Mag World播客推出了一期节目，由主持人Saul和Mike探讨了65年来随机存储器（RAM）容量的演进轨迹 [1]。节目以RAM容量作为计算能力的代理指标，采用指数级标度方法追溯了个人计算机从早期到现代的发展阶段 [1]。
早期计算机的内存极其有限 [1]。Altair是mag 2级计算机，配备数百字节到4KB的可扩展内存 [1]；TRS-80属于mag 3级，拥有4KB RAM [1]；Apple II则达到mag 4级，具备32-64KB的内存 [1]。
随着技术进步，内存容量呈指数增长 [1]。当代消费级计算机已属mag 10级，配置16-32GB RAM [1]，而iPhone作为移动设备代表mag 9级，内存范围在1-12GB之间 [1]。目前尚未出现mag 12级计算机（1TB RAM的设备） [1]。
在数据中心和AI领域，计算资源规模显著更大 [1]。Claude AI运行在mag 12到mag 13级的计算资源上 [1]，而现代数据中心的规模已达mag 16到mag 17级 [1]。超级计算机的计算能力比消费级硬件多出3-4个数量级 [1]。
芯片工艺方面，现代芯片工艺已进步到3纳米（mag负9米） [1]，接近单个原子尺寸（mag负10米）的极限 [1]。节目还涉及Moore定律的演变，指出晶体管密度翻倍周期从1965年的18个月延长至1975年的2年 [1]。
过去三十年间，计算机的应用能力也经历了巨大飞跃，从chiptune音乐发展到4K流媒体视频 [1]。
A Mag World podcast episode explores the evolution of computer memory and computing power over six decades by using RAM capacity as a proxy metric on an exponential magnitude scale [1]. Hosts Saul and Mike trace the progression of personal computing from magnitude 2 to magnitude 17, spanning from early byte-level memory systems to modern data center capabilities measured in petabytes [1].
The discussion traces key milestones in computing history through this magnitude framework. The Altair represented magnitude 2 computing with hundreds of bytes to 4 kilobytes of expandable memory [1], while the early TRS-80 operated at magnitude 3 with 4 kilobytes of RAM [1]. The Apple II emerged as a magnitude 4 computer with 32 to 64 kilobytes [1]. Modern consumer computers typically reach magnitude 10, containing 16 to 32 gigabytes of RAM [1]. Smartphones like the iPhone function as magnitude 9 devices with 1 to 12 gigabytes [1]. The podcast notes that magnitude 12, representing 1 terabyte of RAM, has not yet been achieved in practical systems [1].
The conversation contextualizes these developments within Moore's Law, noting the adjustment from an initial 18-month doubling period established in 1965 to a 2-year cycle by 1975 [1]. Current chip manufacturing has advanced to 3-nanometer processes, equivalent to magnitude negative 9 meters, while individual atoms measure at magnitude negative 10 meters [1].
Advanced systems operate at substantially higher scales. Claude AI runs on computing resources ranging from magnitude 12 to magnitude 13 [1], while modern data centers span magnitude 16 to magnitude 17 [1]. Supercomputers exceed consumer-grade hardware by three to four orders of magnitude [1]. Over the past three decades, computing applications have evolved from chiptune music to 4K streaming video [1].