AI
Coding agents, LLM workflows, and the parts that make them useful instead of just expensive autocomplete
Coding agents, LLM workflows, and the parts that make them useful instead of just expensive autocomplete
CI pipelines, build systems, runners, and the YAML everyone edits but nobody owns
Command-line tools, terminal UX, and the output contracts that keep scripts from hating you
Parsers, type checkers, runtimes, and the design choices that make languages feel coherent
Containers, Dockerfiles, networking, deployment, and the cache misses everyone blames on Docker
Game development experiments, Godot projects, and notes from trying to make code feel playable
Git workflows, commit hygiene, and tools built on top of the boring thing everything depends on
Go code, CLIs, libraries, compilers, and the restraint that makes the language useful
Godot notes from small games, fast iteration, scene trees, and the parts backend work does not prepare you for
LLMs in real software work: useful sometimes, noisy often, better with actual process
Music, listening habits, and the software that quietly shapes both
Next.js outside the Vercel comfort zone, mostly Docker, Nginx, and fewer deployment myths
Logs, metrics, traces, and the awkward truth that you cannot operate what you cannot see
Platform engineering, developer infrastructure, and the work that makes everyone else faster or miserable
Web work without the ceremony: HTML, CSS, and enough JavaScript to get the job done
Yar, a compiled language with explicit errors, generics, structured concurrency, and tooling built alongside it