Hacker News with Generative AI: Hardware

Framework Laptop 12 pre-orders open next week (frame.work)
At the end of our launch livestream last month, we teased Framework Laptop 12, a colorful little laptop that is the ultimate expression of our product philosophy. We received a ton of interest around this product, and we have a lot more to share on Framework Laptop 12… in exactly a week!
Framework Laptop 12 – more details and teardown [video] (youtube.com)
Zero Day in Microchip SAM Microcontrollers (recessim.com)
This write-up will cover analysis of the Microchip (ATMEL) SAM4C32 microcontroller vulnerability that allows an attacker to gain unlocked JTAG access to a previously locked device.
Nintendo unveils Switch 2 ahead of June 5 launch (arstechnica.com)
In a highly anticipated Nintendo Direct presentation this morning, the company revealed new information about the Nintendo Switch 2 hardware and software ahead of a newly announced June 5, 2025, release date.
Nintendo Switch 2 Game-Key Card Overview (nintendo.com)
Game-key cards are different from regular game cards, because they don’t contain the full game data. Instead, the game-key card is your "key" to downloading the full game to your system via the internet.
Measuring Acceleration Structures (zeux.io)
Hardware accelerated raytracing, as supported by DirectX 12 and Vulkan, relies on an abstract data structure that stores scene geometry, known as “acceleration structure” and often referred to as “BVH” or “BLAS”. Unlike geometry representation for rasterization, rendering engines can not customize the data layout; unlike texture formats, the layout is not standardized across vendors.
The Nvidia DGX Spark Is a Tiny 128GB AI Mini PC Made for Scale-Out Clustering (servethehome.com)
With 20 Arm cores connected using C2C to a Blackwell generation GPU, 128GB of LPDDR5X memory, and 200GbE NVIDIA ConnectX-7 networking, the NVIDIA DGX Spark is exciting. At $3999 it is far from cheap. On the other hand, we expect folks to create the most awesome clusters with this.
Attention is NOT all you need: Qwerky-72B trained using only 8 AMD MI300X GPUs (recursal.ai)
We are proud to announce the updated Qwerky-72B and 32B.
Dual RTX 5090 Beats $25,000 H100 in Real-World LLM Performance (hardware-corner.net)
AI enthusiasts looking for top-tier performance in local LLMs have long considered NVIDIA’s H100 to be the gold standard for inference, thanks to its high-bandwidth HBM3 memory and optimized tensor cores. However, recent benchmarks show that a dual RTX 5090 setup, while still pricey, outperforms the H100 in sustained output token generation, making it an ideal choice for those seeking the best possible performance for home use, especially for models up to 70B parameters.
Measuring Acceleration Structures (zeux.io)
Hardware accelerated raytracing, as supported by DirectX 12 and Vulkan, relies on an abstract data structure that stores scene geometry, known as “acceleration structure” and often referred to as “BVH” or “BLAS”. Unlike geometry representation for rasterization, rendering engines can not customize the data layout; unlike texture formats, the layout is not standardized across vendors.
Show HN: Terminal dashboard that throttles my PC during peak electricity rates (naveen.ing)
Windows on ARM on a Smart Watch (gus33000.me)
Many years of spending, sorry, losing time making Windows on ARM run on anything not a computer, I started getting bored
Notes on the Pentium's microcode circuitry (righto.com)
Most people think of machine instructions as the fundamental steps that a computer performs. However, many processors have another layer of software underneath: microcode. With microcode, instead of building the processor's control circuitry from complex logic gates, the control logic is implemented with code known as microcode, stored in the microcode ROM. To execute a machine instruction, the computer internally executes several simpler micro-instructions, specified by the microcode.
Nvidia's latest AI PC boxes sound great – for data scientists with $3k to spare (theregister.com)
Nvidia's latest AI PC boxes sound great – if you're a data scientist with $3,000 to spare
Build an 8-bit computer from scratch (eater.net)
Nvidia GPU roadmap confirms it: Moore's Law is dead and buried (theregister.com)
As Jensen Huang is fond of saying, Moore's Law is dead – and at Nvidia GTC this month, the GPU-slinger's chief exec let slip just how deep in the ground the computational scaling law really is.
Tearing Down the Unitree Go2: A Robotics Expert's Deep Dive [video] (youtube.com)
Xfinity XB3 hardware mod: Disable WiFi and save 2 watts (github.com)
Comcast has a prepaid "Xfinity NOW" service that's cheaper than normal Xfinity, with unlimited data instead of a 1.2TB/month cap.
Open source, 3D-printable smart chess board (thangs.com)
Print. Click. Play. OpenChess is a fully open source smart chessboard designed to make interactive, intelligent gameplay accessible to everyone. By combining low-cost electronics, 3D printing, and customizable software, OpenChess empowers makers, educators, and chess lovers to build their own connected chess experience — without the high price tag.
An Ode to the Game Boy Advance (brainbaking.com)
In March 2001, Nintendo introduced an advanced portable model to the gaming market with the release of the Game Boy Advance (GBA, codenamed Advanced Game Boy or AGB). Equipped with a modernized 32-bit ARM CPU running at twice the speed of the Game Boy Color (GBC), this small device was more than capable of playing SNES-like games—still at the price of only two AA batteries.
Lilygo T-Deck Pro is a mobile dev kit with ePaper display, QWERTY keyboard, 4G (liliputing.com)
Last summer LILYGO launched a pocket-sized mobile communications device called the T-Deck Plus that looks like a phone, but is really more of a mobile dev kit with a 2.8 inch IPS LCD display, a BlackBerry keyboard, and support for WiFi, Bluetooth, and LoRa wireless connectivity… but no support for cellular networks.
Ploopy Classic 2 open source trackball (ploopy.co)
The Classic 2 is a brand new revision of the original Classic, and kits are available today starting at $144CAD!
Commodore C64 still used in Indiana bakery as cash registers (tomshardware.com)
6502 as a Service (emulationonline.com)
The 6502 is now available in the Chiplab! This allows you to upload your programs to be queued for running against a real 6502 chip. For each cycle of execution, you can observe the values of all the busses of the physical chip.
U.S. Atari parts store still open after 41 years, spent $100K+ designing parts (tomshardware.com)
Swapping Apple's M4 SSD – Faster, Bigger, Cheaper [video] (youtube.com)
Lvgl: Embedded graphics library to create beautiful UIs (github.com/lvgl)
LVGL is the most popular free and open source embedded graphics library to create beautiful UIs for any MCU, MPU and display type. It's supported by industry leading vendors and projects like  Arm, STM32, NXP, Espressif, Nuvoton, Arduino, RT-Thread, Zephyr, NuttX, Adafruit and many more.
Bolt Graphics Zeus a New GPU Architecture with Up to 2.25TB of Memory and 800GbE (servethehome.com)
Bolt Graphics Zeus The New GPU Architecture with up to 2.25TB of Memory and 800GbE
Real Time Chess – A physical chess board without the concept of turns (github.com/misprit7)
Chess is boring. I'm boring too so I enjoy it anyways, but I can't help but think "I could design it better." Normally in chess players move sequentially in turns, but this introduces a huge latency bug that the developers of chess forgot to patch: you spend literally half the time waiting for your opponent!
Show HN: Physical Pomodoro Timer with ESP32 and e-paper screen (github.com/Rukenshia)
This is the repository for an ESP32 based pomodoro timer. It uses an ePaper display and a rotary dial for input. The code in this repository will not be ready-to-use, as some assets and fonts have been removed. However, if you really want to you should be able to adapt the code to your needs.