Allegedly trashing Airbnbs to test robots puts startup in legal trouble
Ars Technica AI1d agoIncidentA startup is facing legal trouble over allegations that robot testing damaged an Airbnb property. The lawsuit seeks $12,000 from the company, according to the provided article summary. The available excerpt does not identify the startup, describe the robot, detail the alleged damage, or state whether a court has ruled on the claim.
COMPUTEX 2026: AI Moves from the Cloud into the Real World
INSIDE 硬塞 AI1d agoHardwareUnder the theme “AI Together,” COMPUTEX 2026 brings together 1,500 exhibitors across the global AI supply chain. The event focuses on AI computing, robotics, and other applications that move AI beyond cloud services into the physical world. Rather than highlighting one model or product launch, the article frames Taiwan as a key hub in the broader industrial transformation driven by AI.
Beyond Family Privilege: Spencer Huang's Path from Bartending to NVIDIA Robotics
INSIDE 硬塞 AI1d agoBusinessSpencer Huang, son of NVIDIA CEO Jensen Huang, took an unconventional route instead of entering the company directly. He first founded a well-known bar and later pursued an MBA. Huang then joined NVIDIA as an intern and entered its robotics lab, reflecting a start-over-from-the-ground-up approach that differs from the typical narrative surrounding the children of corporate leaders.
NVIDIA Cosmos 3: An Open Omni-model for Physical AI Reasoning and Action
Hugging Face Blog1d agoReleaseHugging Face Blog announces NVIDIA Cosmos 3, described as the first open omni-model for Physical AI reasoning and action. The title indicates a focus on AI systems that interact with physical-world scenarios rather than only text generation. Because the article body was not provided, its architecture, supported modalities, license, downloadable assets, benchmarks, and deployment requirements cannot be verified from the available material.
Startup offers free home cleaning if it can record it all for robot training
Ars Technica AI4d agoBusinessAI training startup Shift is offering free home cleanings while workers wear head-mounted cameras that record household chores. The footage is intended to become training data for domestic robots and related AI systems. The model highlights rising demand for real-world robotics data, while raising privacy questions about recording inside homes.
This AI startup will clean your home for free to train future robots
The Verge AI4d agoBusinessAI training startup Shift is offering to clean homes for free, with a significant condition: it records cleaners at work. The footage captures tasks like scrubbing, vacuuming, dusting, tidying, and washing. Shift says the material will be used to train future robots, raising clear questions about data collection inside private homes.
NASA Unveils $20B Moon Base Plan, Robots to Scout Lunar South Pole This Fall
INSIDE 硬塞 AI5d agoHardwareNASA announced a $20 billion plan to build a phased outpost near the Moon’s south pole. The agency will work with private companies and send robots first for scouting and deployment. The effort is intended to support Artemis crewed missions and prepare for long-term lunar presence after 2032.
Reachy Mini goes fully local
Hugging Face Blog6d agoHardwareHugging Face published a tutorial for running Reachy Mini conversations without cloud audio processing or API keys. The setup uses its speech-to-speech library as a cascaded VAD, STT, LLM, and TTS pipeline exposed through a Realtime API-compatible WebSocket. Recommended defaults include llama.cpp with Gemma 4, Silero VAD, Parakeet-TDT, and Qwen3-TTS, while allowing swaps to vLLM, MLX, Transformers, or hosted Responses API providers.
3D-printable humanoid legs let robotics experiments run wild
Ars Technica AI7d agoHardwareArs Technica reports that Hugging Face has introduced a roughly $2,500 bipedal humanoid robot project built around 3D-printable legs. The effort targets builders and researchers rather than mainstream consumers, lowering the hardware barrier for hands-on robotics experiments. Its broader significance is in open, reproducible embodied AI research, where models and control systems need physical platforms for testing.
This startup is betting India’s gig economy can train the world’s robots
TechCrunch AI7d agoBusinessHuman Archive, founded by Berkeley and Stanford researchers, is using India’s gig economy to gather physical-world AI data. Workers are paid to wear camera-equipped caps and sensor devices while moving through real environments. The company is targeting the growing demand from AI and robotics labs for real-world training data needed to develop physical AI systems.
South Korea Deputy PM Warns of AI Wealth Gap, Proposes "AI Dividend" Policy★ 75
INSIDE 硬塞 AI7d agoRegulationSouth Korea's Deputy PM warned that rapid AI and robot deployment could trigger severe wealth inequality, citing recent Samsung labor disputes. In response, the government is exploring "distribution justice" policies to return excess AI profits to the general public. The initiative aims to prevent technology dividends from being monopolized by tech giants and mitigate social unrest from automation.
