The Dawn of Integrated Systems and Artificial Intelligence
Whether they were standing face-to-face with Yoo Jae-won at the ID Technology headquarters in this 21st-century technopia or watching the New Year’s address from afar through the company broadcast, everyone had the same question mark floating in their heads.
System integration? The words were simple enough, yet grasping their true meaning proved surprisingly difficult. Still, the integration Yoo Jae-won spoke of wasn’t some complicated concept. One only needed to take the words at face value. It meant placing every internet service and every business process of the entire ID Group onto a single server system.
Naturally, the group currently operated on multiple separate servers. Each subsidiary had its own unique methods for handling tasks and computerization, which was exactly why he had brought the topic up. Even the ERP system introduced several years ago saw wildly different levels of adoption across subsidiaries. Email and ID Talk had allowed them to manage digital processes without major inconvenience so far, but Yoo Jae-won judged that they had now reached their limit.
“By the end of this year, every service and every business processing system within ID Group will be perfectly integrated into one server and one system.”
He hadn’t said they should integrate. He had declared they would. Discussions with the presidents of each subsidiary had already concluded. Integration would happen without fail before the year was out. The reasons were straightforward: strengthened communication between subsidiaries, the ability to grasp the group’s entire pool of resources at a glance, and the capacity to execute massive strategies with far greater efficiency. Of course, there were countless other advantages he hadn’t bothered to list.
This kind of integrated system had become an urgent necessity after the launch of the Xbox project. Yoo Jae-won had formed a dedicated task force for the Xbox and pulled specialists from various departments to build the organization. Although it was structurally a relatively simple video game console, mass-producing and distributing an entirely new system was an enormous undertaking.
What Yoo Jae-won had issued as a single casual order required countless procedures at the working level. For example, the DVD-ROMs installed in the Xbox were currently manufactured by the home appliance division of ID Electronics. While the subsidiaries were united under the ID Group banner, in practice they were still practically strangers. Ordering components therefore involved the same complex paperwork as dealing with an outside company.
Time was life in this project. He had already driven a stake into the ground by promising to unveil the system at this year’s E3 show. Only six months remained. Fortunately, the Xbox hardware design had been completed long ago, and the final design was finished as well. The custom Android operating system built exclusively for gaming had turned out compact and excellent. In particular, the 3D library GlideX had received additional functions tailored for the Xbox and been meticulously tuned to match the custom hardware.
The reason GlideX had been separated was twofold: to improve gaming performance and to prevent Xbox games from running on PCs without any restrictions. The greatest weakness of the PC platform was software piracy. Thanks to the rise of online gaming, illegal copying had decreased somewhat, but it had not been eradicated. Moreover, console games tended to be heavily single-player, making them especially vulnerable to piracy. The fact that Sony had used easy piracy as their justification for leaving the Android camp said everything that needed to be said.
That was why Yoo Jae-won had deliberately split GlideX itself into PC and Xbox versions, hoping to reduce piracy. In any case, working on the Xbox project had made him acutely aware of the need for system integration, and he had made the firm decision to complete that integration this year.
He did acknowledge that integration carried disadvantages. The most critical one was that if a single security system was breached, everything would be breached. A system failure could potentially paralyze the entire group. However, the chance of such a catastrophic problem actually occurring was extremely slim.
The internet industry’s number one company, ID Group, faced constant attacks from every kind of hacker imaginable. The problem was especially severe for email.com, which served as the unified ID system. Yet not once had it ever been breached. Yoo Jae-won himself was no ordinary talent when it came to computers. On top of that, the Kaspersky team he had assembled—composed of the world’s top security experts—was practically a dream team.
Besides, the sooner system integration happened, the better. The longer they waited, the higher the integration costs would climb, and the more employees became entrenched in outdated systems, the harder it would be to fix them.
There was another reason integration was necessary.
“We will also be making massive investments into artificial intelligence.”
