Spark creator bags computing gong for making big data a little bit smaller

When Zaharia started work on Spark around 2010, analyzing "big data" generally meant using MapReduce, the Java-based ...
EDN

Advancing AI performance with HBM4, SPHBM4 DRAM solutions

HBM has become one of the most successful and widely adopted examples of chiplet-based integration in AI systems.

Karpathy shares 'LLM Knowledge Base' architecture that bypasses RAG with an evolving markdown library maintained by AI

Karpathy proposes something simpler and more loosely, messily elegant than the typical enterprise solution of a vector ...

Cachee Achieves 28.9-Nanosecond Cache Reads – Verified as Fastest Full-Featured Cache Engine Ever Benchmarked

At 100 billion lookups/year, a server tied to Elasticache would spend more than 390 days of time in wasted cache time.
TMCnet

Aerospike NoSQL Database 8 Delivers Durable and Low-Latency Memory Store for LangGraph Agentic AI Workflows

MOUNTAIN VIEW, Calif., March 25, 2026 (GLOBE NEWSWIRE) -- Aerospike Inc. today unveiled LangGraph integration for the award-winning NoSQL Database 8, delivering persistent memory to stateless AI ...
IEEE

Bauhaus: Restructuring Vector Database for LLM Retrieval on CXL-Based Tiered Memory

Abstract: Retrieval-augmented generation pipelines store large volumes of embedding vectors in vector databases for semantic search. In Compute Express Link (CXL)-based tiered memory systems, ...