These days, Aaron Stark often found himself tossing and turning at night.
In the pitch-black darkness, every ti he lay in bed staring at the ceiling, one question kept circling in his mind.
‘Should I just start an AI company myself…?’
In truth, his heart was already leaning toward “yes,” but for so reason, he hesitated to make a final decision.
There were two big reasons for this hesitation.
First, ‘Now is not the ti.’
His hands were already full with Teslaan, Space Z, SolCity, and more.
Even 24 hours in a day weren't enough as it was; adding another company would be too risky.
And the second reason was...
‘It’s not my idea.’
According to Stark’s original plan, he was supposed to join Next AI and rework their system in his own style.
Starting from scratch would take more ti and resources than remodeling sothing that already existed.
But now, that entire plan had veered far off its original course.
And the one who twisted it all wasn’t Stark himself.
It was Ha Si-heon.
‘The idea itself is attractive, but…’
His pride wouldn’t allow him to just swallow it whole.
‘I don’t have a follower’s temperant.’
By nature, Stark had a deep-seated resistance to following other people’s directions.
No matter how good the proposal was, he would instinctively feel a sense of resistance.
This ti was no different.
His resistance toward Ha Si-heon’s proposal gradually turned into outright aversion, and then into suspicion.
‘What if all of this was part of Ha Si-heon’s script?’
What if Ha Si-heon had deliberately blocked him from joining Next AI and guided him to start a separate AI business instead?
Of course, the chances were slim.
It was Stark who had first approached Ha Si-heon, and the idea Ha Si-heon had suggested had co up spontaneously during their conversation.
But still… it had to be that way… and yet…
His instincts kept sounding an alarm.
‘There’s sothing here.’
That unpleasant feeling he had when talking to Ha Si-heon.
The odd sensation that he wasn’t in control.
Usually, Stark was the one who dominated the conversation.
Whoever the opponent was, if he pushed hard enough, they would eventually follow his lead.
That’s why so many people called him the “Bull.”
But Ha Si-heon didn’t follow.
On the contrary, it was the exact opposite.
Ha Si-heon had never used a strong word.
A gentle tone, a subtle smile, a soft suggestion.
But when Stark looked back after their conversations, he always realized that his thinking had shifted.
‘This is like…’
Suddenly, a bullfighting match popped into his mind.
A charging bull losing its direction after being lured by a small cloth.
And just at that mont, Ha Si-heon reached out to et him.
‘Maybe…’
It might be his chance to confirm things.
Whether Ha Si-heon’s suggestion to “start an AI company yourself” had really been a coincidence, or if it was a carefully disguised intention.
So, “Only seven minutes left.”
Stark pulled out a tactic he was very familiar with.
A ti limit.
By doing this, they would have no ti for small talk and would have to get straight to the point.
In other words, Ha Si-heon wouldn’t be able to use the excuse of having “co up with it on the spot.”
Moreover…
This way, Stark could firmly hold the reins of the conversation.
“Seven minutes, huh…”
“Now there are only six minutes and thirty seconds left.”
Stark kept reminding him of the ti, creating pressure.
And inside, he smiled.
Ha Si-heon would surely bring up the AI topic.
And he would try to casually throw out the suggestion as if it was nothing.
But with only six minutes left, he would naturally feel rushed and make mistakes, and in that process, his true intentions would slip out.
Or so Stark thought.
But the response he got was completely unexpected.
“I recently read an article about the Eagle launch and was concerned that you’re too fixated on the rocket’s guidance system. There were even articles comparing it to a ‘ballet in the sky’…”
Out of nowhere, he started talking about reusable rockets.
“Is that… what you ca here to talk about?”
He had clearly expected an AI discussion…
But Ha Si-heon just shrugged and calmly replied.
“No. But I wouldn’t have finished what I wanted to say within six minutes anyway, so let’s talk about it next ti.”
Ha Si-heon effortlessly slipped out of the trap Stark had set with the ti limit and continued.
“Controlling a rocket flying at thousands of kiloters per hour at an altitude of several kiloters can’t be solved with just simple calculations or sensor values. You have to consider gravity, air resistance, engine thrust, even the shifting center of mass as fuel is consud—all in real ti. So I can understand why you’re obsessed with the aerial maneuvering aspect…”
Ha Si-heon added with a wry smile.
