Andrej Karpathy’s “LLM Wiki” vision just escaped the whiteboard and turned into a real desktop app.
Meet Tolaria.
A native knowledge workspace where humans + AI agents work together inside plain markdown files.
No cloud lock-in.
No weird proprietary format.
No accounts.
Just files you actually own.
But the craziest part is how it was built:
→ 100,000+ lines of code
→ Tauri + React + Rust stack
→ 3,000+ tests with 85% coverage
→ 70+ architecture decision records
→ 9.9/10 code health score
And every vault is a Git repo with built-in visual history.
It even ships with an MCP server out of the box, so Claude Code can directly read + edit your knowledge base like it’s a second brain.
This feels less like “note-taking software” and more like the blueprint for AI-native operating systems.
Open source. Free forever.
Insane work by Luca Rossi.
Repo👇
Chez Spotify, 96 % des ingénieurs codent désormais avec Claude Code.
La fréquence des PR a explosé de +60 % et ils réalisent 4 500 déploiements par jour.
Niklas Gustavsson, Chief Architect de Spotify, vient d'expliquer exactement comment ils ont fait ça, lors d'une intervention de 27 minutes sur la scène d'Anthropic:
Gardez-la précieusement en Signet 🔖
« Plus de 99 % de nos ingénieurs utilisent des outils de codage basés sur l'IA. L'adoption a décollé après la sortie d'Opus 4.5. »
Ça vaut plus que tous les cours à 600 $ que t'as failli acheter.
( Il n'y a pas encore de transcription officielle qui existe pour cette vidéo, elle a été transcrite par moi.)
El futuro de la agentic AI ya no es texto… es un centro de mando.
Ya no más terminales caóticas, logs infinitos ni tener que adivinar qué está haciendo cada agente.
[Hermes War Room] es el dashboard visual definitivo para orquestar tu flota de agentes Hermes:
- Chat directo con el orquestador
- Kanban en tiempo real (Todo → Ready → Running → Blocked)
- Sala de Operativos con avatares, callsigns y burbujas de pensamiento
- Flechas animadas de delegación entre agentes
- Contrata, reentrena o despide agentes al instante
Todo en vivo.
Todo estratégico.
Todo cinematográfico.
REPOOO👇
TENCENT ACABA DE DROPEAR LA BOMBA para todos los que hacen AI Agents:
Un sandbox que:
- Arranca en menos de 60 ms (hasta 50x más rápido)
- Usa solo 5 MB de RAM por instancia
- Puedes correr +2.000 sandboxes en un solo servidor
- Seguridad de verdad (microVMs con KVM + RustVMM)
- y 100% compatible con E2B SDK.
Self-hosted, open-source y GRATIS.
REPOOO👇
This is the most complete Claude Code setup that exists right now.
27 agents. 64 skills. 33 commands. All open source.
The Anthropic hackathon winner open-sourced his entire system, refined over 10 months of building real products.
What's inside:
→ 27 agents (plan, review, fix builds, security audits)
→ 64 skills (TDD, token optimization, memory persistence)
→ 33 commands (/plan, /tdd, /security-scan, /refactor-clean)
→ AgentShield: 1,282 security tests, 98% coverage
60% documented cost reduction.
Works on Claude Code, Cursor, OpenCode, Codex CLI. 100% open source.
Pour moins de 2 euros, ce petit composant USB-C me dit quand mes agents IA travaillent et quand ils attendent une réponse de ma part.
Je l’ai mis en place avec Claude, ça fonctionne même avec plusieurs terminaux (en local ouremote), et ça ouvre des possibilités assez folles pour presque rien.
Si ça vous intéresse, demandez-moi le fichier Blinky.MD : il permet à l’IA de faire le setup automatiquement.
This is the only article you’ll need to build your own MCP server. It covers:
> need of MCP
> how MCP works
> core concepts along with visualization
> build your own mcp server
> theory + coding
🚨BREAKING: Someone built an entire AI trading firm and open-sourced it.
A full team of AI agents working together exactly like a real Wall Street firm.
It's called TradingAgents. Here's how it works:
The Analyst Team reads the market:
→ Fundamentals Analyst — reads company financials, finds red flags
→ Sentiment Analyst — scans social media for market mood
→ News Analyst — monitors global news and macro events
→ Technical Analyst — reads MACD, RSI, price patterns
Then the Researchers debate:
→ A bull agent argues for the trade
→ A bear agent argues against it
→ They go back and forth until one side wins
Then the Trader decides.
Then Risk Management approves or kills it.
Then the Portfolio Manager signs off.
Every trade goes through 6 layers of AI review before it executes.
Works with Claude, GPT, Gemini, Grok, and local models via Ollama.
Three lines to run your first analysis:
ta = TradingAgentsGraph(config=DEFAULT_CONFIG.copy())
_, decision = ta.propagate("NVDA", "2026-01-15")
print(decision)
29.9K stars. Backed by a published research paper.
Free & open source.
link in comment
🚨Breaking: ByteDance released an AI employee that works 24/7 — fully open-source.
It researches. Codes. Builds websites. Creates slides. Generates videos.
And it does everything on its own computer.
Meet DeerFlow 2.0 🦌
This isn't a chatbot.
It's an autonomous AI system that plans, delegates, executes, and delivers.
You give one prompt → it spins up sub-agents → executes tasks in parallel → returns a finished output.
No babysitting. No step-by-step prompting.
What makes it insane:
→ Creates a team of AI sub-agents automatically
→ Runs code, debugs errors, retries until it works
→ Builds full websites, reports, dashboards & slides
→ Uses memory to learn your style & workflow
→ Reads files + works inside its own sandbox
→ Searches web, calls tools, executes commands
How it works: Lead agent → creates plan
Sub-agents → work in parallel
Results → merged + synthesized
Output → fully finished deliverable
Not notes. Not drafts.
Actual production-ready output.
Wild part:
DeerFlow 2.0 launched Feb 28, 2026
Hit #1 on GitHub Trending the same day
Complete rewrite from v1
Built by ByteDance
MIT Licensed — fully open source
22.7K stars already. And climbing fast.
We're entering the era of AI coworkers, not AI tools.
Link in comments👇