Today's guest is Adam Tornhill, founder of CodeScene, author of Your Code as a Crime Scene, and one of the very few repeat guests of the show.

With Adam, we talked about how his team refactored the whole Street Fighter 3 code base, which is more than 300,000 lines of code in just a few days and spending a few thousand dollars in tokens, and measured how that led to better and cheaper AI development of new features.

And then we discussed how the CodeScene team does agentic development, what the process looks like, what the bottlenecks are, and what are the crucial parts that Adam believes we still need to do as humans.

(00:00) Episode start

(03:38) Introducing Adam Tornhill

(04:12) Refactoring 300K lines of Street Fighter with AI

(09:50) The refactoring playbook: AI-invented patterns

(21:41) Generalizing the approach to enterprise code bases

(27:12) Why documentation of the "why" matters more than ever

(32:25) CodeScene goes fully agentic: the MCP rewrite story

(38:54) The development process: uncertainty drives rigor

(47:51) Validation is the new bottleneck

-

Show notes:

StreetFighter case study:

https://codescene.com/blog/case-study-refactoring-at-scale-with-agents

One of Adam's blog posts about coding processes:

https://adamtornhill.substack.com/p/controlling-the-uncertainty-machine

MCP-guided Claude vs. unguided Claude experiment on refactoring tasks (Code health-guided MCP did 2-5x more code health improvements):

https://codescene.com/blog/making-legacy-code-ai-ready-benchmarks-on-agentic-refactoring

-

Today's episode is brought to you by Notion.

Learn more about Notion’s Developer Platform today at https://ntn.so/refactoring

-

You can also find this at:

• 📬 Newsletter: https://refactoring.fm

• 🎧 Spotify: https://open.spotify.com/show/7Luds9dmzZDoDC8Q7EMbSw

• 📱 Apple: https://podcasts.apple.com/us/podcast/refactoring-podcast/id1719137305

—

For inquiries about sponsoring the podcast, or appearing as a guest, email: luca@refactoring.club

Today's guest is Rahul Raina, co-founder of TRM Labs, which is a YC-backed company that works on solutions and products for security, fighting criminal networks, AI attackers, and more.

With Rahul, we had a great chat that covered a lot of ground, starting with the tough reality of cyber defense today, and why attackers by default always have the upper hand when AI and technology improves. And then we moved to a deep dive into how the TRM team works and how they rebuild their whole product development workflow for AI, including reshuffling all the technical roles, compressing the developer process steps, and creating dedicated artifacts for human reviews.

(00:00) Episode start

(03:50) Introducing Rahul Raina

(04:08) Why AI widens the gap between attackers and defenders

(08:54) The asymmetry: attackers don't play by the rules

(17:09) Inner loop vs outer loop: where the real bottleneck is

(20:54) Redesigning the shipping framework around jobs, not roles

(24:18) How to transition people: go slow before you go fast

(34:24) What to hire for: problem solving, agency, humble craft

(42:14) How builders review work: human-readable artifacts, not code

(46:59) The harness interviews you: AI asks questions humans wouldn't

(57:20) Build tools for understanding, not just for producing code

-

Today's episode is brought to you by Notion.

Learn more about Notion’s Developer Platform today at https://ntn.so/refactoring

-

You can also find this at:

• 📬 Newsletter: https://refactoring.fm

• 🎧 Spotify: https://open.spotify.com/show/7Luds9dmzZDoDC8Q7EMbSw

• 📱 Apple: https://podcasts.apple.com/us/podcast/refactoring-podcast/id1719137305

—

For inquiries about sponsoring the podcast, or appearing as a guest, email: luca@refactoring.club

Today's guest is Dennis Pilarinos, who is the founder and CEO of Unblocked and a long-time developer tools expert, having founded BodyBuild and having served as a director at Apple, Amazon and Microsoft.

With Dennis, we talked about the latest trends in software development. We discussed why more tokens are not translating into more productivity, how to guarantee code quality with AI and what's the role of context in making your agents succeed. And we also explored how the Unblocked team is doing it, navigating their setup and a few use cases about how their shipping features in production.

(00:00) Episode start

(03:50) Introducing Dennis Pilarinos

(04:40) Token spend is exploding

(09:03) Measuring developer productivity is still unsolved

(12:31) Code review is the new bottleneck

(18:01) Code quality depends on context, not skill alone

(21:07) The content curation trap: why markdown files go stale

(26:40) What sources to plug into your context engine

(33:47) How the Unblocked team ships: risk assessment and auto-merge

(43:29) Releasing faster: skip design partners, ship to everyone

(49:20) Remote agents and the future of context-powered coding

-

Today's episode is brought to you by Notion.

Learn more about Notion’s Developer Platform today at https://ntn.so/refactoring

-

You can also find this at:

• 📬 Newsletter: https://refactoring.fm

• 🎧 Spotify: https://open.spotify.com/show/7Luds9dmzZDoDC8Q7EMbSw

• 📱 Apple: https://podcasts.apple.com/us/podcast/refactoring-podcast/id1719137305

—

For inquiries about sponsoring the podcast, or appearing as a guest, email: luca@refactoring.club