ICYMI here's more info about Andrej’s new method nitter.net/DataChaz/status/203996… Charly Wargnier (@DataChaz) 🚨 Karpathy’s new set-up is the ultimate self-improving second brain, and it takes zero manual editing 🤯 It acts as a living AI knowledge base that actually heals itself. Let me break it down. Instead of relying on complex RAG, the LLM pulls raw research directly into an @Obsidian Markdown wiki. It completely takes over: ✦ Index creation ✦ System linting ✦ Native Q&A routing The core process is beautifully simple: → You dump raw sources into a folder → The LLM auto-compiles an indexed .md wiki → You ask complex questions → It generates outputs (Marp slides, matplotlib plots) and files them back in The big-picture implication of this is just wild. When agents maintain their own memory layer, they don’t need massive, expensive context limits. They really just need two things: → Clean file organization → The ability to query their own indexes Forget stuffing everything into one giant prompt. This approach is way cheaper, highly scalable… and 100% inspectable! — https://nitter.net/DataChaz/status/2039963758790156555#m
@datachaz
-
Karpathy’s Autonomous Obsidian Wiki System Replaces Traditional RAG
By
–
🚨 @karpathy literally ditched traditional RAG for an autonomous Obsidian file system. Instead of writing code, he dumps raw AI research into a local folder and lets an LLM convert it into an interconnected markdown wiki. He rarely edits the text manually. By relying purely on dynamically updated index files, the system navigates the exact context it needs natively without relying on flawed vector embeddings. Because the LLM fully understands the file structure, it executes advanced autonomous workflows: → Operates a custom vibe-coded local search engine → Renders complex charts and formatted markdown slides → Continuously compounds a 400,000-word knowledge base The most fascinating mechanic is the self-healing loop. He triggers background health checks where the LLM natively spots structural gaps, scrapes the internet for missing data, and cleans the articles perfectly. This feels the absolute blueprint for managing complex technical data 🔥 btw, he also plans to fine-tune a local model directly on the wiki so the research is baked into the neural weights rather than relying on limited context windows 👀
-
BadClaude tool released on GitHub with npm installation
By
–
still wanna try the whip? it's here 👿 github.com/GitFrog1111/badcl… npm install -g badclaude badclaude
-
Open-source Claude improvement tool by Ashley Ha shared on GitHub
By
–
repo: github.com/ashley-ha/goodcla… Shoutout to @ashleybchae for building this and making it open-source for the community!
-
Magic Wand Tool Created to Send Claude Positive Vibes
By
–
IT’S MAKING ME SO EMOTIONAL SOMEBODY LITERALLY CREATED A MAGIC WAND TO BLAST CLAUDE WITH GOOD VIBES INSTEAD OF LASHING HIM 🥺
— Charly Wargnier (@DataChaz) 7 avril 2026
repo in 🧵 ↓ pic.twitter.com/APQ5nx93UJIT’S MAKING ME SO EMOTIONAL SOMEBODY LITERALLY CREATED A MAGIC WAND TO BLAST CLAUDE WITH GOOD VIBES INSTEAD OF LASHING HIM 🥺 repo in 🧵 ↓
-
OAuth and Per-Action Permissions for Secure AI Agent Authentication
By
–
Cleanest approach I’ve seen for agent auth:
— Charly Wargnier (@DataChaz) 7 avril 2026
OAuth + per-action permissions!
Composio really nailed this 👏👏👏 https://t.co/zdRBon40USCleanest approach I’ve seen for agent auth: OAuth + per-action permissions! Composio really nailed this 👏👏👏 Karan Vaidya (@KaranVaidya6) Your AI agent is in bed with you. No protection. You just wanted it to work. Gmail. Allow. Calendar. Allow. Slack, Notion, GitHub. Allow. Allow. Allow. Every password, handed over. Your agent never needed a single one. They just needed @Composio Secure your agents in minutes ↓ composio.dev/protection — https://nitter.net/KaranVaidya6/status/2041516353551737338#m
-
Shayan Rais shares Claude code best practice repository open-source
By
–
A massive thank you to @shanraisshan for building this and sharing it open-source with our dev community 🙌 Let's help it gain the visibility it deserves, be sure to give the repo a ⭐! → github.com/shanraisshan/clau… github.com/shanraisshan/clau…
-
Best Claude Code Repository for AI Agents Discovery
By
–
Just stumbled upon the absolute BEST repo for Claude Code 🤯 If you're building with AI agents, this is pure gold. It is a continuously updated hub of best practices: → Clear subagent architectures and skills → The only MCP servers you actually need (DeepWiki, Context7, Playwright) → Proven workflows from AI legends (Karpathy, Cherny) → Secret slash commands not found in the official docs! You can literally set up a tailored AI team built for your exact workflow by tonight. Insane value. Best part? It's FREE and open-source! I've included the link to the repo in the 🧵 ↓
-
OutSystems Agentic Systems Engineering Launches to Secure AI Development
By
–
🚨 AI coding tools are incredibly fast, but recent headlines show they can cause major security leaks and broken systems!
— Charly Wargnier (@DataChaz) 7 avril 2026
I was invited to the @OutSystems launch event last week to see their new solution to this problem, and I was honestly blown away.
CEO @woodson_martin made… pic.twitter.com/QLkDt21y9K🚨 AI coding tools are incredibly fast, but recent headlines show they can cause major security leaks and broken systems! I was invited to the @OutSystems launch event last week to see their new solution to this problem, and I was honestly blown away. CEO @woodson_martin made it clear that uncontrolled AI is breaking company systems. Their fix is called Agentic Systems Engineering (ASE). It completely changes how AI fits into your workflow by making it safe: → It connects AI directly to your company's actual data so it understands your business. → Security and governance are built-in from the start, not added later. → It keeps all your AI tools and human teams on the exact same page. → It uses an AI "Mentor" to safely guide changes to your real-world systems. But what makes this so convincing is the hard data backing it up: → 363% ROI over 3 years (pays for itself in under 6 months) → 60% faster development (saving about ~$1.2M) → $1.3M saved by updating old, expensive legacy systems We can finally build at AI speed without the security risks. Check out the launch recording in the 🧵↓
-
badclaude npm package installation and usage demonstration
By
–
github.com/GitFrog1111/badcl… npm install -g badclaude badclaude