Workbench
Build logs, debugging notes, architecture sketches, and rough technical field notes from projects in progress. Smaller and messier than essays, but more useful than a changelog.
Why building a small OS is a good way to understand bootloaders, memory, interrupts, and how software starts running.
A field note on why QEMU makes OS dev approachable, why serial logs matter, and why a blank screen can have many causes.
Notes on why writing pixels directly to framebuffer memory makes OS development feel visual for the first time.
Dense UI is not automatically bad. Finance tools need speed, filters, drilldowns, saved views, exports, and auditability.
Why service layers help avoid hardcoded UI data and keep enterprise finance modules consistent.
Notes on assistant families, prompt precheck, model routing, usage tracking, admin activation, and plan requests.
Different tasks need different models, and cost, latency, quality, modality, and plan limits all matter.
Product notes on why productivity apps need opinionated workflows, calendar-style organization, and sticky daily loops.
Clone projects are not the destination, but they honestly show the path from layouts and APIs to real systems work.
Exploratory papers, especially in quantum computing architecture, improve tradeoff thinking and connect research back to builds.
Workbench is not blog
Blog posts are polished essays and finished explanations. Workbench entries are build logs, debugging notes, architecture sketches, raw experiments, half-formed ideas, and what-I-learned-today notes.