Executive-level risk management meeting.
Department Head O convened only ten director-level people and held the meeting in secret. It was because I had started spouting insane nonsense again.
My sell report was hard enough to accept as it was, but on top of that, we had the added assignment of having to make a decision on it within a week.
Because of that, even the division head, who had an overseas schedule, had no choice but to cancel it in a hurry and be dragged here.
“Good. I think the explanation is over, so now speak freely and share your opinions.”
After listening to the presentation that had gone on for about an hour, the division head said that.
While everyone was merely blinking, Director Choe of the overseas team sighed and said to me,
“Does that make any sense...?”
Starting with him, voices of concern poured out.
-Team Leader Lee, even to me, this makes no sense...
-A total liquidation of Envidi...
-This is the age of AI. Of course a company selling jeans and pickaxes is bound to soar...
I knew exactly how they felt.
Envidi’s semiconductors were in a class of their own. With a 94% share of the GPU market, they could be said to have effectively monopolized the pickaxes in this AI gold rush.
In other words, unless the age of AI came to an end, they couldn’t fail.
“Impossible... absolutely impossible. Even the United States has to bow before Envidi chips, and another country is going to surpass this level of technology? That is absolutely impossible.”
Pushing past Director Choe’s adamant opposition, I emphasized it again.
“They can. It turns out there isn’t no way at all.”
“No, what’s your basis?”
“From what I’ve heard through several routes, DeepSync has succeeded in developing its own AI. This model will be announced to the world in a week.”
“Haa... fine, then let me ask you just two things.”
Director Choe, his face flushed, pressed me.
“First. Where did you hear that information?”
“I can’t reveal the details, but... from someone quite trustworthy.”
“It’s not as if this is some public-interest whistleblowing. Is there a reason you can’t even disclose it to us?”
“Director...”
“Fine. Let’s say that’s true. You’ve always heard information we couldn’t access, and your market predictions have always been right. But listen. If there really is such high-level information, how is it that only you know about it?”
“...”
“If what you’re saying is true, this is an enormous shock that could overturn the entire AI landscape. I doubt Wall Street sources are worse than you, so why is the market reaction this quiet? Can there be data in this world that only you know?”
I smiled bitterly.
In truth, a shocking announcement of this magnitude was bound to have warning signs. For instance, a competing company might have succeeded in a technological leap, or blueprints might have been leaked—there was usually some kind of indication.
But this was not even a competing company. It was just a startup. For such a company to have surpassed the technology of the big tech firms could only sound absurd.
So I decided to be shameless.
“Yes. It’s top-secret material that only I know and Wall Street does not.”
“Wh-what?”
“I heard it through our external manager, Changcheng Fund. They spend enormous amounts not only on investments, but also on information that might help their investments. They said it was data they barely obtained by lobbying high-ranking officials inside and outside the Communist Party.”
Fortunately, it worked extremely well.
At present, Changcheng Fund was posting by far the most overwhelming returns among our external managers, and when I passed it off as information they had obtained through lobbying, no one could refute me.
Seizing the momentum, I presented Changcheng Fund’s recent liquidation of Envidi and its short position as supporting evidence.
“This makes no... No, fine. If it was information control by the Chinese authorities, then Wall Street might not know. Then let me ask just one last thing.”
“Yes.”
“Does developing AI with 8 billion even make sense...? If that’s true, shouldn’t we be selling not only Envidi, but every Nasdaq stock?”
The conference room sank into a long silence.
In truth, this was the most fundamental fear. If AI development was a technology that could be reached so easily... it meant the entire Nasdaq was a bubble. But how could anyone possibly accept that?
“That is not the case.”
I spoke calmly.
“In the early days of the Industrial Revolution, no matter how much Britain ran its factories, it couldn’t beat China’s household handicraft industry. During the transition period, the steam engine couldn’t exert its power. I think it’s the same now. AI is still in its infancy. It may be possible to catch up to the current level of technology through some entirely unforeseen method, but in the end, the more advanced the AI business becomes, the more demand there will be for advanced chips. By then, it will reach a level of technology that household handicraft cannot keep up with.”
If my memory was right, that was how it had been.
If you raised a few fine horses, they might be faster than early automobiles. But once automobile technology advanced into the age of airplanes and fighter jets, what ultimately became most important was engine technology.
In the future I knew, China would pursue closely, but in the end, the United States would lead AI hegemony. China knew this too, and later poured its efforts into developing advanced computing chips.
“Division Head. May I say something?”
When everyone was wrapped in silence, Department Head O opened his mouth.
“Go ahead.”
“In truth, whether Team Leader Lee’s words are right or wrong is not what’s important here. The MIT paper by DeepSync’s CEO has already been released to the public.”
“This is publicly disclosed technology?”
“Yes. It’s open AI technology, so anyone can make it as long as they follow the code. I also checked the contents of the paper before opening this meeting.”
One month before the model announcement, DeepSync CEO Liang Wen had disclosed his technology in an MIT paper. It meant anyone was welcome to verify it. It was interpreted as a sign of immense confidence in the result.
