果的新TurboQuant AI算法将存储成本降低六倍, 并提高速度八倍. Alphabet's new TurboQuant AI algorithm cuts memory costs by six times and boosts speeds eightfold.

字母表推出了一个新的AI记忆压缩算法, Alphabet has introduced a new AI memory compression algorithm called TurboQuant. 这项技术通过至少减少六倍的密钥值缓存内存需求, 降低计算成本, 并提高速度至八倍而不会损害数据质量. The technology reduces computational costs by cutting key-value cache memory needs by at least six times and boosting speeds up to eightfold without sacrificing data quality. 这一进步加强了Alphabet在定制芯片和人工智能模型中获得的现有成本优势,并可能巩固其在AI领域的领导地位. This advancement strengthens Alphabet's existing cost advantages derived from its custom-built chips and AI models, potentially solidifying its leadership in the AI sector.