I Automated 80% of My Daily Development Work with Claude Code
Build an AI Agent system from scratch with Claude Code to automate memory management, literature digestion, and SOP routing.
Grimoire
Long-form analysis at the intersection of AI and life sciences — all spells, sorted by time.
All articles
Your Cells Are Faking Oxygen Starvation: Three Truths About the NAD⁺ Collapse
You took NMN for three months and still look the same in the mirror. Why? Three landmark studies piece together how NAD⁺ collapses with age: how cells fake hypoxia, who eats your supplements, and whether topping NAD⁺ back up can truly rejuvenate old cells.
Blood, Brain Scan, and a Human Trial: NAD⁺ Decline Finally Shows Up in Real People
We've heard plenty of mouse stories about NAD⁺ falling with age. Three studies finally move it onto real humans: a tube of blood, a brain scan, and a randomized double-blind trial each confirm NAD⁺ truly declines in us, while honestly flagging that the clinical payoff is still unsettled.
Put Claude Code in Your Pocket: A Complete Guide to the Telegram Plugin
The Claude Code Telegram plugin extends your desktop AI assistant to mobile devices via MCP. This guide covers four-step installation, secure pairing with allowlist configuration, group and multi-bot setup, so you can remotely operate your desktop Claude session from your phone with full skill and tool permissions.
The Atomic Rules of AI Skill Composition: From Atoms to Compounds
A single AI Agent is reliable. But your cognitive bandwidth caps at 5 parallel threads. Skill Graphs 2.0 introduces a three-layer architecture — atoms (deterministic), molecules (fixed sequence), compounds (strategic autonomy) — turning 5 compound-level tasks into 500 work units. This article breaks down the framework, practical principles, and known limitations.
250 Documents Can Plant a Permanent Backdoor in AI: The Real Threat of Training Data Poisoning
Anthropic, the UK AI Safety Institute, and the Alan Turing Institute have demonstrated that inserting just 250 carefully crafted documents into training data can plant a permanent backdoor in LLMs ranging from 600M to 13B parameters. Bigger models and more data do not dilute the poison.