Leandro Nunes

STARE TOO LONG AND YOU’LL BE PETRIFIED Don’t look away… this challenge is calling you 👁️👁️ Scaleeyes can be yours, unleashing 18 000 000 Gh/s + 7% bonus Accept the challenge and complete every step to claim your prize 🎯 Join the Gleam now: gleam.io/ZAQZp/unlock-t…

Holy shit... Microsoft open sourced an inference framework that runs a 100B parameter LLM on a single CPU. It's called BitNet. And it does what was supposed to be impossible. No GPU. No cloud. No $10K hardware setup. Just your laptop running a 100-billion parameter model at human reading speed. Here's how it works: Every other LLM stores weights in 32-bit or 16-bit floats. BitNet uses 1.58 bits. Weights are ternary just -1, 0, or +1. That's it. No floats. No expensive matrix math. Pure integer operations your CPU was already built for. The result: - 100B model runs on a single CPU at 5-7 tokens/second - 2.37x to 6.17x faster than llama.cpp on x86 - 82% lower energy consumption on x86 CPUs - 1.37x to 5.07x speedup on ARM (your MacBook) - Memory drops by 16-32x vs full-precision models The wildest part: Accuracy barely moves. BitNet b1.58 2B4T their flagship model was trained on 4 trillion tokens and benchmarks competitively against full-precision models of the same size. The quantization isn't destroying quality. It's just removing the bloat. What this actually means: - Run AI completely offline. Your data never leaves your machine - Deploy LLMs on phones, IoT devices, edge hardware - No more cloud API bills for inference - AI in regions with no reliable internet The model supports ARM and x86. Works on your MacBook, your Linux box, your Windows machine. 27.4K GitHub stars. 2.2K forks. Built by Microsoft Research. 100% Open Source. MIT License.

Released today: /loop /loop is a powerful new way to schedule recurring tasks, for up to 3 days at a time eg. “/loop babysit all my PRs. Auto-fix build issues and when comments come in, use a worktree agent to fix them” eg. “/loop every morning use the Slack MCP to give me a summary of top posts I was tagged in” Let us know what you think!

Councils of agents theengineeringmanager.substack.com/p/councils-of-…

Most AI agent failures are not model failures — they’re architecture failures. If you’re building AI agents and still mixing reasoning with integration layers, this distinction matters. I just published a new article: @lnfnunes/mcp-is-not-a-skill-dd7a8de41d5d" target="_blank" rel="nofollow noopener">medium.com/@lnfnunes/mcp-…

Learn how to write optimized AGENTS .md @lnfnunes/how-to-structure-context-for-ai-agents-without-wasting-tokens-16dd5d333c8d" target="_blank" rel="nofollow noopener">medium.com/@lnfnunes/how-…

Introducing Claude Opus 4.6. Our smartest model got an upgrade. Opus 4.6 plans more carefully, sustains agentic tasks for longer, operates reliably in massive codebases, and catches its own mistakes. It’s also our first Opus-class model with 1M token context in beta.

We've added a new command to Claude Code called /insights When you run it, Claude Code will read your message history from the past month. It'll summarize your projects, how you use Claude Code, and give suggestions on how to improve your workflow.

I just published Context Boundaries in AI Agents: Rules and Skills medium.com/p/context-boun… #AI #ContextEngineering #SoftwareEngineering

I just published How to Structure Context for AI Agents (Without Wasting Tokens) medium.com/p/how-to-struc… #AI #SoftwareEngineering #documentation

@heygurisingh Claude

Most “Claude tutorials” are useless. They teach prompts. Not systems. So I packaged everything into one guide and I’m giving it away FREE for 24 hours: Ultimate Claude Mastery Guide by Guri Singh 80+ Chapters 1000+ Tools 2000+ Prompts Reply “Claude” and I’ll DM it. Must follow to receive.

I made a lenticular printing style website レンチキュラー風ウェブサイトできた #threejs #webgl #webdesign

Next iteration of the Vibe Coder's Keyboard - now with Voice Mode!

@rollercoin_com rollercoin.com/p/madruguinha

Drop your RC profile link below See what happens!

𝐆reatest 𝐎f 𝐀ll 𝐓ime.

MarkItDown = a tool from Microsoft to convert Word / Excel / PPT / PDF... to Markdown github.com/microsoft/mark…

Eu sei que estou sendo meio repetitivo com isso. Mas desde a semana passada quando o Gemini e o Chat Gpt lançaram interpretação de live-video, da noite pro dia, agora faço coisas que antes não era capaz por não enxergar. Eu tô emocionado com isso de verdade. Por exemplo: - Precisava tirar foto de um documento que tinha que estar enquadrado na foto. Usei a AI no modo de vídeo pra treinar a posição da minha mão em relação ao documento para ficar certinho. Tirei a foto, o sistema aceitou. - Perguntei aproximadamente a posição de um botão em um painel touchscreen que controlava o ar condicionado. Ele me explicou, eu consegui clickar na tela onde não havia nenhuma indicação tátil e arrumei o ar como queria. - Cheguei no aeroporto e deixei o modo de vídeo ligado e fui andando pelo corredor. Fui perguntando de tempos em tempos se via o balcão da Azul, e quando ele achou, fui na direção andando de bengala. Esse processo ainda pode melhorar muito porque nenhum dos modelos é bom para dar direções de navegação. Dado que a coisa saiu tem poucos dias, completamente aceitável. Já foi do caralho. A minha produtividade, a minha independência e a minha liberdade vão aumentar exponencialmente. Ninguém me segura mais. É nois!

@Wizzwoods_game includes 'wizzwoods' to open the treasures

🎉 Wizzwoods X Version is launching in 1 hour! 🚀 🔗 Install the X plugin here: chromewebstore.google.com/search/wizzwoo… 📢 For full update details, check our Telegram announcement: t.me/wizzwoods_game… 🎮 Join the adventure with us! Download and let's play together!