The Internet can't stop watching Figure AI's humanoid robots handling packages
Ars Technica AI13d agoRelease人形機器人新創公司 Figure AI 近期推出了 24 小時不間斷的直播，展示其人形機器人在模擬倉庫環境中搬運與處理包裹。這個直播迅速在網路上走紅，吸引了大量觀眾駐足觀看。這項企劃不僅展示了機器人在實際工作場景中的穩定度與技術進步，也揭示了人類對於「人形」機器人從事日常勞動時所產生的獨特情感連結與好奇心。
The Autonomous Drone Tech Stack & Economics of Drones — Yaroslav Azhnyuk, The Fourth Law & Guest Host Noah Smith, Noahpinion
Latent Space15d agoCommentary本期 Latent Space 訪談了烏克蘭無人機新創 The Fourth Law 創辦人 Yaroslav Azhnyuk 與客座主持人 Noah Smith。他們深入探討了從寵物攝影機轉向 AI 導引武器的技術歷程，解析了現代自主無人機的技術棧與低成本製造經濟學。Azhnyuk 警告，西方國家在面對這場由 AI 驅動的無人機軍備競賽中顯得過於遲鈍，亟需正視其對未來國防的衝擊。
Gemini Robotics-ER 1.6: Powering real-world robotics tasks through enhanced embodied reasoning★ 85
Google DeepMind Blog50d agoReleaseGoogle DeepMind 發表全新 Gemini Robotics-ER 1.6 模型，專為自主機器人設計。該版本專注於提升「具身推理（Embodied Reasoning）」能力，特別加強了空間推理與多視角理解。這項技術突破將使機器人能更精準地整合不同角度的視覺資訊，並在真實世界中執行更複雜的物理任務。
Import AI 451: Political superintelligence; Google's society of minds, and a robot drummer ★ 75
Import AI (Jack Clark)64d agoCommentary知名 AI 科技週報《Import AI》第 451 期聚焦三大前沿議題。首先探討「政治超智慧」對人類社會與民主體制的衝擊，並提出科技是否已無法收回的詰問；其次介紹 Google 借鑑閔斯基理論的「心智社會」多智慧體協作研究，展示群體 AI 解決複雜問題的潛力；最後則分享了結合強化學習與精準物理控制的機器人鼓手技術。
LeRobot v0.5.0: Scaling Every Dimension★ 80
Hugging Face Blog85d agoReleaseHugging Face 發表開源機器人庫 LeRobot v0.5.0 重大更新。本次更新以「全面擴展（Scaling Every Dimension）」為核心，大幅提升了數據集規模、支援更多樣的模型架構（如 Diffusion Policy 與 ACT）、擴展了相容的實體硬體與模擬環境。這標誌著開源社群在具身智能（Embodied AI）領域邁出關鍵一步，降低了開發者進入實體機器人研究的門檻。
Bringing Robotics AI to Embedded Platforms: Dataset Recording, VLA Fine‑Tuning, and On‑Device Optimizations★ 75
Hugging Face Blog89d agoTutorial本指南介紹 Hugging Face 與 NXP 的合作成果，展示如何將視覺-語言-動作（VLA）模型應用於嵌入式機器人平台。內容涵蓋使用 LeRobot 進行資料集錄製、對 VLA 模型進行客製化微調，以及利用 NXP eIQ 工具包在 i.MX 晶片上進行量化與硬體加速優化，實現低延遲、低功耗的邊緣端機器人控制。
D4RT: Teaching AI to see the world in four dimensions★ 80
Google DeepMind Blog137d agoReleaseGoogle DeepMind 發表全新技術「D4RT」，成功將動態三維空間與時間（4D）的重建與追蹤整合。相較於以往技術，D4RT 實現了高達 300 倍的運算加速。這項突破不僅能讓 AI 更精準地理解動態物理世界，也將大幅推進自動駕駛、機器人互動及 AR/VR 領域的發展。
NVIDIA Cosmos Reason 2 Brings Advanced Reasoning To Physical AI★ 85
Hugging Face Blog147d agoReleaseNVIDIA 於 Hugging Face 宣布推出 Cosmos Reason 2 模型，專為「物理 AI（Physical AI）」與具身智能設計。該模型結合了先進的推理機制，使機器人與自主系統能更深層地理解物理規律、進行複雜的空間時間規劃。此模型的發布將大幅降低開發者構建具備物理常識與決策能力 AI 應用的門檻。
NVIDIA brings agents to life with DGX Spark and Reachy Mini★ 80
Hugging Face Blog148d agoReleaseHugging Face 宣布與 NVIDIA 合作，將 NVIDIA 全新的 DGX Spark 運算平台與 Pollen Robotics 的 Reachy Mini 機器人進行深度整合。此合作旨在降低具身智能（Embodied AI）的開發門檻，讓開發者能透過 Hugging Face 的 LeRobot 開源機器人框架，輕鬆將 AI 代理（Agents）部署至實體硬體，實現具備物理互動能力的實體 AI 應用。
Google's year in review: 8 areas with research breakthroughs in 2025★ 85
