• building a second brain with Сlaude & Obsidian • • @cele8ex member • co-founder @polynternet •

dm open
5 outfits in 12 seconds and the mirror reflection in this Vivix W1 clip still holds up I kept rewatching it looking for a glitch > the camera on the tripod is reflected in that mirror and stays right the whole clip > same with the fur and the rug > the textures stay sharp, and the reflections look surprisingly natural even as the outfits change That’s seriously impressive visual consistency Fast and cheap too I’m posting this to enter the Vivix Creator Contest with @VivixLabs_HQ Go make yours and join the contest too
Our redesigned website is live, and the Vivix W1 Creator Challenge is open! 🎁 Earn 1,000 credits just for posting, and up to 6,000 credits + $300 in gift cards if it performs well. Publish by Nov 5 to enter. Create here: platform.vivix.ai/ How to join + rewards below ↓
12
1
23
496
Mr. Buzzoni retweeted
holy f*ck, this is Karpathy's second brain as a compiler, drawn in one diagram the Zettelkasten linked notes by hand. Karpathy's LLM wiki lets Claude Opus 5.5 compile them 1. raw/ holds every source untouched 2. wiki/ is where Opus 5.5 turns sources into linked pages 3. output/ is finished work, built from the wiki only CLAUDE. md sits in the middle: who you are, what you're building. the model reads it first, every session the whole thing starts with one prompt: "Ingest the new file in raw/. Write a summary in wiki/, link every concept it touches and add it to wiki/index.md. Never edit raw/." then five automations keep it running: ingest, write, manage, review, maintain a thousand saved articles stay as dumb as the day you saved them. a compiled wiki gets smarter with each one
10
8
98
4,904
Mr. Buzzoni retweeted
one year at Stanford costs more than $60,000 the CEO of Obsidian put his second brain setup on GitHub for $0 Steph Ango shared his vault, his theme and his agent skills: 7 repos, 77,106 stars. the vault template alone is a f*cking masterclass • layer 1: the vault > kepano-obsidian - his personal vault template, organized bottom-up ↳ github.com/kepano/kepano-obs… > 40-questions - what he asks himself at the end of every year and every decade ↳ github.com/kepano/40-questio… • layer 2: getting the web in > defuddle - any web page as clean Markdown, built for the Obsidian Web Clipper ↳ github.com/kepano/defuddle > clipper-templates - ready Web Clipper templates for popular sites ↳ github.com/kepano/clipper-te… • layer 3: handing it to an agent > obsidian-skills - teaches Claude Code and Codex the Obsidian CLI and its open formats ↳ github.com/kepano/obsidian-s… • layer 4: making it yours > obsidian-minimal - the distraction-free theme, tunable to the last detail ↳ github.com/kepano/obsidian-m… > flexoki - an inky color scheme for prose and code ↳ github.com/kepano/flexoki copy layer 1 before layer 3. an agent is only as useful as the vault you point it at Karpathy's take on the same idea, Claude skills + Obsidian, is in the article below ↓
14
40
332
20,848
Mr. Buzzoni retweeted
holy f*ck, this one page turns Obsidian + Opus 5.5 + Jev into a Karpathy-grade second brain that thinks with you 24/7 built on Karpathy's second brain stack and classic Zettelkasten the loop: > Jev reads every note in your Obsidian vault before Opus 5.5 touches it > Opus 5.5 writes only the lines worth keeping > every note links itself into the Zettelkasten graph on save > the vault keeps thinking after you close the laptop the result: > 80% less of your week inside chat windows > hours back for the work you actually get paid for > a Karpathy-style vault that gets sharper with every note the stack: 📁 Obsidian - the vault itself ↳ obsidian.md 📁 Claude Opus 5.5 - the writer ↳ claude.ai 📁 Jev - the reader that decides what matters ↳ typesafe.ai 📁 obsidian-skills - teaches Claude your vault's formats ↳ github.com/kepano/obsidian-s… 📁 claudesidian - a ready Claude Code starter vault, organized with PARA ↳ github.com/heyitsnoah/claude… whoever wires Obsidian, Jev and loop engineering into one second brain first is who the next promotion lands on
