S StackNova.

Learn Modern
Engineering

02 / FEATURED ARTICLES

Read these first.

Five articles we keep coming back to - one story about what Claude Code actually changes, one focused essay on Copilot vs Claude Code, the short read on the toolchain every web developer should know in 2026, the long view on how we got to LLMs, and the runtime-internals deep dive on the JavaScript event loop.

The migrated 2018 snake game in mid-play at playsnake.online - a top-down brown sand arena framed by green bamboo and jungle leaves.
AI · Case study Level 101 · 7 min

Claude Code is a Magician

A real 2018 snake game - ASP.NET Core, gulp, .cshtml, scattered JavaScript - migrated to HTML5 + TypeScript + esbuild in a single Claude Code session. Twenty minutes of work, screenshot by screenshot.

Read the article →
Two workbenches side by side - an engineer's precision workbench labelled COPILOT THE ENGINEER, and an artist's studio labelled CLAUDE CODE THE ARTIST.
AI · Temperament Level 101 · 8 min

Copilot vs Claude Code: engineer vs artist

A focused take on the one axis that most decides which AI pair programmer fits which task. One tool ships clean parts; the other proposes whole designs. Knowing which is which is most of the choice.

Read the article →
A spiral-bound grid notebook on a dark wooden desk, handwritten in black ink: Good context = better answers, and below it, underlined twice, Garbage in = garbage out. A black and gold pen rests on the page, with a brass globe paperweight, a leather notebook and a cup of coffee around it.
AI · Context Level 101 · 7 min

Garbage in, garbage out

AI is not magic, it is a reflection - the four properties that make context good, the four ways it goes bad, and why pasting everything you have makes the answer worse.

Read the article →
A painted tree infographic in a wide valley titled How We Got to LLMs, with the line: many roots, one trunk, many branches. Wooden plaques stack up the trunk from Early AI through Machine Learning, Neural Networks, Deep Learning, Transformers and Foundation Models, with LLMs at the crown and a single green shoot above it reading LLM Agents.
AI · History Level 101 · 14 min

How We Got to LLMs

AI did not start in 2022, and an LLM is not a new kind of thing. Rules written by hand, then rules learned from data, then layers that learn the features too, then attention in 2017 - one trunk from the 1950s to the model you typed into this morning.

Read the article →
Philip Roberts on stage at JSConf EU 2014, presenting his Loupe call-stack visualiser projected on the screen behind him.
Web · Runtime Level 401 · Expert · 11 min

The event loop in JavaScript

How a single-threaded language handles concurrency without blocking. The call stack, the host environment, the task and microtask queues, and the algorithm that ties them together - paired with Philip Roberts' definitive JSConf talk.

Read the article →
04 / LEARNING PATHS

Seven tracks. Pick one
and commit.

Every path is a curated sequence - built to take you from first principles to shipping in production, without the detours.

The library
VOL. 02 · MAY 2026
PATH 0157 articles

AI

Build with the new generation of agents. From prompting fundamentals to design systems for AI surfaces.

Open this path →
PATH 0214 articles

Cloud

Containers, orchestration, and the runtime your services actually run on.

Open this path →
PATH 039 articles

Data

Schemas, queries, and the storage decisions that outlast every framework. The data layer, without the hand-waving.

Open this path →
PATH 0414 articles

Design

Design systems, UX and UI fundamentals, and the tools design work ships through - plus how design hands off to engineering.

Open this path →
PATH 0519 articles

Engineering

Distributed systems, instrumentation, and the platform craft beneath every product.

Open this path →
PATH 069 articles

Marketing

Launch it, write the content, run the ads, read the SEO, wire the analytics - the growth stack modern builders ship alongside the product.

Open this path →
PATH 0724 articles

Web

Modern build tooling and the frameworks shaping the next decade of the browser.

Open this path →
05 / FEATURED MAPS

Start here.

Curated reading orders across the library - each one sorts a subject into layers and tells you where to start. Maps can pull articles from more than one Path, which is the point of having them. All maps →

MAP · AGENTIC AI / 7 MIN READ

Agentic AI stack

The agent slice of the AI library, walked in full - the idea, the context window, the extensions, the architectures past one agent, and what production takes - with three reading orders.

Open the map →
MAP · AI / 7 MIN READ

AI stack

Every AI Path article outside the agent layer, sorted by layer - from the engine and the platforms up to the developer tools - with a hand-off to the Agentic AI stack and four reading orders.

Open the map →
MAP · AI × DESIGN / 8 MIN READ

Claude Design for designers

The adoption route for a designer meeting Claude Design - the tool, the design-system language it reads, making yours machine-readable, and the fidelity ladder it hands back.

Open the map →
MAP · DESIGN / 9 MIN READ

Design stack

Every Design Path article sorted across eight layers - language and tokens, craft, vocabulary, fidelity, tools, handoff, brand and graphics - plus the AI design layer on top, with three reading orders.

Open the map →
MAP · WEB / 6 MIN READ

Internet search

The route through the search series - the pipeline inside a search engine, the two files that steer its crawler, the honest fundamentals of ranking well, the tags that measure the traffic, and the AI answer layer on top. Fourteen articles across Web, Marketing, and AI.

Open the map →
MAP · MARKETING / 6 MIN READ

Marketing analytics

Getting traffic and measuring where it came from - the tagged links you send, the campaign they roll up into, the two ways people find you organically, and how you measure and optimize what comes back. Nine articles, the sibling of the internet-search map.

Open the map →
MAP · ENGINEERING / 7 MIN READ

Modern engineering

The platform craft underneath modern software - orchestration with Aspire, observability across services, and API documentation on top of OpenAPI. Three layers, every Engineering Path article on that surface - the Aspire deep dives and recipes included.

Open the map →
MAP · LIFECYCLE / 9 MIN READ

Ship a product

The flagship route across every Path - design, build, ship, and grow a product end to end, with AI woven through, plus three reading orders for founders, developers, and managers.

Open the map →
MAP · WEB / 7 MIN READ

Web stack

Every Web Path article outside the search series, sorted by layer - language, build tools, frameworks, API docs, orchestration, and the browser platform the result ships into - with a one-line reason and three reading orders. Pulls in API doc pieces from Engineering and hands the search series to the Internet search map.

Open the map →