Google DeepMind Blog161d agoCommentaryGoogle DeepMind 發表 2025 年度回顧，總結了今年在 8 大關鍵領域的研究突破。內容涵蓋 Gemini 多模態模型的演進、AlphaFold 3 帶來的生醫革命、AlphaProof 在數學推理的進展，以及 AI 在氣象預測、材料科學和機器人控制等領域的實際應用。這份回顧展示了 AI 如何從單純的語言助手，加速轉化為推動人類科學探索的強大引擎。
Run Isaac 0.1 on Replicate
Replicate Blog188d agoReleaseReplicate 平台正式上架 Isaac 0.1 模型。這是一款輕量級且具備定位能力（Grounded）的視覺語言模型（VLM），專為真實世界的空間感知與物體識別而設計。開發者現在可以透過 Replicate 的雲端 API 輕鬆部署並整合此模型，適用於機器人、物聯網及空間運算等需要即時視覺理解的應用場景。
Join the AMD Open Robotics Hackathon
Hugging Face Blog200d agoReleaseAMD 與 Hugging Face 攜手推出「AMD Open Robotics Hackathon」，旨在推動開源機器人與實體 AI（Embodied AI）技術的發展。參賽者將利用 Hugging Face 的 LeRobot 開源機器人框架，並結合 AMD 的強大運算平台（如 ROCm、Ryzen AI 或 Kria 開發套件）來開發創新的機器人應用。活動提供豐富的硬體與獎金回饋，是機器人開發者與 AI 研究人員展身手的絕佳機會。
Building a Healthcare Robot from Simulation to Deployment with NVIDIA Isaac★ 75
Hugging Face Blog216d agoTutorial本文介紹 Hugging Face 開源機器人庫 LeRobot 與 NVIDIA Isaac 平台的整合應用。透過 Isaac Sim 的高保真物理模擬環境進行安全訓練，再將模型無縫遷移至實體醫療機器人上。此方案降低了醫療自動化設備的開發門檻，為手術輔助、照護等場景提供高效的端到端開發路徑。
How to Build a Healthcare Robot from Simulation to Deployment with NVIDIA Isaac for Healthcare★ 70
Hugging Face Blog216d agoTutorialNVIDIA 介紹如何利用其 Isaac 機器人平台開發醫療照護機器人。文章詳細說明了從 Isaac Sim 虛擬醫院環境的建立、利用 AI 與強化學習訓練機器人安全互動，到最終透過 NVIDIA Jetson 或 IGX 邊緣運算平台進行「模擬到真實（Sim-to-Real）」部署的完整工作流，為醫療自動化與輔助照護提供安全且高效的開發路徑。
LeRobot v0.4.0: Supercharging OSS Robot Learning★ 75
Hugging Face Blog221d agoReleaseHugging Face 旗下的開源機器人學習庫 LeRobot 迎來 v0.4.0 重大更新。本次版本專注於降低機器人 AI 的開發門檻，優化了從資料收集、模型訓練到實體部署的完整工作流。新版本擴展了對多種平價開源硬體的支援，並加強了與 Hugging Face Hub 的整合，讓開發者能更輕鬆地共享機器人資料集與預訓練模型，加速機器人領域的開源生態發展。
Gemini Robotics 1.5 brings AI agents into the physical world★ 85
Google DeepMind Blog221d agoReleaseGoogle DeepMind 發表 Gemini Robotics 1.5，旨在將 AI Agent 的能力帶入實體世界。透過此系統，機器人將具備更強大的環境感知、多步驟任務規劃、邏輯思考、工具使用以及實體行動能力。這項進展代表著「實體代理（Physical Agents）」時代的開啟，能更有效率地解決現實世界中複雜且多步驟的實體任務。
`LeRobotDataset:v3.0`: Bringing large-scale datasets to `lerobot`★ 75
Hugging Face Blog259d agoReleaseHugging Face 宣布推出 LeRobotDataset:v3.0，這是其開源機器人庫 LeRobot 的重大更新。新版本重新設計了數據集格式，旨在解決大規模機器人數據（如影像、關節狀態、動作）的儲存與 I/O 瓶頸。v3.0 提供了更高效的串流載入、統一的機器人本體架構 Schema，並無縫整合至 Hugging Face Hub，為訓練下一代具身智能（Embodied AI）奠定基礎。
Asynchronous Robot Inference: Decoupling Action Prediction and Execution★ 75
Hugging Face Blog327d agoOpinion傳統機器人控制常受限於 AI 模型推論速度，導致動作不流暢。Hugging Face 提出「非同步機器人推論」架構，將「動作預測（AI 模型）」與「動作執行（硬體控制）」解耦。此方法允許硬體以高頻率（如 100Hz+）持續運行，而較慢的 AI 模型則在背景非同步更新動作指令，大幅提升了機器人在實時環境中的反應速度與操作流暢度。
Reachy Mini - The Open-Source Robot for Today's and Tomorrow's AI Builders★ 75
Hugging Face Blog328d agoReleaseHugging Face 與 Pollen Robotics 合作推出全新開源機器人「Reachy Mini」，專為 AI 開發者與研究人員設計。該機器人主打高性價比、易於組裝，並深度整合 Hugging Face 的 LeRobot 開源機器人庫。開發者可輕鬆透過模仿學習（Imitation Learning）訓練 AI 模型，控制實體機器人完成各種操作任務，加速具身智能（Embodied AI）的普及。