📰 Dev.to · Harrison Guo

CodeGraph is the LLM-symbol-graph my prior retrieval post argued should exist. Read against its own SQLite index: why its architectural choices are right, and w

Go channels are synchronization primitives, not queues. They deliver backpressure only when the producer is bounded — and that's where the real OOM hides.

Same model, same test cases, 20% better results. 7 out of 8 fixes were pure code, zero LLM cost. Here's exactly what I changed and why it worked.

Most discussions of Node vs Go concurrency stop at 'async vs threaded.' The real split is deeper — where does context switching happen, and what is the unit of

gRPC interceptors are where cross-cutting concerns live — auth, tracing, retry, metrics, rate limiting. Most examples online show toy single-interceptor demos.

Go 1.18 shipped generics in March 2022. Four years later, the honest picture: they're useful for a narrower set of problems than the community hoped, they fit n

Most performance advice for Go is ritual — 'use sync.Pool,' 'avoid interface boxing,' 'preallocate slices.' Useful sometimes, hollow often. A production enginee

sync.Pool is the Go feature most likely to be used incorrectly. A working engineer's guide to when pooling buffers actually saves GC pressure, when it just adds

After Four Pillars of Concurrency, the natural question: what does a system actually look like when it uses all of them deliberately? IronSys is a composite blu

Docker is not a virtual machine. Kubernetes is not a container tool. Fifteen years in, both are misunderstood — and misused — as a result. A working engineer's

The 'which language is fastest' benchmark wars miss the real question. Rust, Go, Java, and Python aren't competing on the same axis. They're tuned for different

How to use OpenAI's Codex plugin inside Claude Code — turning Claude Opus and GPT-5.4 into a dual-brain coding system. Setup, commands, rescue workflows, and wh

Inside query.ts — the 1,729-line async generator that is Claude Code's beating heart. 10 steps per iteration, 9 continue points, 4-stage compression, and stream

Why blind fail-fast during leader election causes retry storms, and how bounded retry budgets, failure boundaries, and error normalization create predictable di