AlphaDev
Paper preview

Paper summary
DeepMind's deep RL agent discovering faster sorting algorithms from scratch, now in LLVM.
Ask this paper
01
Assembly-level discovery: Searches over CPU assembly instructions rather than high-level code, finding micro-optimizations humans would miss.
02
LLVM integration: Discovered sorting routines were integrated into the LLVM C++ standard library - the first major AI-discovered algorithm in production compiler infrastructure.
03
Human-beating benchmarks: Found 70% faster sorting for very small inputs and 1.7% faster for large inputs, running billions of times per day worldwide.
04
Algorithm discovery AI: A proof point for AI-driven algorithm discovery that would later be extended to matrix multiplication (AlphaEvolve) and other primitives.