to date.
Nystrom described the improvement as a "step function," based on usage patterns compared to previous versions. "It's that feeling of when the product is being pulled out of you rather than you trying to push," Nystrom explained on a VB Beyond the Pilot podcast. "We knew from that moment, really early on, that we had something. Now it's, 'How could I ever use Notion without this feature?'"
The initial approach involved complex AI techniques, reflecting the common assumption that sophisticated problems require sophisticated solutions. However, the team found that LLMs responded better to simpler, more direct instructions. This shift highlights a crucial aspect of working with AI: the importance of understanding how these models process information and tailoring the input accordingly.
The success of Notion's AI agents has significant implications for the future of productivity software. By simplifying the interaction between users and AI, Notion has made powerful AI tools more accessible and user-friendly. This approach could pave the way for wider adoption of AI in various fields, as developers learn to prioritize simplicity and clarity in their AI integrations.
The development also underscores the ongoing evolution of AI. While complex algorithms and vast datasets remain essential, the ability to effectively communicate with AI through simple prompts and clear instructions is becoming increasingly important. This shift requires a new set of skills, focusing on prompt engineering and human-AI interaction.
Currently, Notion is focused on further refining its AI agents and exploring new ways to integrate AI into its platform. The company's experience demonstrates that sometimes, the most significant breakthroughs come from simplifying the complex.
Discussion
Join the conversation
Be the first to comment