At the mention of artificial intelligence, the people gathered in the main auditorium of ID Technology headquarters nodded with understanding. Their reaction was completely different from when the topic of system integration had come up. That was only natural. Artificial intelligence had been an important issue in Silicon Valley for quite a long time. Serious research had begun around the same time computers first appeared.
The greatest strength of computers was their ability to perform repetitive tasks at high speed. Human greed did not stop there. People wanted computers to handle even more work, which naturally led to research into artificial intelligence. However, there had still been no meaningful results. Among the countless venture companies in Silicon Valley, several were studying AI, but they remained at the foundational stage, and commercialization was nowhere in sight.
It was in this environment that Yoo Jae-won had brought up artificial intelligence. For those who loved computers, AI was naturally an object of fascination, so the eyes of the ID Technology headquarters employees gathered in the auditorium sparkled with excitement.
“As those in the know are already aware, our ID Group has already completed two artificial intelligence algorithms.”
The bomb Yoo Jae-won dropped next sent the ID Technology staff into a frenzy. They turned to one another, rapidly exchanging whispers of “Us?” and “When?” Although the employees had been busy creating the various programs and products Yoo Jae-won demanded, not a single one of them had ever heard anything about research into artificial intelligence. Even the Hi-Tech Research Lab had been completely in the dark.
“The profanity filtering system and the image search algorithm. When you look closely, both are technologies that utilize artificial intelligence.”
Yoo Jae-won revealed one of the ID Group’s long-held secrets. The profanity filtering system had been running since 1992, when 2CH.com was first created. Knowing that the biggest drawback of an anonymous community was the outpouring of base emotions—especially profanity—Yoo Jae-won had created a monitoring program under the pretext of filtering curses. When users posted profanity in articles or comments, it automatically replaced the text with ‘***’.
By now the profanity filtering system had been applied across all of ID Group’s internet services, including the newly launched TokTok the previous year.
“Does anyone here know how the profanity filtering algorithm works?” Yoo Jae-won asked the employees in the auditorium.
A murmur immediately rose from every corner.
Negative blocking was a simple method: create a database of keywords used in profanity, then block any text that matched those words. It was convenient, reliable, and didn’t consume much system resources.
The rebuttal came immediately. The voice was loud and received plenty of support, showing just how high everyone’s expectations of Yoo Jae-won were. Still, negative blocking wasn’t exactly wrong.
“Negative blocking. That is correct,” Yoo Jae-won said, causing the murmurs to grow even louder. After all, negative blocking was quite far removed from what people thought of as artificial intelligence.
As if he had expected their reaction, Yoo Jae-won smiled faintly and continued.
“In negative blocking, the most important factor is keyword management. Those of you who have explored internet battles know this well—people come up with truly unimaginable combinations to express negative emotions. Our ID Group’s filtering system boasts an iron wall in this regard. Is it thanks to the dedication of the hundreds of monitoring staff members? No. The team responsible for filtering in North America numbers fewer than one hundred people.”
One hundred people might sound like a lot, but they didn’t all work at the same time. Following the basic eight-hour shift policy, they worked in rotations, and hours could be adjusted flexibly according to personal circumstances. Therefore, the number of people on duty at any given time was around forty.
The management of blocking keywords was handled by machine learning, which had grown sophisticated enough to be called artificial intelligence. After more than seven years of silent learning, it would not be an exaggeration to say it was now more familiar with internet slang than even the named netizens of 2CH.com.
“There is one more. Image search.”
Image search also had a strong connection to artificial intelligence. If you clicked the image search button on the NextCom search bar and uploaded an image or entered its URL, results would pour in. What had once seemed strange to netizens had now become a frequently used feature. From the names of weeds to historical figures no one could remember, everything could be found easily.
Of course, netizens were using it in ways quite different from Yoo Jae-won’s original intentions—mainly to find adult videos. When they couldn’t remember an actress’s name, a quick image search would locate even those with relatively low recognition.
