A bidirectional pixel that is screen and camera in one

I paired a local LLM with Frigate and Home Assistant, and my smart cameras finally understand what they are looking at

Người đam mê nhà thông minh tự lưu trữ đã tích hợp một LLM cục bộ (Qwen3.6-35B-A3B qua llama-server) với Frigate NVR và Home Assistant để tạo ra các bản tóm tắt sự kiện camera thông minh có ý nghĩa nhờ AI. Hệ thống sử dụng tích hợp Frigate HACS, tích hợp LLM Vision với adapter MMPROJ, cùng blueprint AI Event Summary để tự động gửi thông báo chi tiết khi camera phát hiện hoạt động, hoàn toàn không phụ thuộc vào dịch vụ đám mây.

Lập trình viên muốn tự chủ hóa hệ thống an ninh thông minh và tối ưu hóa hiệu suất AI trên thiết bị cá nhân sẽ tìm hiểu cách kết hợp LLM với Frigate và Home Assistant để tự động hóa cảnh báo thông minh từ camera mà không cần phụ thuộc vào dịch vụ cloud.

This Steam Controller mod uses vibration to guide the gamepad back to its charging dock on its own

Một modder tên Ray Foss phát triển công cụ Auto-Charge Tracker dựa trên trình duyệt, sử dụng camera tracking và động cơ rung để tự động điều hướng Steam Controller quay trở lại dock sạc. Dự án được chia sẻ trên GitHub.

Lập trình viên nên đọc bài này để khám phá cách kết hợp camera tracking và haptic feedback—các kỹ thuật cơ bản trong AI và IoT—để tạo ra giải pháp tự động hóa thú vị cho thiết bị ngoại vi, mở rộng kiến thức về cách ứng dụng công nghệ sensor và haptic trong thiết kế phần mềm thực tế.

Firefly AIBOX-9075 Edge AI box features Qualcomm IQ-9075 SoC with 200 TOPS NPU, 36GB LPDDR5

Firefly has launched the AIBOX-9075, an industrial Edge AI box powered by the Qualcomm DragonWing IQ-9075 SoC. It delivers up to 200 TOPS of AI performance via a Hexagon Tensor Processor NPU, paired with 36GB LPDDR5 ECC memory and 128GB UFS 2.2 storage. The device targets edge AI workloads including private LLM deployment, robotics, and computer vision. It supports popular AI frameworks (TensorFlow, PyTorch, ONNX) and LLMs such as DeepSeek-R1, Llama, and Gemma, achieving up to 22 tokens/s with Llama2-7B. Connectivity includes dual 2.5GbE with TSN, optional Wi-Fi 6, 4G/5G, and 8x GMSL2 camera inputs. The rugged aluminum enclosure operates from -40°C to 85°C. It runs Ubuntu and Yocto Linux and is priced at $1,239 from the Firefly store.

Robotaxis drives miles just to get cleaned and charged; this new startup wants to fix that

Aseon Labs, a Redwood City startup from Y Combinator's 2026 spring cohort, has raised $10M in seed funding to build parking space-sized automated pods that clean, inspect, and charge robotaxis in-city. The core problem they address is 'deadhead miles' — empty trips robotaxis make to distant depots — which hurt fleet utilization and profitability. Their pods use robotic arms, cameras, and vision-language-action AI models to handle routine maintenance autonomously, while flagging complex issues for human handling at central depots. The units are designed as temporary structures to avoid lengthy permitting and can be relocated if a site underperforms. No robotaxi contracts have been signed yet, but the company plans to build five prototypes and grow its team with the new funding.

Aseon Labs raises ten million dollars to build parking-space-sized pods that charge and clean robotaxis

Aseon Labs has raised $10M in seed funding to build parking-space-sized automated pods that charge, clean, and inspect robotaxis in place. The Y Combinator-backed startup aims to solve the deadhead miles problem — robotaxis driving empty to distant depots for servicing — by deploying modular pods across city parking lots and gas stations. Using robotic arms, computer vision, and vision-language-action models, the pods can handle most routine maintenance autonomously. The company estimates its solution could cut reset costs by 50%, reduce downtime by 65%, and add over $50K in annual revenue per vehicle. Aseon Labs is pre-product with five prototypes as its immediate goal, targeting a market that Goldman Sachs projects will reach $415B by 2035.

Berlin’s Almetra raises €16.3M to turn factory video into live data

Berlin startup Almetra (formerly Deltia) has raised €16.3M in Series A funding to expand its AI-powered factory floor analytics platform. The company mounts cameras above assembly lines at manufacturers like Bosch, Siemens Energy, and ABB, converting video footage into live production data — cycle times, output rates, equipment utilisation — without requiring IT system integration. Customers report productivity gains of 15–19%. The round was led by blisce/, with participation from Merantix Capital and others. Almetra has been accepted into Google DeepMind's Robotics Accelerator and an AWS/Nvidia/MassRobotics Physical AI Fellowship, positioning it as a potential data source for industrial robotics. The company plans to use the funding to expand into the US and build out robotics applications.

AI brings object-level vision prosthetics closer to reality

Researchers at EPFL's NeuroAI Lab have developed AI-based topographic neural network models that predict optimal brain stimulation patterns to evoke perception of complex visual objects — such as faces and houses — rather than just simple light flashes. The models were validated in live trials on sighted monkeys in Amsterdam, showing that model-guided cortical stimulation can bias visual object perception in predictable ways. While the team cannot yet create object perception from scratch (without any visual input), this is the stated next goal. The approach could also extend to improving cochlear implants for auditory prosthetics.