The Intelligent Counter
After the family gathering at Tiffany’s maternal home, Yoo Jae-won slightly revised the operating policy of the information team. Previously, the team had operated in a passive manner—simply retrieving whatever Yoo Jae-won asked for. Now he granted them the freedom to act autonomously and proactively, provided they remained within legal boundaries. Autonomy, of course, meant they had to prove their own capabilities. Remington mentioned that Levin, the team leader, had become quite busy, and today he had delivered an extraordinary piece of intelligence: something concerning Apple.
At the mention of Apple, Steve Jobs was the first name that came to Yoo Jae-won’s mind. There had been that one time when a meeting had been arranged only to be unilaterally canceled. It had confirmed for him that while Jobs was indeed an innovator, he possessed a rather peculiar personality. There had been no news of Steve Jobs returning to Apple, and even cross-referencing his past memories confirmed that this was not yet the moment for such a return.
“In that case, Apple’s new product must be the first-generation Power Mac, correct?”
According to the timeline within his Memory Palace, the likelihood that Apple’s upcoming release was the Power Mac exceeded ninety percent. The report Levin had sent reinforced this assessment, complete with compelling evidence suggesting Apple had signed a supply contract with the PowerPC camp.
PowerPC was the microprocessor Apple had developed jointly with IBM and Motorola. Its foundation lay in IBM’s POWER microprocessor, originally designed for mainframe CPUs. Because the development costs and performance made it wasteful to restrict the chip to mainframes alone, IBM had sought other applications and eventually connected with Apple. At the time, Apple had been feeling the performance limitations of Motorola’s 68-series CPUs and was actively searching for a new processor. The industry had already shifted toward HPC, delivering performance gains of one or two hundred percent, leaving Apple in a desperate position. Yet Apple could not simply abandon Motorola, its longtime ally, in favor of Intel or AMD CPUs.
Intel and AMD enjoyed a relationship with Yoo Jae-won’s Android OS that went beyond mere compatibility—they were practically inseparable. The industry had taken to calling it the AIA Alliance, though the name was merely an acronym derived from Android, Intel, and AMD rather than any formal pact. Still, the partnership was undeniably close. Android actively supported the latest technologies from both companies, which in turn optimized their hardware for the operating system. This arrangement worked because over ninety percent of PC shipments were IBM-compatible systems and more than ninety-five percent of personal operating systems ran on Android. Apple’s market share had long since fallen below ten percent, and even its vaunted Mac OS was struggling to keep pace with Android.
Under these circumstances, adopting Intel or AMD CPUs would feel like a betrayal to Apple’s loyal fanbase. Fortunately for Apple, IBM’s POWER CPU happened to meet its requirements perfectly. By adding features suited for PC use, the PowerPC 6xx series was born.
“The upcoming PowerPC will see a significant performance boost over the original, won’t it?”
After all, copper interconnect technology existed. Yoo Jae-won had made the technology available to any semiconductor manufacturer willing to pay, all in the name of accelerating the IT revolution. Even Ilseong Electronics had purchased a license, so IBM and Motorola were certainly no exceptions.
“Let’s see.”
Refusing to rely on speculation alone, Yoo Jae-won immediately opened the directory containing the legal team’s reports and searched for HPC-related materials. He had meticulously archived every report he received on his hard drive, allowing him to retrieve information even offline.
“As expected, they’re here.”
Both IBM and Motorola appeared prominently on the list of license purchasers, and their acquisition dates predated even Ilseong Electronics’.
“Hmm. Still, I wonder how much the performance will actually improve.”
Copper interconnect technology was no magic wand. The degree of performance gain depended entirely on each company’s development capabilities. Even so, IBM remained a traditional powerhouse that warranted caution. Moreover, PowerPC was built upon the CPU architecture IBM proudly used in its mainframes, which meant exceptional performance. Its RISC instruction set enabled faster clock speeds, and a single physical CPU could handle as many as eight threads.
In the past, Apple had boasted that the Power Mac was two to three times faster than PCs using Intel’s latest CPUs. While the advantage did not appear across every application, it held true in specific scenarios, so the claim was not false advertising.
