Alfred walks through building production-ready AI agents using Laravel paired with Claude API. Covers tool use, conversation memory, and structured outputs. Key insight: treating agents as stateless HTTP workers with Redis-backed memory gives the best scalability.

Opinion piece arguing that traditional dev agencies will be disrupted by no-code + AI combos. Alfred shares revenue numbers from his own pivot. Controversial take: "Junior devs who can prompt-engineer will outship senior devs who can't."

Spicy thread ranking AI coding assistants. Alfred puts Cursor at #1, Copilot at #3 (controversial), and says Windsurf is underrated. Key quote: "The IDE war is over. The agent war is just beginning."

Alfred guests on Laravel News Podcast discussing how PHP/Laravel remains relevant in the AI era. Argues that PHP's simplicity is an advantage for AI-generated code. Interesting data point: 40% of his AI agent code is now generated by AI itself.

Weekly roundup covering new releases: Claude 4 Opus launch, Cursor 0.45 with multi-file agents, and a new open-source RAG framework called "Haystack 2.5". Alfred's pick of the week: using Claude artifacts for rapid prototyping.

Alfred reflects on patterns from 50+ client AI projects. Most common failure: trying to automate everything at once. Most successful approach: find the one painful manual process and nail it.

Developer shares their migration from Vercel to self-hosted Coolify on Hetzner. Honest pros/cons — cheaper and more control, but more maintenance. Good for side projects and small teams.

Alfred argues that startups building thin wrappers around AI APIs are doomed as the APIs themselves add those features. Recommends building on primitives, not platforms.

Official walkthrough of Astro 5.0's Content Layer API. Unified interface for content from any source — files, APIs, databases. Fully type-safe with Zod schemas. Game changer for content-heavy sites.

Newsletter deep-dive on Claude 4's improvements for code generation, specifically PHP. Extended context window means entire Laravel codebases can be analyzed. 40% improvement in code accuracy for Laravel-specific tasks.