“For a gymnast, fancy moves in the air are important, but what’s even more crucial is the landing. In the case of the Eagle, compared to its state-of-the-art guidance system, its landing system is practically like a folding chair.”
It was an annoyingly accurate point.
In truth, Stark had been so focused on the guidance and navigation systems that he had relatively neglected the landing system.
Ha Si-heon followed up with other advice, but Stark hardly heard it.
‘Is he really not going to ntion it at all?’
Before he knew it, there were only two minutes left.
At this point, it was too late to bring up the main topic.
Just then—
“Oh, by the way, I heard so interesting news. Apparently, Gooble’s AlphaGo secretly played a match against a European Go champion and won decisively.”
That was what Ha Si-heon said the mont the tir ended.
“It’s still top secret, but I heard it from one of the startups I invested in. A college friend working at Gooble leaked it to them, and they told not to tell anyone outside. But I thought you might find it interesting…”
Stark’s eyes widened.
‘AI beat a human at Go?’
That wasn’t just a technical advancent.
It was a paradigm shift.
He had thought that level of AI wouldn’t be possible for another few years…
He desperately wanted to hear more.
But Ha Si-heon glanced at the wall clock and opened his mouth.
“Ti’s up. Well then, I wish you a smooth launch.”
The smile at the corner of his lips was deeply irritating.
‘As I thought…’
He felt like the bull in a bullring.
But he couldn’t just not hear the rest of it.
In the end, Stark quickly stepped over to his desk, turned on the speakerphone, and gave instructions to his secretary.
“Tell the team the eting will be delayed by 30 minutes.”
Then he turned back to Ha Si-heon, who raised his hand awkwardly with an embarrassed look.
“You really didn’t have to do that because of . I know you’re busy…”
His tone was polite, and his expression showed genuine concern.
But sohow… that made him feel even more insincere.
Stark forced a smile as he continued speaking.
“Please, continue talking about Go.”
“In fact, that’s all I know.”
He had gone out of his way to make ti to hear more, only to learn that what Ha Si-heon had already said was everything?
Just when Stark thought he had been played, Ha Si-heon continued with a serious expression.
“But if this is true… it’s a pretty serious matter. AlphaGo’s victory ans that AI has broken through a threshold it had never crossed before.”
That was true.
Until now, AI had been stuck at “rule-based thinking” and couldn’t even begin to emulate things like intuition, inference, or contextual understanding.
That’s why AI could sowhat handle gas like chess or poker, which were based on numbers, but it was hopeless at Go, which required intuition.
However, “Deep learning overca that limit through reinforcent learning. And if it even managed to beat a human opponent…”
It ant that technology previously thought to be only in the research stage was already effective in real-world settings.
A chill ran down Stark’s spine.
“Once the effectiveness of reinforcent learning is proven, AI developnt will no longer follow a linear path but will accelerate exponentially.”
What Ha Si-heon was trying to say was clear.
“To catch up with Gooble, this is the last chance.”
They had to act imdiately.
“With reinforcent learning, it’s not simply about having more data. You need to build a superior model faster based on that data, then use that model to acquire even better data and further strengthen learning. The gap widens with every cycle. If you start a year late, you’re not chasing ‘just one year’—you’re trying to close an exponentially growing gap.”
That was true.
A gap that opened now would quickly beco impossible to close.
Ha Si-heon then presented a concrete strategy to catch up to Gooble.
“At this point, it’s too late to start developnt from scratch. It’s better to acquire an existing company in the industry and imdiately scale it up. You need to integrate computing infrastructure, model architecture, and data feedback loops to enhance overall capabilities.”
There was no choice but to buy up and rge multiple deep learning companies.
Ha Si-heon even nad specific acquisition targets.
“There’s a company called ta Minds. They don’t just do simple natural language processing; they focus on core structures for ‘understanding,’ like multitask learning, domain transfer learning, and multimodal fusion.”
He raised another finger and continued.
