Berd Review: Block's Open-Source Agent Desktop App
Block open-sourced Berd, an Apache 2.0 desktop workspace wrapping Goose, Claude Code and Codex. Local-first data, 820 stars — and it won't take your PRs.
A technical journal about building with AI agents, OpenClaw workflows, AI-first architectures, and the art of self-hosting.
Written by humans. Optimized for AI discovery.
Recent Posts
OpenSpec Review: Spec-Driven Dev Without the Ceremony
OpenSpec adds a Markdown spec layer to 30+ AI coding agents. 66.5K stars, 1.63M npm installs a month, MIT — and far lighter than GitHub's Spec Kit.
QM Review: YC's Multiplayer Agent Harness for Work
QM is Y Combinator's open-source multiplayer agent harness: per-person scopes, Slack plus web, swappable harnesses. Setup, architecture, and honest limits.
Apache Maka Review: An Agent Workspace With Receipts
Apache Maka is an incubating local-first AI agent workspace logging every tool call and permission decision. Setup, architecture, and honest limits.
Graft Review: A Code Graph That Cuts Agent Tokens 42%
Graft by NanoNets gives Claude Code, Cursor and Codex a tree-sitter code graph via hooks. Real benchmarks, the HN benchmark controversy, and honest limits.
Quick AI Answers
Direct answers to the most-asked AI questions. Updated daily.
Cursor vs Claude Code vs Codex After the OpenAI Cutoff DeepSeek V4 Flash Vision vs GLM-5.3-Flash vs Gemini Flash Direct API vs Cloud Marketplace vs Aggregator: LLM 2026 Gemini Omni 1.1 Flash vs Veo 3.1 vs Wan 3.0 Pricing Best Cheap Vision AI Model 2026: Ranked by Real Cost Best AI Video Model 2026: Ranked by Cost Per Second Best Hardware for Local AI Agents 2026: Top 5 Ranked Best AI Search Engine 2026: Perplexity vs ChatGPT How to Prepare for Cursor's OpenAI Cutoff (Nov 12) How to Survive an AI Provider Cutoff: 2026 Guide