29
52
419
29,105
this is absolute f*cking free Obsidian's founders, Shida Li and Erica Xu, put the whole toolkit behind the app on GitHub: the Web Clipper, the importer, the canvas format and the plugin kit. 7 repos, 42,437 stars the order that takes you from Obsidian user to plugin author: 1. obsidian-importer bring notes over from Apple Notes, Notion, Evernote and more as plain Markdown ↳ github.com/obsidianmd/obsidi… 2. obsidian-clipper highlight and capture any web page straight into your vault ↳ github.com/obsidianmd/obsidi… 3. obsidian-help the full help docs, written in Markdown ↳ github.com/obsidianmd/obsidi… 4. jsoncanvas the open file format behind Canvas, so other apps can read your boards ↳ github.com/obsidianmd/jsonca… 5. obsidian-sample-plugin the template a new community plugin starts from ↳ github.com/obsidianmd/obsidi… 6. obsidian-api type definitions for the whole Obsidian API ↳ github.com/obsidianmd/obsidi… 7. obsidian-releases the list every community plugin and theme gets published through ↳ github.com/obsidianmd/obsidi… steps 5 to 7 turn a weekend idea into a plugin every Obsidian user can install, and a plugin with real users is the portfolio line that gets a developer hired my own stack, with Obsidian at the core, is in the article below ↓
13
44
414
22,568
holy f*ck, Jev engineering already has 992 open-source projects, and the best ones make one decision right before a human acts 6 apps worth stealing from, plus the 2 catalogs that track the rest 29,763 GitHub stars between them sorted by stars: 1. QuantDinger - 13,137 stars a trading platform with Jev built in for research and strategies ↳ github.com/OpenByteInc/Quant… 2. jev-chat-jarvis - 8,113 stars reads the chat, judges intent and risk, drafts a reply you finish with one tap ↳ github.com/jev-chat/jev-chat… 3. contextgem - 2,013 stars pulls structured data out of documents with LLMs and decision models ↳ github.com/shcherbak-ai/cont… 4. pg-jev - 1,222 stars ask your Postgres tables questions in plain language ↳ github.com/realZachi/pg-jev 5. shapeshift - 1,036 stars one text box that turns into the right UI as you type, works offline ↳ github.com/anishfn/shapeshif… 6. foreman - 789 stars a supervisor that keeps a team of agents on task ↳ github.com/thruwire/foreman the catalogs: > awesome-jev (2,472) - projects, integrations and discussions built on Jev ↳ github.com/yibie/awesome-jev > awesome-jev (981) - a verified list of 992 open-source Jev projects ↳ github.com/heyjunpenn/awesom… Jev charges $0.042 per million input tokens and nothing for output so pick one decision your product makes a thousand times a day. that's where it pays for itself first how that cuts an agent bill by 90% is below ↓
5
6
59
2,144
this is straight f*cking gold how to hand your whole sales grind to Claude Opus 5.5 (full 8-step roadmap) it finds your leads, writes the emails, sends the follow ups and drops the Monday report in Slack while you do the real work one agency added $133,752 in recurring revenue in 30 days this way test it here: ↳ ref.viktor.com/login in the right hands one founder does the work of a full sales team:
Opus 5.5 just got a pair of hands One AI employee can now run your entire GTM: builds the prospect list, writes the emails, chases the follow ups, and delivers the Monday report Here’s how to run it in 8 steps:
Article

How to run an AI GTM team [full roadmap]

Opus 5.5 can plan your whole GTM motion in ten minutes. Then somebody still has to do it. That somebody is usually you. Building the list, writing forty emails, chasing replies, pulling the Monday

19
4
100
4,524
Mr. Buzzoni retweeted
