Jonathan Beard's blog about systems, performance, and tech.
AI gets you to a working draft fast. The trouble is what the missing fifth contains, the edge cases and operational reality that used to teach engineers their judgment, and who builds that muscle now.
A generated answer is an abstraction over reasoning you never see. It works until it leaks, and when it leaks it hands you a bill for exactly the understanding it let you skip.
Adding cores quietly stopped working, and Amdahl told us why forty years early: scaling stalls on the one dependency you can't parallelize away, not the resource you keep buying.
When can you abstract something away for free? The condition has two names, context-free and zero mutual information, and almost nothing in a real system meets it.
Every abstraction is borrowed complexity: a quiet debt that comes due at 3am, in the layer you chose never to learn. On staying ahead of the bill.
How a wedding photo booth led me to a Nikon Zf, a bag of cheap vintage glass, manual focus, and a running habit of photographing tiny plastic llamas wherever work sends me.
A systems architect's questions, pointed one level up: from cores and caches to minds, institutions, and the societies they run on.
Managing Parallel, Part 4: The Machine Underneath. Memory ordering (x86-TSO vs Arm), NUMA topology, and cache-line atomicity
One pool of storage that Linux, macOS, and Windows can all share: NFS for the compute nodes, SMB for the laptops, and the UID-alignment trick that stops NFS permissions from ruining your week. Built for home ML clusters and home labs.
Managing Parallel, Part 3: When Parallel Goes Wrong. The short list of broken contracts behind most concurrency bugs
Managing Parallel, Part 2
A step by step guide to using C++ in C.
Managing Parallel, Part 1
How to use templates to build functions by type size
How to host pages with GitHub pages, use jekyll and bootstrap together to make awesome pages
The folly of inventing secure languages
Data wrangling, why should you learn basic coding?
What is stream processing and how does it relate to data-flow?
A case for more programming languages
Tutorials and full code examples of how to use pipes and FIFOs to connect C/C++ compiled applications to a thread running inside a JVM using Java.
How to get flex and bison to work together in C++
How to time software properly: from rdtsc to clock_gettime and everything in between, plus timer synchronization.