“Will they use the same ‘two to three times’ line again this time?”
As PowerPC gained speed through copper interconnects, so too did Intel and AMD. The performance gap would likely remain similar to before.
“Come to think of it, it’s about time we released a new product ourselves.”
There was little urgency in Yoo Jae-won’s voice. If PowerPC had been armed solely with its strengths, the market might have flipped overnight—but its weaknesses were equally formidable. The greatest drawback was its extraordinarily high development difficulty. Designing the chip was challenging enough, but creating software capable of fully extracting its designed performance was equally arduous. Higher clock speeds also meant greater power consumption and heat output. The decisive reason the Power Mac never achieved mainstream success was its exorbitant price, compounded by the fact that Apple’s existing management team remained unchanged.
Until last year, Apple’s CEO had been John Sculley, a Wharton MBA who had previously served as president of PepsiCo. A marketing genius, he had masterminded the blind taste-test campaign that pitted Coca-Cola against Pepsi. Apple’s executives had brought him in believing marketing was where they lagged behind the PC camp, yet that very decision ultimately drove Steve Jobs out of the company. Worse, Sculley’s marketing efforts had underdelivered while a string of strange failures accelerated Apple’s decline.
“The man actually had vision, surprisingly enough.”
Among Apple fans, John Sculley was often blamed as the architect of the company’s downfall. While not entirely wrong, Yoo Jae-won believed Sculley was not without merit. The first PDA, Newton, had emerged under his leadership, and he had also pushed tablet development. The problem was that his ambitions exceeded the technology available to realize them.
Apple’s current CEO was Michael Spindler. The upcoming Power Mac would be his first major product—and his last. Though he belonged to Apple’s inner circle and had delivered the substantial Power Mac, he would not remain in the CEO seat for long. Sales of the Power Mac proved disappointing, and the day of Steve Jobs’ triumphant return drew ever closer.
“Perhaps we should target professionals this time.”
The New Egg 2 had been designed for general consumers and continued to sell well. It boasted the most refined hardware among IBM-compatible PCs, maintaining steady demand from those seeking premium models. Even the highest-spec New Egg 2 would remain relevant for another year or two. Professionals, however, for whom time equaled money, were always willing to open their wallets for even marginal performance gains.
“Let’s take a look.”
Yoo Jae-won flipped open his Shell Book and brought up multiple documents simultaneously. The hard drive whirred so loudly that Kim Dae-seok, seated beside him, turned to look. Yoo Jae-won longed for SSDs, but they remained far out of reach for now. The documents that appeared with such clamor contained compiled lists and detailed specifications of engineering samples from CPU, RAM, and various hardware manufacturers. Without operating system support, even the newest hardware was little more than a paperweight. That was why it had become standard practice for Android to receive the latest engineering samples and detailed roadmaps from hardware makers.
“Pentium 3 is scheduled for release this summer, I see.”
Thanks to this arrangement, Yoo Jae-won could examine Intel’s latest product specifications and even its architectural details from the comfort of his seat. While the name itself had been disclosed to the public when the roadmap was announced at the beginning of the year, the detailed performance tables had never been shared externally. The Pentium 3 featured a base clock speed of 400 MHz, with flagship models reaching 600 MHz. It was not merely a clock speed increase. The chip incorporated new MMX instructions optimized for multimedia processing, and both L1 and L2 cache capacities had been doubled, delivering substantial performance gains over the Pentium 2. The report included not only verbal descriptions but also benchmark comparisons across games and professional applications. In Doom 2, the Pentium 3 showed roughly a thirty percent improvement, while 3D Max rendering performance exceeded seventy percent. The magnitude of these gains depended on how effectively the newly added MMX instructions were utilized. Doom 2 lacked an MMX patch, so its improvement reflected only the higher clock speed, whereas 3D Max benefited from Intel’s specially tuned drivers that actively leveraged the new instructions.
“It would be good to make full use of MMX.”