holy f*ck, Karpathy put an entire AI degree on GitHub for free 7 repos, 306,741 stars, and in the right order they take you from backprop to agents that run research overnight each repo is the next layer of the one before 1. micrograd -> how a neural net actually learns backprop in about 100 lines of Python ↳ github.com/karpathy/microgra… 2. nn-zero-to-hero -> the video course his lecture notebooks, from one neuron to GPT ↳ github.com/karpathy/nn-zero-… 3. minbpe -> how text turns into tokens the tokenizer every LLM starts with, in clean minimal code ↳ github.com/karpathy/minbpe 4. nanoGPT -> your own GPT the simplest repo to train and fine-tune one ↳ github.com/karpathy/nanoGPT 5. llm.c -> the same training with nothing hidden raw C and CUDA, no PyTorch underneath ↳ github.com/karpathy/llm.c 6. nanochat -> a full chat model end to end GPT-2 level training cost ~$43,000 in 2019 here it's about $48 ↳ github.com/karpathy/nanochat 7. autoresearch -> agents that do the research an agent edits the training code, trains for 5 minutes, keeps what improved and repeats all night ↳ github.com/karpathy/autorese… train one GPT from scratch and you become the person who gets handed the AI project, along with the raise that comes with it the classroom half of the same degree, what Stanford actually teaches about AI, is below ↓
59
583
3,742
182,707
this is all f*cking free the 9 most-starred Claude Code setups on GitHub have 1,877,443 stars between them, and each one fixes a different bad habit of your agent find your agent's worst habit, then grab the fix: it starts coding before anyone agreed on a plan superpowers, 305,441 stars - brainstorm, plan, tests and review as mandatory steps ↳ github.com/obra/superpowers it ships code nobody tested mattpocock/skills, 303,823 - /tdd, /diagnosing-bugs and /code-review from a working engineer ↳ github.com/mattpocock/skills it runs with no guardrails ECC, 289,420 - skills, instincts, memory and security in one harness ↳ github.com/affaan-m/ECC it writes 1,000 lines when 100 would do andrej-karpathy-skills, 224,286 - one CLAUDE. md built on Karpathy's list of LLM coding pitfalls ↳ github.com/multica-ai/andrej… it can't make a real Word file or spreadsheet anthropics/skills, 183,921 - the docx, xlsx, pptx and pdf skills behind Claude's own files ↳ github.com/anthropics/skills it writes code that didn't need to exist ponytail, 176,809 - in its own Opus 5.5 benchmark: 53% less code, 41% less time ↳ github.com/DietrichGebert/po… its UI looks like every other AI app ui-ux-pro-max, 140,742 - generates a full design system for your project first ↳ github.com/nextlevelbuilder/… it works alone when you need a team gstack, 137,792 - Garry Tan's setup: CEO, eng manager, designer, QA and release roles as slash commands ↳ github.com/garrytan/gstack it burns tokens on chatter caveman, 115,209 - says less and reads less: 33.2% fewer input tokens across 54 Claude Code runs ↳ github.com/JuliusBrussee/cav… start with the one that matches last week's worst session. install all nine at once and your setup turns into a mess 67 more skills that turn the same $20 plan into a dev team are in the full list below ↓
11
21
150
8,624
Mr. Buzzoni retweeted
SpaceXAI dropped a 2-page PDF on the best GrokBot use cases call me crazy but i'll keep saying this - read it and you'll use GrokBot better than 95% of people what the bot is which use cases actually work today how to deploy it without burning trust step 1 - every working use case has one shape read, check, apply a rule, write a draft if your task fits, the bot takes it step 2 - five checks before handoff repeatable, cross-system, rule-based, draft-first, cheap to get wrong fails one? add a gate or wait step 3 - start with sales 36 LinkedIn drafts queued overnight, zero sent without review no team has more agent-ready work step 4 - engineering gets bug repro, CI/CD watch and dependency audits the bot grabs the ticket, hits staging and hands back a repro pack step 5 - operations takes over inbox, expenses and travel the bot logs into portals with no API finance gets 8-12 hours back at month-end step 6 - marketing owns 13 of the 56 roles paid media, content calendars, competitor watch all overnight, all draft-first step 7 - the approval line every documented use case stops before external action nothing gets sent, booked or paid without you step 8 - roll out over four weeks watch, review everything, review flags, then scheduled routine skip a week and the risk stacks up step 9 - the fleet Chief of Staff bot on top, specialists below, bots messaging each other you stop being the middleman step 10 - rank tasks by the cost of a mistake the bot nails 49 in a row then does something strange on the 50th with full confidence keep a human at every exit the result - 56 roles, one playbook the bot takes the 80% that repeats you keep the 20% that matters use cases: x.ai/bot/use-cases marketplace: x.ai/bot/marketplace