“For image search as well, the database is crucial. At first a person could manage it manually, but once the data volume becomes vast, that becomes impossible. So we replaced it with mechanical learning, and the results have exceeded expectations.”
Yoo Jae-won spoke as if he were officially introducing artificial intelligence because of an unexpectedly positive response. This was despite the fact that the algorithm using AI for image search had been the highly praised Golden Code even in his previous life.
“Some of you may be thinking, ‘What’s so grand about artificial intelligence for mere image search?’ However, this algorithm’s applicability is remarkably versatile.”
Once this vision presentation ended, dozens of media outlets would undoubtedly claim that it was still too early for artificial intelligence. Yoo Jae-won was kindly opening his bag of tricks for their sake.
“Let me give a direct example. What would happen if image interpretation artificial intelligence were combined with radiology—the academic field currently receiving the most attention in medicine?”
Radiology was the discipline that used equipment such as X-rays, ultrasound, CT, and MRI to extract images, then diagnose diseases, determine treatment methods, and monitor post-treatment progress. Reading and interpreting those images was a human task that required a high level of training for accurate judgment. Nevertheless, misdiagnoses still occurred frequently, and for patients, a misdiagnosis could be fatal.
If artificial intelligence were introduced to assist with image interpretation and diagnosis, it would bring tremendous progress to medical development. Patients would avoid the slim chance of misdiagnosis, doctors could escape from overwhelming workloads, and ID Group would gain a new source of revenue. Of course, there were prerequisites. The skepticism about whether the AI’s analysis could be trusted would need to be completely resolved.
In his previous life, Google had staged an event of the century to prove that their artificial intelligence was excellent and trustworthy. The first attempt had been Go. From there, they had to take a long and painful detour. Contrary to internal predictions of an overwhelming victory, the AI had suffered one loss. It was a match that proved the phrase “divine move”—which had once been nothing more than a simple idiom—truly existed.
Afterward, Google had to invest nearly fifteen more years to further advance their artificial intelligence. In the end, AlphaGo’s failure had brought Google an enormous benefit. Because they worked relentlessly to overcome that failure, they succeeded in creating artificial intelligence that surpassed the technological singularity faster than any other company.
Yoo Jae-won’s ultimate destination was also the creation of artificial intelligence that had crossed the singularity. That was why he had chosen the first vision declaration of the new millennium to plant the seeds for AI early on. It might have been too soon. While consumers could quickly experience SNS through the TokTok service, artificial intelligence would, for the time being, only manifest through the filtering system and image search.
But that didn’t matter. Yoo Jae-won wasn’t thinking about commercial applications of AI right now. It was enough that he had made massive future investments in artificial intelligence seem natural and reasonable for the group. For now, combining AI with the newly integrated system to extract various forms of data that humans could never provide was sufficient.
The next day.
As expected, Yoo Jae-won was the rock star of Silicon Valley. Less than a day after the vision presentation, related articles poured in.
Artificial intelligence suddenly shot to the center of public attention. Along with it, the number of people visiting 2CH.com just to test the profanity filter or trying out NextCom’s image search function skyrocketed. Among companies listed on Nasdaq, those conducting research related to artificial intelligence even saw their stock prices rise.
Having turned Silicon Valley upside down, Yoo Jae-won did not immediately launch an artificial intelligence business division. Looking at his major plans for the year 2000, April would see the Korean general election, June would bring the Xbox announcement at E3, July was his own wedding, and the year-end target was the completion of system integration. And one task that could not be overlooked was his involvement in the American presidential election.
Although eleven months remained until election day, the American political world was already in full campaign mode. Having chosen Al Gore after deep consideration, Yoo Jae-won was determined to thoroughly prevent Bush from becoming president of the United States, just as he had in the past. That work had already begun.
After all, the man who had delivered the 2000 vision declaration yesterday was now aboard his private jet, flying toward the East Coast.