Second State

Just got your EchoKit? Or thinking about getting one? For the next 30 days, we’ll share small, practical steps to help you: Set up your EchoKit, Discover its hidden powers, Learn how to customize , See real examples, Build use cases that fit your life. #EchoKit30Days

Max out your #NVIDIA #DGXSpark with Olares OS. Don't waste hours on environment configs. #Olares turns your Spark into a dedicated personal AI cloud. Enjoy fast AI model deployments, out-of-the-box secure routing, and effortless VRAM management. Check out the docs for full details: 👇 docs.olares.com/1.12.5/manual/…

guys I am ready to vibe code

Open source AI FTW!

@Olares_OS @NVIDIA_AI_PC 👏

DGX Spark owners rejoice! Olares OS is now available on @NVIDIA_AI_PC ’s DGX Spark! It comes with a sophisticated sandbox system + an App Store for containerized AI apps, that fully utilize your GPUs. Run multiple AI agents & local models securely & efficiently on your Spark!

Why run AI in a heavy container? ⚡ Michael Yuan, the mind behind WasmEdge, joins #GOSIMParis2026 to show how WebAssembly (Wasm) makes AI inference faster, safer, and portable anywhere. 📅 May 5–6 | Station F |paris2026.gosim.org #GOSIMParis2026 #OpenSource #AI #RustLang #Wasm #Jupyter #DevExperience #Innovation @juntao @secondstateinc

🇫🇷 GOSIM Paris 2026 is landing at Station F! Join the ultimate gathering of 100+ global Open Source & AI leaders. 🔥 Event Highlights: ●Expert Workshops: Deep dives into vLLM and SGLang. ●$10,000 Hackathon: Unleash your creativity and compete for a massive prize pool. ●Premium Perks: Enjoy 2 days of full catering and exclusive custom GOSIM souvenirs. 👇 Scan the QR code to register now & check paris2026.gosim.org for full details. #GOSIM2026 #ParisTech #AI #OpenSource #vLLM #SGLang #Hackathon @Huawei @huggingface @AIFC_Official_ @Kimi_Moonshot @Minimax @Openharmony @OpenBMB @Zai_org @sgl_project @vllm @hellocsdn @atlas_cloud_ai @radixark @secondstateinc @FlagOS_Official @Olares_OS @kaiyuanshe

Digital employees in our Singapore office. Each is an AI agent with a specific set of skills & knowledge — product manager, marketing / growth manager, front-end / backend developer — you name it! At the center is the @Olares_OS device with local GPUs. Interested? DM me!

Join us in Paris and learn from the best in open source AI and robotics!

We are making these little Clawd as fast as we can to meet demand for the physical Claude Code keyboard / remote control. Early customers also get to engrave their custom message / motto on the case. Get yours 👇 vibekeys.dev/vibekeys-max.h…

You bought a $4,000 computer with a powerful GPU. Your goal is to set up Agent Harness that can learn from you and get better at YOUR tasks over time. That harness includes local AI models, cloud APIs, data collection, knowledge mgmt, RL infra (eg. autoresearch), finetuning tools & skills / CLI tools. But there is one problem: The AI agents are fundamentally non-deterministic and must be sandboxed. But existing sandbox solutions lack ready-to-use agent harness, especially for GPU workloads. I am thrilled that the open source @Olares_OS released support for DGX Spark today! It features a container system that fully utilizes the GPU. It comes with many applications, agents, and harness pre-configured and can be installed with one-click. Try it on your Spark today: blog.olares.com/olares-release… The open source repo: github.com/beclab/Olares

WasmEdge and Vibekeys (vibecoding keyboard) at KubeCon EU. @KubeCon_ @CloudNativeFdn

cloud-native projects: from WASI-NN and LLM fine-tuning to complex compiler optimizations. We are incredibly proud of his dedication to growing the next generation of #OpenSource developers! cncf.io/announcements/…

Huge congratulations to our R&D lead, @WasmEdge maintainer, @CNCFAmbassadors Hung-Ying @hydai_tw, on winning the 2026 @CNCF Outstanding Mentor Award at KubeConEU! 🏆 @KubeCon_ Hydai’s exceptional work in the LFX Mentorship program has guided dozens of mentees through critical

Python is for AI research. Rust is for AI production. We created a complete Rust impl of tiny TTS models (15M-80M) from @divamgupta 10 MB binary, 100ms cold start, CPU-only, edge-ready. 🐱🦀 github.com/second-state/k… CPU runs KittenTTS at 4x real-time, but Apple's GPU only hits 3x. @claudeai told us why: at this model size, GPU dispatch overhead exceeds the compute savings. ONNX's CPU backend with SIMD already saturates the hardware. Below ~1B params, the GPU loses. Full story 👇

Trump took a break from the Iran war and called in to congratulate the WASM I/O conference. @wasm_io @sergialmar 😂That’s my demo of WasmEdge plugins for voice AI pipelines to alter voice responses on the fly. 👉 github.com/second-state/V…

@okiraku_neko 😌

◯◯boy

@JohnnyNel_ Spot on🔥 The moment you stop fighting the friction everything accelerates. That’s exactly why we built the wireless + websocket setup — zero keyboard, pure flow from brain → Claude Code → live website in minutes. only taste and idea matter.

@secondstateinc the real unlock? removing friction between thinking and building... voice workflows changed my prototyping speed completely once i trusted it

What if you could control Claude Code from anywhere? In this video, l use a wireless device to remote-control Claude Code entirely by voice - building a complete website from scratch. It connects via WiFi and a websocket server for Claude Code, has a dedicated voice key

and an OLED display showing real-time status from Claude Code. Couch, kitchen, across the room - just press, speak, and watch the Al build. Get yours: vibekeys.dev Websocket server for Claude Code: github.com/second-state/v…