Google TurboQuant: AI Memory Usage Reduced 6x, Speed Increased 8x

Google Research unveils TurboQuant compression algorithm, reducing LLM memory 6x and increasing speed 8x without accuracy loss or retraining.