Fireside chat
summary
History of LLVM Clang Swift and MLIR leads to a new stack: a compiler-first language built on MLIR that keeps python compatible syntax while adding optional static types and compile-time metaprogramming. The team demonstrates high-performance matrix kernels and heterogeneous-accelerator code generation, claims 10–100k× speedups when hot paths get type annotations, and reports about 150,000 users within six months plus a 22,000-person Discord. Mojo targets Jupyter, interoperates with PyTorch/JAX layers, and focuses on predictable, extensible backends for accelerators.
AI-written summary · about
chapters (8)
- 0:00Introduction to Chris Lattner
- 3:54Engineering career philosophy
- 8:26Swift and community growth
- 11:05The vision behind Mojo
- 18:53Language design and adoption
- 23:06Managing project expectations
- 27:22Hardware-software co-design
- 30:39Q&A: Mojo in practice
related talks
53mLLVM
48mCross-Domain Compilation: Exploiting Synergies Across the CS Community
32mGetting started with LLVM using Swift
01h 04mJust-in-Time Compilation: The Next Big Thing?
34mAsk Me Anything with Chris Lattner
11mEnabling Composition of Kernels and Compilers
06mLightning Talk: Keeping Up to LLVM IR: An Introduction to LLVM IR
01h 28mA Tour of Julia