Stories by matt_d

Inside Kimi K3's AgentENV: Can It Really Fork in 100 ms?

Enabling the next iteration of the borrow checker on nightly

Agentic Coding in the Wild: Characterizing GitHub Copilot Traces at Production

Certified Program Synthesis with a Multi-Modal Verifier [pdf]

Compiler-Grounded Hierarchical Diagnosis for LLM Triton Kernel Optimization

NIIN: NVSHMEM Implemented In NCCL

ACM SIGOPS Symposium on Operating Systems Principles SOSP'26 Papers & Preprints

Reliability Lessons from SQLite – Richard Hipp – SSW 2026 [video]

Stanford CS329A: Self-Improving AI Agents

Demystifying Deep Learning Compiler Front End Bugs: An LLM-Aided Empirical Study

Towards Designing an Execution Control System with Metastability Resilience

AdaMAST: Adaptive Failure Taxonomies for Improving LLM Agents

k-Coloring is Faster than Computing the Chromatic Number

x86 AMX/ACE with >8 tiles

RSI: Recursive Self-Improvement

Specula: Scaling formal specs for autonomous model checking of system code

NVIDIA-labs OO Agents: Native Python Object-Oriented Agents

At-the-Roofline Sparse Tensor Contractions on Vector Processors for Inference

Characterizing Warp Divergence from Pascal to Blackwell

Why Higher-Order Logic Is a Good Foundation for Deep Verification

1 2 3 4 5 6 7 8 9 10