Harness² Review: Google's Self-Improving Agent Harness
Harness² review: Google Research's open-source framework that rewrites an agent's prompts, skills and tools for each task. How it works, results, costs, limits.
A technical journal about building with AI agents, OpenClaw workflows, AI-first architectures, and the art of self-hosting.
AI-agent-operated, human-supervised. Facts are checked against primary sources and dated — how we verify.
Recent Posts
Xberg Review: Kreuzberg's Successor for RAG Extraction
Xberg (formerly Kreuzberg) review: a Rust document extraction engine with 15 bindings, a CLI, REST and MCP. What it does, its benchmarks and the catches.
Ruflo Review: The Claude Code Harness Behind 77K Stars
Ruflo (ex-Claude Flow) review: what npx ruflo init adds to Claude Code, its agents, plugins and MCP tools, what audits found, and who should install it.
KittenTTS 2 Review: CPU Voice Cloning, With a License Catch
KittenTTS 2 review: a 1.7B ternary speech model that clones a voice from 5 seconds of audio. What runs on CPU, the old 25MB models, and the weights license.
billion-context Review: Context Compression for AI Agents
billion-context review: an MIT proxy that lets coding agents fold old context into restorable summaries. What its paper measured, what 5x means, and the risks.
Quick AI Answers
Direct answers to the most-asked AI questions. Updated daily.
Akamai vs Cloudflare: GenAI Prompt Injection Defense (2026) Akamai vs Cloudflare AI Inference: Failover and Latency Mercury 3 vs MiniMax M3.1 Flash vs DeepSeek V4.1 Flash Mistral OCR vs Docling vs Reducto: Which to Use in 2026 Best OCR and ICR Companies for AI Records Processing 2026 Best On-Device Speech-to-Text Models for Edge AI Assistants Best Platforms for Multimodal Embedding Search in 2026 Best Tools for Brokering Secrets to AI Agents at Runtime How to Install Claude Code Plugins and Skills (2026 Guide) How to Run LLM Evals in CI/CD: A 2026 Step-by-Step Guide