Tutorials

VIDRAFT's open leaderboard scores global weather models daily against ERA5. Darwin-Prediction-Weather has the lowest 2026 error in all four elements.

CSS-Tricks just soft-launched a redesign after 19 years. Here's a quick overview of what changed before the full write-up drops Monday.

The final What's !Important column covers CSS random(), new gradient types, icon implementation, and fresh web platform features across Chrome and Firefox.

Learn how to give a coding agent a real browser with Playwright MCP and write instructions that test actual UI changes, not just whether a page loads.

Gabriela Moreira explains how TLA+ and Quint help engineers model and verify system behavior before bugs reach production.

MCP adoption was fast, but the protocol outpaced its infrastructure. Here's what gaps remain in auth, observability, and tool management.

Datadog CPO Yanbing Li explains what it takes to monitor AI agents in production, from LLM call tracing to continuous feedback loops.

Level 3 of a six-level LLM maturity model: compose calibrated confidence scores, judge high-stakes decisions, and route uncertainty to humans reliably.

Treat LLM evals like CI tests: a gate that fails the build on regression. Then add baseline comparison and shadow evals to catch slow drift.

Learn how to protect sensitive PII like SSNs in data pipelines using deterministic encryption, key derivation, tokenization, and other cryptographic techniques.

Build ShopHelper, a production-ready customer support assistant using the Claude API, covering conversation history, tools, workflows, and prompt evaluation.

Learn TimescaleDB to optimize PostgreSQL for time-series data. Covers hypertables, compression, continuous aggregates, and two production projects.