InfoQ

Google Releases Gemma 3 270M Variant Optimized for Function Calling on Mobile and Edge Devices

FunctionGemma is a new, lightweight version of the Gemma 3 270M model, fine-tuned to translate natural language into structured function and API calls, enabling AI agents to "do more than just talk" ...

HBF memory stacks could improve GPU performance and AI workloads

HBF memory stacks could improve GPU performance and AI workloads, offering ten times HBM capacity in real systems ...
Fireship on MSNOpinion

OpenAI solves the stack memory problem

For years, software stacks kept getting more complex. OpenAI is moving in the opposite direction. This video breaks down how ...
IEEE

VCS (Vertical Copper Post Stack) for on-Device AI Memory Solution

Abstract: With the advancement of on-device AI, we have developed a new memory package platform by applying copper post to meet the growing demand for high-bandwidth memory. The development of a new ...
IEEE

Low-Power SRAM Memory using Sleepy-Stack Hybrid Power Gating Technique

Abstract: Nowadays, with everything being digital, memory is an essential part of any integrated device. It makes a substantial contribution to the circuit power overall as well. The growing demand ...