“Then does Department Head O believe we should respond as Team Leader Lee says?”
“Yes. I don’t know what will happen in the long term, but in the short term, it was a level of technology that absolutely cannot be overlooked. For now, I think we should liquidate our stake and stay on the sidelines until DeepSync’s announcement. And...”
“And?”
“In truth, today’s meeting wasn’t held simply because of the matter of liquidating Envidi. There is an even more important issue.”
“...There’s an issue more important than that?”
“Yes. The domestic stock market due to AI is an even bigger concern.”
Department Head O said that and nodded toward me. It meant I should say it myself.
“Team Leader Lee, fine. You explain it yourself. What else did you say to Department Head O?”
“...To start with the conclusion, I think Korea needs to grow its AI market as well.”
“What?”
“You’ll know once it’s announced in a week, but the DeepSync model will probably not be very different compared to existing models like Gemini and GPT. In other words, it’s a level of technology that latecomers can sufficiently challenge.”
“Even so, how are we supposed to grow that? We’re an investment firm that invests in those with potential. That’s not our role.”
“But it will probably have a major impact on the domestic market overall as well. Our stocks aren’t safe either.”
I smiled bitterly.
If something rises, something else is bound to fall. The more firmly the AI mainstream theory took hold, the more several domestic businesses grew endangered. First, the platform market, which could be called a direct target of AI, was shattered.
In the 2010s, Goggle’s search engine share in Korea was only 5%, but ten years later, it had leapt into the 30% range.
Many experts were saying in unison that if it gained the wings called AI, Goggle’s domestic market share would rise even higher.
Perhaps because this outlook had been reflected, the stock price of Korea’s top platform, Naive, had been hitting the ground day after day.
Recently, the KOSPI had surged about 50%, making it an era where anything rose if you bought it, but Naive’s stock price remained at its price from five years ago. Compared to the peak it reached during COVID, it was still cut in half.
I can see it... I keep seeing Cyworld in them.
“Semiconductors are not safe either. At present, Hoam and Seongyeong are continuing a symbiotic relationship by supplying various products to big tech companies, but... if the prime contractors become technologically independent or switch subcontractors, they could collapse in an instant.”
“So?”
“From what I’ve heard recently from officials inside and outside the government, the government also has great interest in AI development.”
“Then they should just invest through R&D. Don’t tell me they’re going to take our seed capital again?”
“This time, I think it would be better for us to invest more aggressively.”
“What?”
“If Korea misses AI hegemony anyway, most of the domestic stocks we hold will walk the path toward delisting. If that’s the case, it would be better to take a shot.”
“...What are you suggesting we do?”
“If my memory is right... No. After DeepSync’s announcement, I’m sure the government will also come out with measures to foster AI. I think we should prepare as much seed capital as possible and move in step with the government’s actions.”
The division head quietly turned his head.
But even the directors who opposed liquidating Envidi wore the same expressions. They all agreed that if Korea missed AI hegemony, domestic industry would lose its competitiveness.
The division head quietly let out a sigh, then asked me,
“Then how much seed capital do you want us to invest?”
I stated the amount I had in mind without the slightest hesitation.
-No, that’s absurd!
-Team Leader Lee, that’s way too much!
-If we invest that money and fail, the pension fund will go bankrupt before domestic industry does!
The amount was somewhat large.
The directors who had briefly gone along with my view now glared and rose up in protest.
*
“Allow me to introduce our AI model, developed over ten years. Its name is DeepSync! In hopes that it will swim freely in the sea of information, the logo is shaped like a very cute whale. Hahaha!”
One week later, the DeepSync model that was announced shocked global society.
CEO Liang Wen, who had remained calm throughout, could not hide his excitement as the model reveal drew near.
The model he presented was more than enough to display his confidence. DeepSync, developed with 8 billion, had outstanding performance.
In the American mathematics competition test, it scored higher than GPT. In coding as well, it recorded a better score than GPT.
“If possible, we would like to aim for a high score on the American SAT within the year as well. I’d like to test how well our DeepSync can handle written-answer questions too.”
Major foreign media outlets were left speechless each time DeepSync’s capabilities were revealed one by one.
All of this development cost only 8 billion...
DeepSync had completed, with 8 billion, an LLM model that Meta’s LLaMA had completed by investing 89 trillion won, or 65 billion dollars. The difference between 80 trillion and 8 billion was jaw-dropping.
DeepSync’s announcement ended after roughly two hours, and the resulting global repercussions only amplified as time passed.
Netizens all over the world rushed in to doubt its technology and bombard it with nasty questions, but... unfortunately, DeepSync answered every question without difficulty except for Tiananmen and President Xi. Some of its answers surpassed GPT and Gemini.
The suspicious gazes that had poured in so fiercely gradually turned into shock as time passed.
[Breaking News - Envidi Stock Opens Down 18%!]
[Broadcom, Marvell, TSMC, and Others Also Plunge in the 15% Range!]
And this became an enormous nuclear bomb in the Nasdaq market that opened eight hours later.