AI Hardware & Devices
-
Midjourney Medical puts a 500,000-transducer full-body scanner inside a spa
The company best known for turning text prompts into surreal images just shipped a physical machine. Midjourney Medical, announced June 18, is a full-body ultrasonic scanner — you stand in a water-filled chamber for about 60 seconds and it builds a 3D map of your insides. No radiation, 500,000 transducers, roughly 17GB/s of data per… Continue reading
-
Dune is a context-aware Mac keypad whose three keys adapt to your app and Claude
Most desk gadgets do one fixed thing. Dune, a context-aware keypad for Mac, takes the opposite approach: it sits beside your keyboard and changes what its three keys do in real time based on whatever app is in the foreground. ## What Dune does The three keys remap themselves as you switch apps — one… Continue reading
-
Monako Glass Puts Claude Code and Codex on a 48g Pair of Glasses
Most smart glasses are built for photos and notifications. Monako Glass is built for one thing: running AI coding agents on your face. The 48-gram wearable runs a full Buildroot Linux OS with a waveguide display, so you can drive Claude Code, Codex, or any coding agent hands-free, anywhere. ## What Monako Glass does Instead… Continue reading
-
NVIDIA’s GR00T N1.6 gives humanoid robots full-body control from one vision-language-action model
NVIDIA’s Isaac GR00T N1.6 is the latest open foundation model for humanoid robots — a vision-language-action (VLA) model that turns camera streams, robot state, and plain-language instructions into one unified control policy. ## What changed from N1.5 N1.6 isn’t a rewrite; it’s a sharpening. Architecture, data, and modeling improvements let it beat the previous N1.5… Continue reading
-
NVIDIA’s GRAIL generates humanoid robot skills from 3D assets and video
Teaching a humanoid to pick up a box and carry it across a room usually means hours of teleoperation or motion capture. NVIDIA’s GRAIL goes the other way: it generates loco-manipulation behavior — locomotion plus object handling — directly from 3D assets and motion priors pulled from video. ## The data bottleneck it targets Humanoid… Continue reading
-
Google Gemini Intelligence for Android arrives this summer, hitting Pixel and Samsung first
Google just rebranded Android. At The Android Show I/O Edition on May 12, 2026, Google’s product VP said it out loud: “We’re moving from an operating system to an intelligence system.” That sentence is the whole strategy. Gemini Intelligence is the new name for everything Gemini does on your phone, and it ships this summer… Continue reading
-
Google retires Chromebook for Googlebook, betting the laptop on Gemini
Google just made its biggest hardware reframing in a decade. It unveiled Googlebook on May 12 — a line of AI-native laptops built from the ground up around Gemini. ChromeOS is being quietly phased out. The new platform runs on an Android foundation. First units ship fall 2026, made by Acer, Asus, Dell, HP, and… Continue reading
-
Needle by Cactus Compute squeezes Gemini 3 tool calling into 26M parameters
Cactus Compute, the YC-backed on-device inference startup, open-sourced Needle today: a 26M-parameter Simple Attention Network that does single-shot function calling on phones and smartwatches. No MLPs, just attention and gating — the team argues FFN params are wasteful at this scale, and cross-attention is the right primitive for routing a query to the right tool.… Continue reading
-
MiniCPM-V 4.6 packs Qwen3.5-2B-level vision into a 1.3B model
OpenBMB open-sourced MiniCPM-V 4.6 on May 11. A 1.3B-parameter multimodal model built on SigLIP2-400M plus Qwen3.5-0.8B, aimed squarely at the edge — phones, laptops, consumer GPUs. The trick is in the visual encoder. LLaVA-UHD v4 brings intra-ViT early compression with a hybrid 4x/16x token compression ratio, cutting vision encoding compute by more than 50% versus… Continue reading
-
NVIDIA releases CUDA-Oxide 0.1, the first official Rust-to-CUDA compiler
NVIDIA quietly shipped CUDA-Oxide on May 7, 2026, and it pulled 323 points on Hacker News inside a day. This is the first NVIDIA-sanctioned path to writing CUDA kernels in pure Rust — no DSLs, no FFI bindings, no awkward wrappers. You write safe, idiomatic Rust, and a custom rustc back-end emits straight to NVIDIA… Continue reading
Top AI Product 