All Posts

Is Rails Slow? I Built the Same Blog API Eight Times to Price the Framework

"Rails is slow" is an unanswerable question, because a Rails app and a Go app are never doing the same work. So I narrowed it: inside one runtime, what do the framework and the ORM cost? Four runtimes, each built twice, all returning identical JSON from the identical four SQL statements, measured three times on Docker Linux with MySQL. The magic tax in CPU per request came out at ×7.67 for Ruby, ×6.14 for Python, ×2.62 for Node and ×2.51 for Go — and I counted, object by object, exactly what Active Record builds on every request.

ruby rails activerecord orm benchmark performance django sequelize gorm mysql docker

2026년 9월 1일

bitchat Architecture: How Do You Deliver a Message With No Internet and No Radio Range?

bitchat is a P2P messenger with no accounts and no servers. Nearby devices form a BLE mesh; distant peers are reached over Nostr. The center of this piece is the four-layer store-and-forward stack that answers "the recipient is not here right now" — a sealed outbox, courier envelopes addressed by nothing but an HMAC tag plus spray-and-wait, GCS-filter public history sync, and Nostr mailboxes — traced through the code.

bitchat bluetooth-mesh nostr noise-protocol p2p store-and-forward swift architecture

2026년 7월 25일

Hallmark Architecture: How Do You Stop Designs From Looking 'AI-Generated' — With a Prompt?

Hallmark is Together AI’s anti-AI-slop design skill. It ships almost no executable code — its entire intelligence lives in 106 markdown files. This piece dissects the SKILL.md router and its progressive context loading, the narrowing decision pipeline, and — at its center — the 58 slop-test gates, contrasting it with skill-injecting projects like Superpowers and SkillSpector.

hallmark prompt-engineering ai-agent skills design claude-code llm architecture

2026년 7월 24일

Llama 3 (2024) Paper Notes

Paper notes on The Llama 3 Herd of Models: 15T tokens, a 405B flagship, 128K context, and a design philosophy of 'managing complexity' (dense over MoE, SFT → rejection sampling → DPO instead of PPO) — the paper where an open-weight model first reaches GPT-4-class performance. It completes Llama 2's skeleton through scale, data, and simplicity.

llm paper-reading transformer llama-3 dpo alignment open-source meta

2026년 7월 24일

Llama 2 (2023) Paper Notes

Paper notes on Llama 2 covering its commercial license, a base model scaled to 2T tokens with 4k context and GQA, and the Llama 2-Chat alignment pipeline (SFT → two reward models → iterative RLHF with rejection sampling + PPO → GAtt). It fuses LLaMA's open base lineage with InstructGPT's RLHF at production quality.

llm paper-reading transformer llama-2 rlhf alignment open-source meta

2026년 7월 13일