
AI 101 What's a CLAUDE.md file? 📋 It's a short note in your project folder that Claude Code reads at the start of every session. Write your rules once, and every new session starts with them. Keep it under 200 lines. michellelinmd.com/blog/what-…

UCSF Prof of Emerg Med | Founder @aliemteam | Author, Tricks of Trade in EM Building in public: An AI-native OS for academia & leadership.

AI 101 What's a CLAUDE.md file? 📋 It's a short note in your project folder that Claude Code reads at the start of every session. Write your rules once, and every new session starts with them. Keep it under 200 lines. michellelinmd.com/blog/what-…

AI 201 Build a PubMed literature monitor for your topic 📚 Claude Code searches PubMed for new papers, grades each one read now, skim, or skip, and sorts the keepers into themes with a short summary. Mine sorted 300 papers into 11 themes. michellelinmd.com/blog/pubme…

AI 101 5 things to do when setting up Claude Code 🛠️ AI chat is mostly a brain with eyes. Claude Code is an AI agent that adds arms. Before you let it do things for you, pick a plan, set your privacy, learn the icons, and add Git and GitHub. michellelinmd.com/blog/set-u…

AI 101 What are Git and GitHub? 🗂️ Git is a paper trail for your files. Each commit saves a version with a note on what changed. GitHub is like a private Dropbox for your files and that paper trail. It updates when you push. michellelinmd.com/blog/git-v…

AI 101 Context window: the AI's working memory within one conversation 🧠 Your messages, its replies, and every file it reads fill it up. Too full, and earlier instructions slip. When a session gets long, I start fresh from a summary. michellelinmd.com/blog/what-…

AI 101 What's a .md file? 📝 It's Markdown: plain text with a few symbols for structure. # makes a heading. ** makes bold. AI writes in it and reads it easily. It's how I hand notes from one AI session to the next. michellelinmd.com/blog/what-…

AI 101 AI doesn't read words. It reads tokens 🧩 A token is a small chunk of text, often part of a word. "Pneumonia" is 1 token. "Bronchiolitis" is 4. Tokens set how much the AI can hold at once, and what API use costs. michellelinmd.com/blog/what-…

AI 201 Have a second AI check the first 🔍 I have Codex, OpenAI's version of Claude Code, review what Claude Code builds: as an editor, an adversarial reviewer, and a skeptic. It finds holes the first AI missed. I still make the final call. michellelinmd.com/blog/secon…

AI 101: A .env file is where my app passwords and keys live. 🗝️ Claude Code only needs each password's name. The real password stays in a file on my computer, so it never has to show up in the code or our chat. michellelinmd.com/blog/what-…

Claude Code put my WordPress app password right in our chat. 😬 I never shared it with anyone. It still counts as a leak. And changing my login WordPress password wouldn't have fixed it. michellelinmd.com/blog/claud…

AI 101: an app password is a second password to your account. 🔑 It's how Claude Code builds my WordPress courses in 10 minutes instead of 2 hours. It also works as me, so I treat it like my real password. michellelinmd.com/blog/what-…

AI 101: What is an API? 🔌 It's a way for a program to use an online tool without opening its website. Example: Every week, a program on my Mac searches PubMed for new papers mentioning #PECARN through its API. I never type a search. michellelinmd.com/blog/what-…

AI 101: Before your first Claude Code session, learn 3 Mac skills: a screenshot straight to your clipboard, a hotkey for dictation, and copying a file's path. Your Mac already has all three built in. michellelinmd.com/blog/3-mac…

🚨 New PECARN Study Alert! 🚨 The PECARN STArT trial asked an important question: Could IV arginine help children and young adults with sickle cell disease recover faster from acute pain episodes? jamanetwork.com/journals/jam… #SPARC #PECARN #SickleCellDisease

Intern year is coming. 👀 ALiEM's Bridge to EM is designed to make your transition from medical student to EM resident smooth. Check it out at: emra.org/aliem #BridgeToEM #ALiEM #InternYear #EM #MedEd #EMBound

🚨 New @NEJM landmark trial from @PECARNteam: PRoMPT BOLUS randomized 9,000+ children with suspected septic shock to balanced crystalloid vs 0.9% saline. No difference in kidney injury, mortality, or safety. Either is 👍. 📄 doi.org/10.1056/NEJMoa260196… 🩺 aliem.com/prompt-bolus-pedia…