scripod.com
Spec-driven development: The AI engineering workflow at Notion | Ryan Nystrom

Overview

Shownote

Highlights

Transcript

Chapters

Pins

Spec-driven development: The AI engineering workflow at Notion | Ryan Nystrom

How I AI

2 DAYS AGO
How I AI

How I AI

2 DAYS AGO
OverviewShownoteHighlightsTranscriptChaptersPins

Shownote

Ryan Nystrom is a software engineer at Notion. He joined in December 2024 after Notion acquired Campsite, the team communication platform he co-founded with Brian Lovin. At Notion, he’s been a core builder of Notion AI and the Custom Agents feature launche...

Highlights

Ryan Nystrom, a software engineer and engineering manager at Notion, shares firsthand insights into how AI is transforming engineering workflows—not just as a productivity tool, but as a fundamental rethinking of collaboration, code review, specification, and leadership.
00:00
Whisper and Codex save time compared to traditional methods in spec-driven development
02:48
AI has completely changed my work routine, making me change tools frequently
07:16
A custom Notion AI agent compiles pre-reads from Slack, Notion tasks, merged PRs, and yesterday's meeting transcript
09:00
Detailed stand-up meetings can improve product building, especially for remote teams
11:19
This kind of sharing democratizes information, raising the visibility of everyone's work
12:15
Getting rid of excessive meeting prep and having managers code until stand-ups can prevent burnout
14:26
Line managers, directors, VPs, CTOs, and CPOs should write code
16:13
The 'hot potato' agent runs daily at 9 a.m. to automatically compile project updates, meeting notes, and Honeycomb data into a Slack pre-read
20:39
AI agents complete tasks just-in-time without complaint, leading to a more relaxed work environment and reduced burnout
27:32
Boxy generated a complete pull request with UI verification screenshots in 21 minutes after triggering via Codex comment
29:21
Code review evokes strong emotions like confusion and insecurity
31:07
Ask Codex to build your VM and background agent strategy
32:03
Codex built the feature directly from the spec authored in markdown
35:10
The spec is the source of truth for how a part of Notion AI works, written in plain English
37:53
Verification can be done on the live and working implementation
41:34
A faster CI allows agents to work non-stop and dramatically increases productivity
45:21
Challenging the AI to defend its arguments improves spec quality in CI workflows

Chapters

Introduction to Ryan Nystrom
00:00
How AI has upended 12+ years of the same working routine
02:48
Project Afterburner: Notion’s push to cut CI time to a quarter
04:30
Why high-frequency, high-quality meetings beat lower-frequency standups
09:00
How automated context surfaces every engineer’s work equally
11:10
Why cutting meeting prep is a burnout protection mechanism
12:15
The case for engineering managers writing code
14:26
Inside “Boxy”: Notion’s internal VM-based background agent system
16:13
Old World vs. New World code review
20:30
Prompting Codex from Notion comments
24:51
The emotions around code review
29:20
Quick recap
31:01
Spec-first development: writing and checking agent specs into the repo
32:00
The spec as changelog: version control for how a feature actually works
35:10
How engineers’ roles are evolving
37:53
Lightning round
39:00
Where to find Ryan
45:21

Transcript

Ryan Nystrom: One line that I've been putting in my prompts lately is I literally don't know what I'm doing here. You got to explain it like, I'm a five-year-old. I didn't start with writing code. I didn't start with anything. I just started with an empty ...