Matt Pocock
Become a TypeScript wizard with tips, tricks and tutorials. Plus, updates from the latest TypeScript releases (and other open source awesomeness).
Never Trust An LLM
The video explains that LLMs frequently hallucinate and introduces a taxonomy of these errors (intrinsic vs extrinsic information, factual errors, contextual inconsistency, and fabricated entities). It then discusses why hallucinations occur (training/evaluation incentives, reward guessing) and offers practical ways to work around them (prefer intrinsic information, use search tools, verify critical outputs, and deploy prompts that fetch external data).
Claude Code tried to improve /init... Is it any better?
A reviewer examines the latest init updates for Claude code, tests the new initialization flow in their repo, and weighs design choices for claw.md, skills, and hooks, offering критical feedback and practical adjustments.
Building a REAL feature with Claude Code: every step explained
The video walks through practically using Claude Code in a real project, showing how to structure features as ghost and real lessons or courses, grill ideas with an AFK agent, and ultimately translate insights into PRDs and issues while detailing UI, data flow, and testing considerations.
5 Claude Code skills I use every single day
The speaker, an engineer, outlines how to build and steer AI agents with strict, well-defined workflows. He introduces a framework of interlinked skills (Grill Me, PRD, TDD), a design-tree approach to decompose problems, and practical practices to ensure shared understanding, testability, and durable code as AI-assisted development scales.
Your codebase is NOT ready for AI (here's how to fix it)
The video argues that AI-driven code changes depend heavily on how well your codebase is structured, advocating for deep modular design and clear interfaces to make AI interaction predictable and testable. It then outlines practical benefits and design principles to reduce cognitive load and improve AI usability, finishing with actionable tips and a call to subscribe for more insights.
How to actually force Claude Code to use the right CLI (don't use CLAUDE.md)
The video explains how to deterministically steer Claude-based workflows using claw code hooks, especially to prefer PNPM over npm and to block or redirect certain commands. It walks through building a claw.md with pre-tool and post-tool hooks, turning instructions into deterministic Claude code, testing the hooks locally, and applying the approach to manage instruction budgets and prevent destructive actions.
Never Run claude /init
The video argues against creating or heavily detailing claw.md and agent files, warns about the token cost and drift from keeping context lean, and emphasizes using a structured explore phase plus careful claw.md contents to guide LLMs effectively. It then lays out practical guidance on context management, file architecture, and what belongs in claw.md versus memory and skill blocks, ending with next steps for deeper learning.
Red Green Refactor is OP With Claude Code
The video explains the enduring value of red green refactor in TDD, how it fits into AI-assisted coding, and how to apply an incremental one-test-at-a-time approach to build reliable software with strong feedback loops.
Get daily AI recaps from
Matt Pocock in your inbox
Get AI-powered summaries delivered to your inbox. Save hours every week while staying fully informed.