Even Doom 2 could see further gains if an MMX-compatible patch were provided. The same applied to Android itself—optimizing the OS could improve performance for applications no longer receiving updates.
“And AMD’s offering is the Athlon, I suppose?”
Originally, the chip had been slated to launch under the K7 name. However, after Intel abandoned numeric branding following the 486 in favor of the Pentium brand, AMD had decided to move beyond the K-series and introduce a new brand called Athlon. AMD was confident the Athlon would start at 500 MHz, with flagship models reaching 700 MHz.
“They’ve sharpened their knives properly.”
The attached package photographs showed a large, imposing cooler. While Intel’s solution was little more than a palm-sized heatsink, the Athlon’s cooler was several times larger and made of solid copper. Only the 700 MHz flagship received the full copper cooler; the standard 500 MHz model used an aluminum base with a copper core where it contacted the CPU.
“My, they’ve done proper research on cooling as well.”
It was clear that Terry Micron, AMD’s CTO who had personally delivered the first K6 prototype to Yoo Jae-won’s residence, had been deeply impressed by the cooler he had received as a gift. Unlike architectural or material improvements, a simple enhancement to the cooling system could yield significant performance gains—a revelation that had opened an entirely new world for AMD. Particularly since AMD’s per-clock performance lagged behind Intel’s, the company had invested heavily in raising clock speeds. Higher speeds generated more heat, which the company addressed with an overwhelming cooler, enabling the first-generation Athlon to reach 700 MHz—faster than Intel’s offerings. While the solid copper cooler increased unit costs and reduced margins, AMD had deemed gaining market share the higher priority and boldly created a high-value package.
“No sign of Cyrix?”
A third CPU manufacturer, Cyrix, existed, yet its name did not appear in the new product specification documents sent to Yoo Jae-won. As the industry entered the HPC era, the CPU market was consolidating into a duopoly between Intel and AMD, and Cyrix appeared to be falling behind. Yoo Jae-won easily identified the cause: America’s economic boom had improved consumers’ financial situations, leading more people to choose higher-performance models even at premium prices rather than settling for cheaper, lower-performance options. Sales figures for the Egg series computers and 3D accelerator cards confirmed that mid-range products were selling far better than budget models, while even high-end configurations—once niche—had seen substantial growth.
“Well, we’ll have to go with Intel for the CPU.”
The computers Yoo Jae-won designed always used only the best components. Although the upcoming high-end Athlon boasted 700 MHz, its per-clock processing capability fell short of the Pentium 3. Consequently, benchmark numbers favored the Pentium 3 600 MHz model. Intel’s newly added instructions were also superior. AMD could use Intel’s MMX instructions thanks to cross-licensing and had added its own extensions, branding them 3DNow!. These additional instructions were specialized for 3D processing. The problem was that they shared the same registers as MMX instructions, preventing simultaneous use of both. Furthermore, although AMD had obtained MMX technology from Intel, architectural differences meant Intel’s implementation performed better even with identical instructions. Cooling presented another decisive issue. Yoo Jae-won intended to place great emphasis on design for the new PC that would succeed the New Egg 2. Excessive heat would detract from both aesthetics and stability. Using the Pentium 3 paired with the solid copper cooler from the high-end Athlon would be far preferable. Since the computers he designed were never about cost-performance ratios anyway, aiming for the absolute best made sense.
“For memory, we’ll naturally use Mirae Electronics’ Super Green memory.”
