All posts
Thoughts and rambles on AI tools, LLMs, and automation for people who use love tech but don't have CS degrees.
2026 · 07
- Claude Claude Fable 5 and Mythos 5 are back July 2nd Anthropic has received notice that the Department of Commerce has lifted export controls on Claude Fable 5 and Mythos 5.
- Claude Code Is Steganographically Marking Requests A security researcher claims to have found hidden Unicode markers in Claude Code that subtly alter the date string in its system prompt based on the API base URL's hostname and system timezone—apparently to flag custom gateways, proxies, or resellers (especially Chinese AI-related domains)—raising privacy and transparency concerns about undisclosed steganographic fingerprinting in a widely-trusted developer tool.
- Introducing Claude Sonnet 5 Anthropic has released Claude Sonnet 5, a more agentic and cost-efficient model that approaches Opus 4.8's performance on many tasks, shows improved safety metrics over its predecessor Sonnet 4.6, and is now available across all Claude plans and the API with introductory pricing through August 31, 2026.
2026 · 06
- Qwen 3.6 27B is the sweet spot for local development A hands-on review of Qwen 3.6 27B, a dense local model that punches above its weight for development work, with practical setup instructions for running it via llama.cpp and benchmarks comparing it to frontier models.
- Tiny Air: A Practical MCP Server for Air Quality MCP is still early. Most of the servers I see fall into two camps: wrappers around SaaS APIs (Stripe, GitHub) that are useful but boring, and multi-tool suites that try to do everything and end up doing too much. Tiny Air is an attempt to make it more useful for the everyman.
- Change Hermes Reasoning at Runtime A quick cheat sheet when your Hermes agent is over reasoning.
- Armchair Arena - Test Ollama Models for YOUR use case A self-hosted tiny LLM arena to pick my favorite Ollama model tailored for my unique use case.
- Claude Fable 5: The "Safe" Mythos for Everyone Anthropic's "civilian" Mythos-class model, explained — the silent Opus 4.8 fallback, the specs, the silent-sabotage controversy, and the cost and data-retention catches.
- Figma Design Agent Launch Figma shipped an in-product design agent. Worth a look at where agentic tools are heading inside the apps designers already use.
2025 · 11