Mar 12, 2026 · 29m · no-priors
From Note-Taking App to AI Workspace: The Simon Last Interview
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Notion co-founder Simon Last details how the platform evolved from a productivity workspace into an AI-orchestration engine, exploring enterprise knowledge indexing, autonomous custom agents, and the transformative impact of coding agents on software engineering workflows.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. The hosts hold 22.2% of the talking time here. How this is scored →
The hosts as informed peer 5.2 Guest teaching 4.0 Guest disagreement 1.2 The hosts pushing back 2.6
speaking balance: gold is the hosts, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 4:56 Simon calls out competitor search indexing quality
Simon candidly dismisses incumbents' native search capabilities, stating Notion was baffled by how poorly major tech companies construct their own document indexes.
Hardest push from the hosts ▶ 6:48 Sarah challenges Simon on retrieval and chunking constraintsSarah refuses to accept that tree structure doesn't matter, pressing Simon on how fundamental chunking strategies and pipeline nuances dictate search quality.
Biggest teaching moment ▶ 26:44 Simon redefines Notion's core mission for the AI eraSimon delivers an insightful framing shift, educating the audience on moving from tools where humans do manual work to collaborative environments where humans manage autonomous agent swarms.
The host holds their own ▶ 6:48 Sarah demonstrates deep technical knowledge of RAG pipelinesSarah counters Simon's high-level dismissal of document layout by pointing directly to chunking decisions and indexing architecture, forcing Simon to acknowledge the complex engineering involved.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The hosts as informed peer | Guest teaching | Guest disagreement | The hosts pushing back | Why |
|---|---|---|---|---|---|---|
| Discovering GPT-4 at Mexico Company Offsite | 4 | 3 | 1 | 1 | Sarah kicks off by asking about the origin story of Notion's AI push starting from GPT-4 access at a Mexico offsite. Simon recounts their dual-track vision for an AI writing assistant versus general assistant, gently correcting Sarah's timeline regarding their initial launch date. | |
| Developing Notion Q&A and Cross-Platform Indexing | 5 | 4 | 2 | 4 | Sarah challenges Simon on Notion's ambition to index third-party tools like Slack and Google Drive when those platforms haven't solved cross-platform search natively. Simon acknowledges the boldness of the move, noting how bafflingly poor native indexes often are and emphasizing Notion's craft-driven empirical tuning. | |
| Embeddings, Workspace Structures, and Chunking Strategies | 6 | 4 | 2 | 4 | When Simon suggests workspace hierarchy is irrelevant because embeddings handle context, Sarah pushes back by highlighting technical complexities like chunking strategies and pipeline tuning. Simon concedes that substantial engineering iteration is required under the hood. | |
| Transforming Engineering Workflows with Coding Agents | 5 | 4 | 1 | 2 | Sarah inquires how coding agents reshape engineering team sizes and the gap between median and elite engineers. Simon explains that while team sizes stay small, the ceiling of individual output has scaled dramatically with agent adoption. | |
| Rapid Prototyping, Design Playgrounds, and Automated Verification | 5 | 4 | 1 | 3 | Sarah presses Simon on risk boundaries and data loss protections given agentic development. Simon clarifies their strict review workflows, contrasting intentional test verification with unguided vibe coding. | |
| Personal and Autonomous Custom Agents in Notion | 6 | 4 | 1 | 2 | Simon outlines personal and autonomous custom agents in Notion that can bootstrap integrations. Sarah synthesizes this into the broader definition of agents that leverage code execution as a core primitive. | |
| Notion as a Model-Agnostic AI Platform | 6 | 4 | 1 | 3 | Sarah asks how Notion positions itself against major platform incumbents and how they adapt Notion's internal architecture for models. Simon explains their model-agnostic stance and how they designed custom Markdown and SQLite APIs specifically for agent consumption. | |
| Simon Last's Personal Setup and Continuous Coding Agent Runs | 5 | 4 | 1 | 2 | Simon shares his personal habits of running continuous coding agent sessions and autonomous triage agents with self-updating memory. Sarah probes the mechanisms of supervision and trust calibration. | |
| Spreading Agent Intuition to Non-Technical Teams | 5 | 5 | 1 | 2 | Sarah asks about democratizing agent intuition to non-technical teams and how Notion's foundational mission has evolved. Simon articulates the paradigm shift from building tools for humans to execute work to building tools for humans to orchestrate agents. |
Statements from this episode (19)
Simon Last: GPT-4 Proto-Interface Sparked Notion's AI Pivot
“It wasn't until I played with GPT-IV that it became really, really real. So, you know, we, when we got access to it was sort of like a proto-ChatGPT-like interface. And my co-founder Ivan and I both got access, and it was just immediately clear, like, I would …”
Notion Integrates Slack and Google Drive Into Its Q&A Index
“As soon as we got the Notion index working, it was obvious that, oh, okay, we should index everything else as well. And so we indexed, like, Slack and Google Drive, and we, we're launching new ones on a regular cadence.”
