Blog
Hear from top AI professionals about open-source compilers, cutting-edge hardware, and all things Kernelize.
Apr 14, 2026
10
min read
AI is Revolutionizing Compilers Not Replacing Them
The future of AI-driven model development depends on well-structured open compiler workflows and systems that keep improving over time.
By
Simon
Waters
,
Bryan
Bowyer
Read more
Mar 10, 2026
10
min read
Why CUDA No Longer Scales for New Model Support
The newest models are using more and more Triton, both on the newest GPUs and on previous generations.
By
Simon
Waters
,
Bryan
Bowyer
Read more
Feb 18, 2026
10
min read
The Path to Making AI Hardware Diversity Practical
GPUs are here to stay, so what does that mean for all the new AI inference hardware?
By
Simon
Waters
,
Bryan
Bowyer
Read more
Feb 5, 2026
10
min read
Why Heterogeneous AI Inference Clusters Don't Work (Yet)
Heterogenous hardware should be deployed everywhere, but it is blocked by software
By
Simon
Waters
,
Bryan
Bowyer
Read more
Jan 27, 2026
3
min read
Turning Triton into a Truly Portable Language
Announcing Triton Extensions and Kernelize
By
Simon
Waters
,
Bryan
Bowyer
Read more
Jul 2, 2025
10
min read
Community Beats Code: How to Bridge the NPU Software Gap
Part 3 of our series on the path to efficient AI: How to bridge the NPU software gap.
By
Simon
Waters
,
Bryan
Bowyer
Read more
Jun 18, 2025
10
min read
The NPU Software Gap: How to Build Compatibility for NPU Migration
Part 2 of our series on the path to efficient AI: How to build compatibility for NPU migration.
By
Simon
Waters
,
Bryan
Bowyer
Read more
Jun 18, 2025
10
min read
Beyond CPUs and GPUs: The Rise of Specialized AI Hardware
Part 1 of a three-post series exploring custom hardware for efficient AI.
By
Simon
Waters
,
Bryan
Bowyer
Read more