“The second is Vicari. While ta Minds focuses on implenting understanding, Vicari concentrates on ‘pattern recognition and reasoning.’ Whereas traditional approaches extracted statistical patterns from tens of millions of images, Vicari can infer and generalize from small amounts of data. It aims to build an AI that sees and judges like a human—essentially a kind of ‘sensory organ.’…”
The companies Ha Si-heon suggested were all ones Stark had already been watching closely.
But at that mont, Stark felt a strange unease.
‘He’s trying to use .’
This ti, he was certain.
Ha Si-heon was emphasizing Gooble’s achievents to create a sense of urgency, pushing him to acquire companies, and then driving him straight into the center of the AI battleground.
However, stark was not soone who would simply be “used.”
He looked Ha Si-heon straight in the eyes and asked, “If it’s that important, why don’t you step up and do it yourself?”
It was a question that hit right at the heart.
Surely, Ha Si-heon would be flustered or at least pause for a beat…
“Ah, but are you really okay with that?”
Ha Si-heon’s reaction once again defied expectations.
He wasn’t flustered at all.
On the contrary, his expression and tone almost seed to say, “This is perfect!”
“I actually did consider that. Besides, you’ve been overloaded lately…”
‘This isn’t right.’
In that instant, another thought flashed through Stark’s mind.
What if this feeling of “being used” had been a carefully crafted trick by Ha Si-heon all along…
In the end, what if all of this was a sche for Ha Si-heon to acquire those companies himself?
“Of course, I still think it’s better if you’re the one to do it. After all, it has more impact when soone who symbolizes innovation takes action. I’m not up to the task.”
He was pretending to be humble…
But in reality, in terms of dia buzz, Ha Si-heon was even hotter right now.
After all, he was the man featured in Ti magazine as the one who “brought down the Great Wall of China.”
“Isn’t this a ti when a one-year delay turns into a ten-year gap? Whether it’s or you, the person who can act now should move to stop Gooble. Of course, I don’t mind either way.”
Then, Ha Si-heon’s gaze locked onto Stark.
“The choice is yours.”
Stark was left with only two options.
One.
Jump into the AI battlefield himself under his own na.
But that felt like stepping onto a stage soone else had set, following a path soone else had drawn.
It also ant adding yet another burden to his already packed schedule.
Or.
Leave it to Ha Si-heon.
In that case, the dia would explode once again.
Ha Si-heon would earn titles like “the orca shaking big tech after shaking China,” and most importantly… he would go down in history as “the revolutionary who challenged Gooble’s AI dominance.”
Neither option appealed to him.
But at that mont, what mattered was which one felt “more” unappealing.
Ha Si-heon smiled gently and asked,
“What will you do?”
Finally, Stark made his decision.
“I… will acquire them myself.”
***
Just as expected.
‘There’s no way this glory-hungry man would hand over the spotlight to soone else.’
How do you manipulate a control freak?
Just give them a choice that’s even harder to endure than losing control.
In Stark’s case, that was surrendering the chance to have his na rembered in history.
Sure enough, as soon as I offered to take the role myself, he jumped right in.
But even after making his choice, his face was clouded.
“I have enough acquisition funds… but honestly, I’m at a loss for what to do after the rger. The integration process alone will be tough, and the subsequent learning costs will be enormous…”
The AI war was ultimately a resource battle.
Data, computing power, number of iterations…
All of it required money.
The side that poured in more money had an overwhelming advantage.
And Gooble had almost nation-level financial strength.
To fight such a giant, massive investnt was essential.
“Even if I establish an AI company, will I really be able to gather that much capital…”
In December 2015, Stark’s brand value had already declined significantly.
Too many ventures, too little profit.
So dia outlets had already started labeling him an “overrated genius,” and if he launched yet another new business in that state…
Stark looked unsure of himself, but I answered in a firm tone.
“It’s fine. Didn’t I tell you? This rocket launch will be a success. When it does, you’ll be re-evaluated, and capital will flood in.”
“But… that alone might not be enough.”
That was true as well.
Early investors typically start with small amounts to test the waters.
However, I smiled.
“It will be fine.”
An unimaginable amount of money would soon pour into Stark’s hands.
The reason was simple.
Because I intended to make it so.
User Comments
0 comments from readers