From Swift to Mojo and high-performance AI Engineering with Chris Lattner

Chris Lattner is one of the most influential engineers of the past two decades. He created the LLVM compiler infrastructure and the Swift programming language – and Swift opened iOS development to a broader group of engineers. With Mojo, he’s now aiming to do the same for AI, by lowering the barrier to programming AI applications.

I sat down with Chris in San Francisco, to talk language design, lessons on designing Swift and Mojo, and – of course! – compilers. It’s hard to find someone who is as enthusiastic and knowledgeable about compilers as Chris is!

We also discussed why experts often resist change even when current tools slow them down, what he learned about AI and hardware from his time across both large and small engineering teams, and why compiler engineering remains one of the best ways to understand how software really works.

—
*Brought to you by:*
•⁠ Statsig ⁠ — ⁠ The unified platform for flags, analytics, experiments, and more. Companies like Graphite, Notion, and Brex rely on Statsig to measure the impact of the pace they ship. Get a 30-day enterprise trial here. http://statsig.com/pragmatic

•⁠ Linear – The system for modern product development. Linear is a heavy user of Swift: they just redesigned their native iOS app using their own take on Apple’s Liquid Glass design language. The new app is about speed and performance – just like Linear is. Check it out. https://linear.app/pragmatic?utm_source=gergely&utm_medium=newsletter&utm_campaign=pragmatic-engineer

—
*The Pragmatic Engineer deepdives relevant for this episode:*
• AI Engineering in the real world https://newsletter.pragmaticengineer.com/p/ai-engineering-in-the-real-world
• The AI Engineering stack https://newsletter.pragmaticengineer.com/p/the-ai-engineering-stack
• Uber's crazy YOLO app rewrite, from the front seat https://blog.pragmaticengineer.com/uber-app-rewrite-yolo/
• Python, Go, Rust, TypeScript and AI with Armin Ronacher https://newsletter.pragmaticengineer.com/p/python-go-rust-typescript-and-ai
• Microsoft’s developer tools roots https://newsletter.pragmaticengineer.com/p/microsofts-developer-tools-roots

—
*Where to find Chris Lattner:*
• X: https://x.com/clattner_llvm
• LinkedIn: https://www.linkedin.com/in/chris-lattner-5664498a
• Website: https://nondot.org/sabre

—
*In this episode, we cover:*
(00:00) Intro
(02:35) Compilers in the early 2000s
(04:48) Why Chris built LLVM
(08:24) GCC vs. LLVM
(09:47) LLVM at Apple
(19:25) How Chris got support to go open source at Apple
(20:28) The story of Swift
(24:32) The process for designing a language
(31:00) Learnings from launching Swift
(35:48) Swift Playgrounds: making coding accessible
(40:23) What Swift solved and the technical debt it created
(47:28) AI learnings from Google and Tesla
(51:23) SiFive: learning about hardware engineering
(52:24) Mojo’s origin story
(57:15) Modular’s bet on a two-level stack
(1:01:49) Compiler shortcomings
(1:09:11) Getting started with Mojo
(1:15:44) How big is Modular, as a company?
(1:19:00) AI coding tools the Modular team uses
(1:22:59) What kind of software engineers Modular hires
(1:25:22) A programming language for LLMs? No thanks
(1:29:06) Why you should study and understand compilers

—
See the transcript and other references from the episode at https://newsletter.pragmaticengineer.com/podcast

—
Production and marketing by https://penname.co/.