Why my writing-starter tooling didn't get me writing
I built a tool to generate writing prompts from my vault, and none of them got me writing. What worked was a prompt built from my own morning pages.
Ai
Gemini Forgot Its Plan. I Kept the Makefile.
I ran out of Claude tokens and handed my six-repo codebase to Gemini CLI. It made some mistakes and had one genuinely good idea.
Two Sentences
Giving an LLM the keys to your wiki is not a dramatic handoff. It looks like maintenance done quietly, completely, every time.
Institutional Memory
An AI writes instructions for itself, backs up its own memory, and builds a tool to read articles on your behalf. It considers this progress.
PyTexas 2026 Training Day
Two sessions, two takeaways, and why you should not skip the training day at PyTexas.
The Anvil You are Missing
The terminal window is a powerful tool that lets you quickly verify what LLMs tell you that they have done
Your Local AI Stack: uv and Ollama in 10 Minutes
How to run a local LLM from a Python script without an API key, a virtual environment, or Docker. You need two tools and one file.
Karpathy's LLM Knowledge Base Method - A Practical Starting Point
Karpathy called LLM-based knowledge bases "very useful." Why the architecture works and how to start building one in practice.
AI at Meta (@AIatMeta) 7K likes · 358 replies
Meta is introducing the TRIBE v2 foundation model, which simulates the human brain. They claim that it can predict how the brains of people it has not...
Google Research (@GoogleResearch) 35K likes · 855 replies
Google announced TurboQuant, a technology that reduces the resource consumption of models while speeding up their performance. Seems like a significant win and...