New KV cache compaction technique cuts LLM memory 50x without accuracy loss

MIT researchers developed Attention Matching, a KV cache compaction technique that compresses LLM memory by 50x in seconds — ...

Google PM open-sources Always On Memory Agent, ditching vector databases for LLM-driven persistent memory

Enterprise AI teams are moving beyond single-turn assistants and into systems expected to remember preferences, preserve ...
InfoWorld

The browser is your database: Local-first comes of age

The thick client is making a comeback. Here’s how next-generation local databases like PGlite and RxDB are bringing ...
IEEE

Design and Implementation of a Multi-Protocol Universal Memory Testing System Based on FPGA

Abstract: To address the issues of high cost, poor multiprotocol adaptability, and complex vector programming in existing memory testing machines, this paper designs and implements a universal memory ...
about.fb

Helping People With Memory Loss Find Their Way With AI Glasses

Our AI glasses combine iconic style and cutting-edge technology to help people from all backgrounds connect more easily with one another. We have already seen the positive impact that our wearable ...