That's not the Compute Board @neverlastn (that's here: https://
store.axelera.ai/products/metis
-compute-board-with-arm-based-rk3588
…). This is the carrier board for the Metis + Portenta. But we've an early acess program for anyone who's interested in Axelera+Arduino development, if that sounds appealing? https://
axelera.ai/development-sy
stems/arduino
…
AI HARDWARE
-
Axelera Metis Compute Board Carrier for Arduino Development
By
–
-
AI-powered vein detection device improves blood draw safety
By
–
Esta sorprendente máquina permite a los sanitarios realizar extracciones de sangre de una forma más segura al revelar la posición de las venas.
— Juan Merodio (@juanmerodio) 22 avril 2025
Gracias a su tecnología de proyección y detección infrarroja, mapea en tiempo real el sistema venoso del paciente, reduciendo intentos… pic.twitter.com/jWBqFHtyrXEsta sorprendente máquina permite a los sanitarios realizar extracciones de sangre de una forma más segura al revelar la posición de las venas. Gracias a su tecnología de proyección y detección infrarroja, mapea en tiempo real el sistema venoso del paciente, reduciendo intentos
-
AionLabs Achieves 99.99% Accuracy with SambaNova Cloud Inference
By
–
@AionLabs turned to SambaNova Cloud for 671B-speed inference — and got 99.99% accuracy to match. No bottlenecks. Just breakthroughs. Read more in our blog
-

Hybrid Digital-Analog Quantum Simulation Emulates Magnet Behavior
By
–
Today on the blog we describe a novel hybrid digital–analog approach to quantum simulation that we used to emulate quantum behavior of a magnet, resulting in a notable scientific discovery and new possibilities for beyond-classical applications. More at https://
goo.gle/3EwmZjG -

Llama4 Maverick Achieves 800 Tokens Per Second on SambaNova
By
–
How it feels using #Llama4 Maverick on SambaNova Cloud! ⚡️
— SambaNova (@SambaNovaAI) 21 avril 2025
Kamehameha, at up to 800 t/s am I right. 👀🦾 pic.twitter.com/F5m1G0V6qnHow it feels using #Llama4 Maverick on SambaNova Cloud! Kamehameha, at up to 800 t/s am I right.
-
South Korean Tesollo Unveils DG-5F Humanoid Hand With 20 Joints
By
–
La société sud-coréenne Tesollo a présenté un extrait impressionnant de son DG-5F, une main humanoïde conçue pour imiter fidèlement les mouvements naturels de la main humaine.
— VISION IA (@vision_ia) 21 avril 2025
Elle intègre 20 articulations contrôlées indépendamment, peut saisir jusqu’à 12 kg avec une prise… pic.twitter.com/3QNZUuWN3nLa société sud-coréenne Tesollo a présenté un extrait impressionnant de son DG-5F, une main humanoïde conçue pour imiter fidèlement les mouvements naturels de la main humaine. Elle intègre 20 articulations contrôlées indépendamment, peut saisir jusqu’à 12 kg avec une prise
-

Google Releases Gemma 3 with Quantization-Aware Training for Consumer GPUs
By
–
Google also released a new version of Gemma 3, optimized with ‘Quantization-Aware Training’ This reduces the 27B model's memory requirements, enabling it to run on consumer GPUs, like the NVIDIA RTX 3090, with maintained performance
-
RL, Attention, and Bayesianism Alone Insufficient for AI Progress
By
–
RL is not all you need, nor attention nor Bayesianism nor free energy minimisation, nor an age of first person experience. Such statements are propaganda. You need thousands of people working hard on data pipelines, scaling infrastructure, HPC, apps with feedback to drive
-
TPU Hardware Architecture: From v1 to v4 Optical Interconnect
By
–
If you want to read about them from the hardware perspective, read these three papers: https://
arxiv.org/abs/1704.04760
(about TPU v1, the one that started it off) https://
arxiv.org/abs/2304.01433 (about TPU v4 and its optical interconnect) https://
gwern.net/doc/ai/scaling
/hardware/2021-jouppi.pdf
… (ten lessons) -
Limited Edition AI Hardware Product Launches with 2000 Units
By
–
yeah, I'm on the waitlist, sad its only a limited ~2k units edition.