Wink Pings

The Claude Second Brain Elon Musk Reposted—and 17 Features Most People Have Never Used

Developer Woody built a second brain using Claude Code and Obsidian, and Elon Musk reposted it. Anatoli Kopadze’s list of 17 Claude features then reached a much wider audience. This article highlights the key points: which features are worth enabling first, which ones developers need, and one objection that deserves serious consideration.

Elon Musk reposted developer Woody’s “second brain” setup. It runs on Claude Code and Obsidian.

The workflow isn’t long:

  • Install Obsidian and create a vault
  • Open Claude Code inside the vault
  • Have it create a personal wiki, with a CLAUDE.md at the root
  • Drop source material into /raw and say “ingest this”
  • Claude writes summary pages, traverses the entire vault, finds related notes, and adds backlinks
  • Run “lint the wiki” once a week to clean up dead links, outdated content, and orphan pages

Woody says that now every question he asks points back to one of his own pages.

This setup is very handy. But Claude can do more than this.

In May, Anatoli Kopadze published an article on X titled “Claude Can Do All of This. Most People Have No Idea.” The article put together a list of 17 Claude features, and the tweet announcing it has accumulated 21 million views. Here are a few worth expanding on.

Memory: Projects and Memory

Every new Claude conversation starts as a blank slate. Projects keeps documents, instructions, and frequently used context permanently in one project, so next time you open it you can pick up where you left off. Memory is off by default; once enabled, Claude builds up knowledge about you across conversations, so you don’t have to introduce yourself again in a new chat. These two features are the most worth enabling first.

Build usable things inside a chat

Artifacts can generate genuinely working applications in the conversation sidebar: calculators, habit trackers, dashboards, SVG graphics, and Mermaid diagrams. Available on the free plan. Most people’s impression of Claude stops at text and they’ve never tried asking it to make something clickable.

Role templates: getting Claude to push back

By default, Claude goes along with what you say. Role templates are one of the few ways to get it to push back. The article provides several templates you can copy directly:

  • Cognitive behavioral therapist: asks questions first, doesn’t rush to give advice, and points out your cognitive distortions
  • No-nonsense mentor: tears apart your plan, finds holes, and says the things you don’t want to hear
  • Devil’s advocate: when you’ve already made up your mind, constructs the strongest counterarguments and lists the most realistic ways you could fail

The original article also includes a personal trainer, simulated difficult conversations, and extended thinking, covering more specific scenarios.

Scheduled tasks and desktop access

Scheduled Tasks run automatically once set up. The example in the original article: every morning at 7:30, fetch AI and crypto news from the past 24 hours, pick the five most important items, write summaries and impact assessments, and save them as brief-date.md, readable in under three minutes. Cowork is a desktop app that lets Claude read and write files on your computer directly, without copying and pasting them into the chat box.

For developers

CLAUDE.md goes in the project root, and Claude reads it automatically at every session. Coding standards, glossaries, and tone rules take effect forever after you write them once. Claude Code reads the codebase, writes code, runs tests, reads errors, and fixes bugs in the terminal, looping until the task is complete. Prompt Caching caches repeated context, saving up to 90% on API call costs, with a 5-minute cache window.

The list also includes the Claude in Chrome browser extension, Claude Design visual design tool, and the Skills plugin mechanism, which are suited to specific scenarios.

The full list is in the original article: Anatoli Kopadze’s article

This second brain setup is also controversial. Under Woody’s tweet, someone warned: don’t put Claude on top of notes you wrote yourself; a month later, you won’t be able to tell which notes are your own thoughts and which ones it guessed. This warning deserves serious consideration. Whether to accept the trade-off or reject it depends on whether you care more about efficiency or about your own ideas in your notes.

Don’t enable all the features at once. Start with Projects or Memory—you can get that done in fifteen minutes. Save the rest for when you need them.

发布时间: 2026-11-22 04:16