Skip to content
No PriorsNo Priors

From Coder to Manager: Navigating the Shift to Agentic Engineering with Notion Co-Founder Simon Last

Notion isn’t designing AI agents that just use tools. Their agents can autonomously build their own integrations, as well as write the code needed to finish a task. Sarah Guo sits down with Notion Co-Founder Simon Last to explore Notion’s rapid evolution from a simple writing assistant to a sophisticated platform for custom AI agents. Simon discusses the technical hurdles of indexing disparate data from sources like Slack and Google Drive, as well as the internal shift toward using coding agents to build Notion itself. Plus, Simon elaborates on what he sees as a fundamental transition in productivity: moving from a tool where humans do the work, to one where humans manage a swarm of agents. Sign up for new podcasts every week. Email feedback to show@no-priors.com Follow us on Twitter: @NoPriorsPod | @Saranormous | @EladGil | @simonlast | @NotionHQ Chapters: 00:00 – Cold Open 00:05 – Simon Last Introduction 00:26 – Genesis of Notion AI 04:10 – Challenge of Semantic Indexing and Retrieval 07:16 – The Six-Month Rewrite Cycle 08:12 – Notion’s Coding Agent Era 09:44 – Impact on Team Dynamics 12:49 – Launching Custom Agents 15:39 – Notion as the ‘Switzerland’ for Models 17:33 – Designing APIs for Agent Customers 20:09 – Simon’s Personal Agentic Workflows 24:48 – Notion: Tool for Work is Now A Tool for Agents 27:28 – How Building Has Changed for Simon 29:00 – Conclusion

Sarah GuohostSimon Lastguest
Mar 11, 202629mWatch on YouTube ↗

At a glance

WHAT IT’S REALLY ABOUT

Notion’s shift to agentic work: indexing, APIs, and teams

  1. Notion’s AI journey began with early GPT-4 access in 2022, leading to a fast launch of an AI writing assistant and a longer push toward a general-purpose workspace agent.
  2. A major technical inflection point was building high-quality semantic indexing and retrieval across Notion and third-party sources (e.g., Slack, Google Drive), requiring empirical iteration on chunking, pipelines, and evals.
  3. Notion repeatedly rewrites its “AI harness” roughly every six months to match rapid model progress, and the rise of coding agents has made larger rewrites and more ambitious PRs feasible.
  4. Notion’s product direction is now “tool for humans” plus “tool for humans managing agents,” including personal agents, newly launched custom agents, and agent-friendly APIs (markdown-like pages + SQLite for databases).

IDEAS WORTH REMEMBERING

5 ideas

Ship short-term wins while building toward an agentic north star.

Notion paired a quick-to-deliver writing assistant with a longer-term bet on a general assistant that can use all Notion tools; the latter required years of iteration until models and harnesses caught up.

Indexing and retrieval quality is mostly craft plus relentless empiricism.

Simon argues many companies underperform on search because they don’t iterate empirically; each data source (Slack vs Drive) needs tailored retrieval tactics, chunking, and continuous tuning.

Embeddings reduce dependence on users’ folder/tree organization.

Because retrieval can work from semantic snippets, Notion increasingly advises users not to over-optimize workspace structure—focus on getting information into the system so it can be retrieved.

AI systems demand frequent rewrites to stay aligned with model capabilities.

Notion “rewrites the AI harness” about every six months, treating it as necessary product engineering rather than tech debt, because model behavior and best practices change quickly.

Coding agents increase ambition—but only with strong verification loops.

Simon distinguishes robust agent-driven development (clear specs, tests, safe deploys) from “slop”; PRs are bigger, but expectations for end-to-end testing and review rise accordingly.

WORDS WORTH SAVING

5 quotes

We rewrite our AI harness probably every six months or so.

— Simon Last

You can be like 100 or 1000X engineer if you're using the tools right now.

— Simon Last

If you do it badly, it's all slop.

— Simon Last

We see ourselves as kind of like the Switzerland for models.

— Simon Last

Before AI, our goal was to create the best tool for humans to directly perform their work. And then now the goal is to create the best tool for humans to manage agents to do the work for them.

— Simon Last

Origin of Notion AI and GPT-4 catalystAI Writer → Q&A retrieval → full workspace agent roadmapSemantic indexing across heterogeneous sourcesChunking/retrieval pipeline iteration and eval rigorSix-month AI harness rewritesCoding agents changing engineering output and verification loopsCustom agents, autonomy, and agent-friendly APIs

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.

Add to Chrome