Blog

Notes from building QueryMT

Notes from building QueryMT: experiments with models, agents and tooling, and what actually works in practice.

2026-11-11 25 min read

Does AREX-2 actually self-improve?

AREX-2 against its own base model, Qwen3.8-27B, and against GLM-5.3-flash, on one optimisation task with an external scorer. Nine runs on two local GPUs, with charts, the scoring harness, and the reasoning traces where …

Llm Agents Local-Models Openspec Benchmarks Querymt

2026-03-30 12 min read

qndx: Building a Benchmark-Driven Regex Search Indexer in Rust

Rust Search Indexing Performance Benchmarks Qndx

2026-02-17 4 min read

AGENTS.md and Skills: useful, but narrower than the hype

Llm Agents Agents-Md Skills Benchmark Software-Engineering

2026-01-09 3 min read

Anthropic Just Targeted OpenCode: A Masterclass in Corporate Strategy Disguised as Security

Rust Oauth Security Anthropic Developer-Experience

2026-01-06 4 min read

Stop Paying Twice: Bringing Your AI Subscriptions to Rust

Rust Oauth Security Anthropic Openai

All posts