Super Green memory was not fundamentally different from conventional DRAM. Mirae Electronics had created the brand to distinguish itself from Ilseong Electronics. Its defining characteristics were eco-friendliness and high performance. Memory chips from Ilseong Electronics were so-called tray products—shipped in bulk trays of one hundred units without individual packaging. When sold individually to consumers, these chips were simply wrapped in aluminum foil. In contrast, Mirae Electronics’ Super Green memory offered not only tray products for corporate supply but also retail packaging in single, dual, and quad configurations. While the most visible difference was the clean green boxes, the chips themselves were carefully selected for superior overclocking performance. Pre-HPC computers had matched CPU and memory clock speeds one-to-one. Once HPC pushed CPUs to higher speeds, memory could no longer keep pace, creating a bottleneck. Overclocking memory delivered noticeable performance improvements. Not every motherboard supported memory overclocking—only high-end boards did. Buyers of premium boards were precisely the enthusiasts seeking every possible performance edge, so Super Green memory, released already overclocked, had generated enthusiastic responses among computer hobbyists. The “eco-friendly” keyword also resonated with the times. Since the early 1990s, environmental pollution had become a serious issue in the United States. Semiconductor manufacturing’s use of toxic substances known to cause cancer had drawn particular scrutiny. Mirae Electronics had minimized the use of such substances in its production processes and designed its facilities to prevent human exposure. While this increased construction costs, the company’s appeal to environmental values had earned positive market reception. Ilseong Electronics had also begun emphasizing eco-friendliness, though achieving it without overhauling its entire factory would be impossible.
The final item Yoo Jae-won examined was the 3D accelerator card. VGA cards that supported only 2D no longer existed. Even the cheapest models now included 3D acceleration chips. 3D had become so dominant that supporting it—however poorly—was essential for sales. This shift had triggered a major upheaval in the VGA market. ATI had been the traditional leader during the 2D era, but in the 3D age, two newcomers—NVIDIA and 3dfx—were growing rapidly. Both companies focused exclusively on producing powerful 3D acceleration chips, which they sold to VGA manufacturers. As a result, numerous manufacturers were churning out 3D accelerator cards using their chips, and sales were exploding. Yoo Jae-won had already invested in both companies through ID Investment when they began formal operations. Prioritizing performance, he decided to use NVIDIA’s upcoming Riva TNT, scheduled for release that fall.
“Still, it feels like high performance and design alone have become rather familiar by now.”
Building a computer armed with top-tier performance was no longer difficult. Thanks to the sensation created by the Egg series, more companies were paying attention to design. To maintain overwhelming superiority from this point forward, something additional—a true plus alpha—would be necessary. Ordinarily, such strategic judgments would come from the group’s research institutes or professional consulting firms, but Yoo Jae-won had no need for that. He had already arrived at the answer for the plus alpha.
“It’s time to bring up the word ‘intelligent.’”
“Intelligent” meant smart, and in computing it was typically represented by the prefix “i.” The letter also evoked “internet,” making it doubly appealing. Attaching “i” to Egg felt somewhat awkward, however, so launching an entirely new brand seemed preferable. Yoo Jae-won’s mind brimmed with ideas for intelligent features worthy of the “i” prefix.
“Now, when would be the best time to launch?”
Yoo Jae-won pulled up a 1994 calendar on his Shell Book screen and considered favorable dates.
“Two or three months after the Power Mac launches would be ideal.”
Levin’s report suggested a May or June release. Therefore, Yoo Jae-won’s new computer should launch in August. The timing was simple: the first-generation Power Mac would almost certainly fail to generate meaningful market response. If ID Group’s next-generation computer, armed with intelligent features, launched while the Power Mac was being criticized, the contrast would be striking. The question was whether a targeted product could be prepared so quickly, but for ID Group this posed no problem. All components were already prepared; proper assembly would resolve the hardware side. The real challenge lay in developing innovative software worthy of the word “intelligent,” yet that too was within Yoo Jae-won’s capabilities.
“Heh. This is perfect. Though it does feel rather villainous.”
Yoo Jae-won chuckled, then paused for a moment of self-reflection. Still, unless Apple demonstrated a genuinely different level of innovation, its decline was inevitable. And if it was going to fall, it might as well fall spectacularly. Only then would Apple’s management come to their senses and bring Steve Jobs back sooner. Yoo Jae-won simply wanted to face a worthy opponent. Given how much the IT landscape had changed, there was a real possibility that Steve Jobs’ perception and abilities had upgraded enough to produce something beyond Yoo Jae-won’s expectations. If that challenge spurred ID Group to create even better products, then this would be competition in its most desirable form.