Simon Last: Most Enterprise Software Companies Build Terrible Search Indexes
“But it turns out that most of the companies are pretty bad at making their indexes.”
Simon Last: Enterprise AI Retrieval Requires Source-Specific Tuning
“Like, you know, you can't just apply a one-size-fits-all to, like, querying Slack versus querying Google Drive, let's say. They're completely different kinds of information, and we found that there's just a little bit of, like craft and love that has to go int…”
Simon Last: Vector Embeddings Make Hierarchical Workspace Organization Irrelevant
“I would say, I mean, the interesting thing is that with embeddings, it almost doesn't matter as much anymore, The AI doesn't really care what the tree structure is, for example. The, all, all the AI cares about is that there's a snippet of text that has the co…”
Notion Now Advises Users to Dump Data Without Worrying About Organization
“And so actually, we kind of advise people now, like, don't worry as much about organization. Just find a way to get it all piped in and like thrown in there.”
Simon Last: Companies Must Redesign AI Software Harnesses Every Six Months
“I think this is honestly a really key thing and something that a lot of companies get wrong is just like doing one thing and then just like sticking with it. You really do have to keenly aware of what the current state of the models and the technology is, and …”
Simon Last: AI Coding Agents Can Build More Robust Software
“The upshot is I think if you do it well, you can be much more ambitious about what you're building. And also make it much more robust than you could have done with humans writing it. And then the flip side is if you do it badly, it's all slop.”
Simon Last: Current AI Tools Enable 100x to 1,000x Engineering Leverage
“You can be like a hundred or thousand x engineer for using the tools right now. I think the gap is much bigger. Like the minimum bar has not changed, but the maximum bar has extremely increased.”
Simon Last: AI Agents Now Write All Pull Requests at Notion
“And you know, all the pull requests are now written by agents. They're often like larger and more complex. That's like the worst part. But the better part is that they're often like a much better tested and we can demand sort of a much better testing for the t…”
Simon Last: AGI Will Essentially Be a Coding Agent
“I think of coding agents as, like, the kernel of AGI. AGI will be a coding agent.”
Simon Last: Notion Now Uses Cheaper Chinese Open-Source AI Models
“We actually just released one of them in our agent last week, and we're gonna do all four for sure. They're actually quite good, and they're way cheaper than the Frontier models.”
Notion Built Custom Markdown and SQLite APIs Specifically for AI Agents
“We created sort of a markdown dialect that looks like the default normal markdown, but it's sort of enhanced with All the notion blocks. And the models are really good at it. It works really well. So, so that's how it reads and writes the pages. And then for d…”
Simon Last Let an Autonomous Coding Agent Run for 13 Days
“My personal record is that I've had a coding agent running for, I think it was, 13 days straight without stopping. I'm just basically working through, like, tasks.”
Simon Last: Start Custom Enterprise AI Agents in Approval Mode
“The general pattern I follow is sort of, I build it as a prototype, I have it in sort of like an approval mode, where I'm sort of, you know, watching it closely, and then, but then after it runs a bunch of times, you kind of trust that it's working”
Simon Last: Notion's HR Team Is Top Adopter of AI Agents
“The people team's been amazing. They're actually one of the highest adopters of custom agents.”
Simon Last: Notion Is Now a Tool for Managing AI Agents
“Broadly speaking, 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.”