low-power

Comparing the positioning differences between the ESP32 and STM32: the ESP32 is Espressif's low-cost connectivity solution with integrated WiFi/Bluetooth — higher power draw, but deep sleep can reach 10µA; the STM32 is ST's industrial-grade low-power series, with the L series drawing under 1µA in Stop mode, but requiring an external wireless module. Selection should balance connectivity needs against power targets.

The Bonsai 1.7B AI model can solve simple physics problems efficiently on a low-power Intel N97 processor, achieving inference speeds of about 9.1 tokens per second.

Savartus's Enterprise Laser Storage offers physically immutable optical archive libraries for secure, long-term data preservation, featuring EMP-proof design, 50-100 year media life, and low power consumption.

M5Stack has launched PaperMono, a compact E-Ink development terminal powered by ESP32-S3, designed for low-power IoT projects with features like touch display, NFC, and LoRa connectivity.

Pants for Birds introduces the ADSBee m1421, the world's smallest dual-band ADS-B receiver module, an open-source, low-power hardware designed for embedded aircraft tracking applications.

The article describes the construction of a portable, sensitive, and low-power analog Geiger counter, designed for measuring low radiation levels using a large Geiger-Müller tube.

A hobbyist describes building a low-power llama.cpp server using an Intel N100 motherboard and a refurbished RTX 5060 Ti, sharing performance numbers, power consumption, and model choices.

Details a method to run a 13 million parameter ASR Conformer model directly on a microcontroller, highlighting advances in edge AI deployment.

This study explores the feasibility of classifying ten hand gestures using a single-channel sEMG signal combined with lightweight machine learning models, achieving up to 90% accuracy. It demonstrates potential for cost-effective, low-power gesture recognition.

A new brain-inspired hardware design enables faster and more energy-efficient anomaly detection for AI systems, drawing on principles of neuromorphic computing.

A fully biocompatible multiterminal neuromorphic biodevice using quaternized chitosan is developed, operating at ultralow voltages and simulating visual nervous system processing.

Texas Instruments introduced the MSPM0C1104 microcontroller, a 1.38 mm² device featuring an ARM Cortex-M0+ CPU at 24MHz with 16KB flash and 1KB SRAM, designed for small form factor and low-power embedded applications.

This paper presents low-power analogue neural networks that place trainable nonlinear functions on connections, inspired by Kolmogorov-Arnold networks, enabling efficient continuous control tasks with far fewer nodes and connections than multilayer perceptrons, demonstrated on hardware with projected microWatt power.

MIT researchers have developed a new system-on-a-chip that enables tiny robots to create detailed 3D maps of their environments in real-time using only about 6 milliwatts of power, potentially enabling long-duration autonomous navigation in complex spaces.

A developer ran DeepSeek-V4-Flash on a Raspberry Pi 5 by streaming model weights from an NVMe SSD, achieving 1.3 tokens/second at 8 watts, demonstrating the feasibility of frontier-adjacent open-weight models on low-cost, offline hardware.

A developer successfully ran the 284B-parameter DeepSeek-V4-Flash model on a Raspberry Pi 5 at over 1 tok/s, using an untouched GGUF file from antirez after extensive experimentation.