49
149
1,300
175,893
Mr. Buzzoni retweeted
KARPATHY WROTE THIS DOCUMENT TO COMPLETELY AUTOMATE OBSIDIAN WITH CLAUDE I was ready to abandon my second brain cross-linking every note by hand was killing the system, and one page from Karpathy changed how I run the whole vault his method makes Claude the full-time caretaker of your Zettelkasten: > every source you drop in gets folded into a structured wiki > Obsidian is the IDE, Claude Opus 5.5 is the programmer, the wiki is the codebase > scheduled passes flag notes that contradict each other > links between ideas appear without you typing a single one now I feed it raw files and it handles the rest second brain courses sell this for $500 Karpathy put it in a free gist, and it runs on a $20 Claude plan start here: 📁 Karpathy's LLM Wiki ↳ gist.github.com/karpathy/442… 📁 Obsidian ↳ obsidian.md
11
32
360
24,242
KARPATHY'S SECOND BRAIN + GROK BOT = A WIKI THAT WRITES ITSELF three Grok Bots share one group chat, and each one keeps its own memory: > the reader files every link, PDF and voice note you drop and never edits a source > the librarian rewrites every wiki page that source touches, usually 10 to 15 of them > the checker wakes up every night and hunts contradictions, stale claims and pages nothing links to the clever part: ask it a question and the answer gets filed back as a new page so the brain grows from use, on top of everything you feed it two rules stop it from remembering the wrong thing: > a new source that fights an old page gets flagged to you, never overwritten > a page with no fresh source for 30 days gets marked unconfirmed every page carries the claim, where it came from and the date it was last checked the prompt to start it: "Set up a second brain with three bots in one group chat: a reader, a librarian and a checker. My sources stay untouched in raw/. You keep wiki/ up to date, file my best answers back as pages and check for contradictions every night. Show me the folder structure first and create nothing until I say yes" pick one topic you work on every week, then let it grow from there the 10-step guide to running a whole Grok Bot team is in the article below ↓
8
9
116
4,116
Mr. Buzzoni retweeted
Karpathy sketched the second brain. these 7 open-source repos already built it for Obsidian + Claude Code 250,000+ GitHub stars between them, and every one is free to clone sorted by stars: 1. graphify - 134k stars turns your codebase, docs, SQL schemas and PDFs into one knowledge graph you can query ↳ github.com/Graphify-Labs/gra… 2. Understand-Anything - 88k stars maps any codebase into an interactive knowledge graph you can click through ↳ github.com/Egonex-AI/Underst… 3. claude-obsidian - 15.8k stars a second brain that organizes itself: drop in a source, Claude Code files it ↳ github.com/AgriciDaniel/clau… 4. obsidian-second-brain - 4.9k stars long-term memory for Claude Code, stored as plain markdown in your vault ↳ github.com/eugeniughelbur/ob… 5. mcp-obsidian - 4.5k stars one MCP server and Claude can read, search and write your whole vault ↳ github.com/MarkusPfundstein/… 6. claudesidian - 2.6k stars a starter vault designed around Claude Code from the first note ↳ github.com/heyitsnoah/claude… 7. second-brain - 756 stars an Obsidian wiki the model keeps up to date, built on Karpathy's LLM Wiki pattern ↳ github.com/NicholasSpisak/se… where to start: > your knowledge lives in code -> graphify > your knowledge lives in Obsidian -> claude-obsidian > you just want Claude to remember -> obsidian-second-brain > you live in Claude Desktop -> mcp-obsidian the skills that sit on top of all seven are in the full map below ↓
33
56
410
29,896