A Wall Street Genius's Final Investment Playbook

Chapter 286 : The Invisible Hand (21)

Meanwhile, as the AI power struggle between the United States and China intensified and the AI boom set the market ablaze, the atmosphere inside Gooble was heavier and more complicated than ever.“The cloud business is booming. If you look only at the financial statements, our performance is flawless.”

Gooble’s AI infrastructure business, built on its cloud platform, continued its dazzling growth.

But Gooble couldn’t simply rejoice in this situation.

“The problem is that we’re gradually losing our presence in the market.”

At the very center of the AI craze was not Gooble’s RL, but Next AI and LLM.

Once proudly touted by Gooble, RL had now become all but forgotten, like a relic of the past.

“How ironic. The era of AI we so desperately longed for has finally arrived.”

“And yet, how did we end up being pushed aside like this…”

Just a few months ago, the situation had been completely different.

Back then, there had been a clear competitive landscape: “Stark vs. Gooble,” “LLM vs. RL.”

Gooble was undeniably one of the two main pillars of that AI debate.

But all of that was overturned in an instant.

The cause was the rumor that “China is trying to attract Next AI.”

When that rumor collided with the massive stage of the presidential election, the focus of debate shifted in an instant—from “technological capability” to the frame of “national security” and “technology protection.”

“The only reason people are backing LLM right now is simple. It’s the fear that ‘we must never let China have it.’”

This was a current driven not by the intrinsic value of the technology, but by an emotional impulse—“we can’t let someone else take it.”

“In a situation like this… no matter how much we improve competitiveness or technological strength, it won’t matter. The gap wasn’t created for rational reasons in the first place.”

After much deliberation, Gooble made its decision.

“We need to change direction too. Toward LLM.”

Now that LLM had firmly become the industry’s dominant trend, resisting that flow was unwise.

So Gooble resolved to shift its product strategy toward LLM.

Of course, this was anything but an easy choice.

“No matter what anyone says, the leading player in LLM is Next AI. That makes us latecomers.”

In markets, first movers always hold the overwhelming advantage.

Brand recognition, market share, development speed… they lead in every respect.

For a latecomer to break through in such an unfavorable game, there was only one way.

“We need differentiation. We must leverage the one technology that ensures we have an edge.”

Fortunately, Gooble still had a hidden ace up its sleeve.

“Let’s use RL. We’ll put forward a hybrid model that combines RL with existing LLM.”

What they unveiled was “integrated technology.”

It was an ambition to compete with a level of technological advancement that could surpass Next AI’s model.

By doing so, they could follow the LLM trend while still leveraging Gooble’s unique resources for a leap forward.

And then, at that moment—

“Ha Si-heon has requested a meeting.”

“…?”

The conference room instantly froze.

“…What’s the purpose?”

“He says… he’d like to propose ‘reconciliation and cooperation.’”

“Reconciliation and cooperation…?”

The same thought flashed through everyone’s mind.

“Now, of all times?”

But they couldn’t flatly reject it either.

“If we turn down this meeting… that would be troublesome. Refusing a peace offering would mean he immediately considers us his enemy!”

Working with Ha Si-heon was deeply unsettling.

But ignoring him and turning him into an outright adversary was far too dangerous.

“For now… let’s hear him out.”

Once the meeting with Ha Si-heon was confirmed, Gooble’s leadership convened an emergency session.

The agenda had only one item.

“What exactly does he mean by ‘cooperation’? At this point, he shouldn’t have any need for us…”

In the AI war between Gooble and Next AI, Ha Si-heon was clearly the victor.

Backed by full-fledged national support, why would he bother to reach out to Gooble first?

Countless speculations circulated, but no clear answer emerged.

It was only a few days later, when Ha Si-heon himself appeared and spoke about “the proposal,” that the mystery was finally resolved.

“I want to pursue a joint project with Gooble.”

“…A joint project?”

“Yes. Next AI plans to develop an integrated model that combines LLM with RL.”

In that moment, Gooble’s leadership froze.

The strategy Ha Si-heon had just unveiled was identical to the plan Gooble had painstakingly decided upon.

“A hybrid model?”

“So they’re…?”

The initial reaction was shock and confusion.

But soon, regaining composure, Gooble’s executives ran the calculations—and quickly reached a not-so-bad conclusion.

“In that case… the situation shifts in our favor!”

With RL + LLM, both sides competing with the same hybrid model, the ultimate battle would come down to technological strength.

The deciding factor would be how well each side could master both LLM and RL.

But on the LLM side, thanks to Next AI having already open-sourced their code, there were virtually no barriers for Gooble to follow.

RL, however, was a different story.

Gooble had spent years and enormous resources building unmatched expertise in RL, while Next AI had almost no experience or infrastructure in that field.

In other words, if it came down to a showdown between hybrid models, the advantage was overwhelmingly in Gooble’s favor.

As Gooble’s executives quietly let out sighs of relief and triumph, Ha Si-heon suddenly dropped another bombshell.

“That’s why I propose a partnership. Next AI will provide expertise in LLM, and Gooble will provide expertise in RL.”

A short silence followed.

Gooble’s leadership just stared blankly at him.

Ha Si-heon was asking Gooble to hand over the RL mastery it had built over many years—yet he said it as though he were doing Gooble a great favor.

No matter how they thought about it, they couldn’t find a reason.

Finally, breaking the awkward silence, someone spoke.

“Next AI’s LLM is open-source, isn’t it? That means the technology is already available to anyone.”

“That’s correct.”

“…?”

So, essentially, Ha Si-heon was offering technology anyone could already take for free, while demanding Gooble’s exclusive expertise in return.

One executive, with a look of disbelief, asked, “What reason would we have to agree to such a partnership?”

“Oh, a reason? Of course, there is one. We need RL technology. And Gooble is the leader in that field. For us, there was never any reason to look elsewhere.”

“…??”

It made no sense at all.

“No, obviously from Next AI’s perspective that makes sense. But what I mean is, why would Gooble share its technology with a competitor?”

“Ah, that’s the reason you meant?”

Ha Si-heon gave a short laugh before replying.

“It’s simple. For the future of America.”

“What?”

“You know very well what the current situation is like, don’t you?”

“… …”

At that moment, the expressions of Gooble’s executives hardened.

Only then did Ha Si-heon’s true intention come into view.

“With China aiming for AI supremacy, is this really the time for American companies to weaken each other through internal competition? I believe now is the time to set aside narrow self-interest and join forces.”

Ha Si-heon’s argument wasn’t based on logic or numbers.

What he invoked was none other than “patriotism.”

“Of course, I’d understand if Gooble chooses to reject this proposal. But if that happens, we’ll have no choice but to seek RL expertise from foreign companies. And if that happens, this technology will no longer be America’s alone. That’s the concern…”

The gist was clear: “If Gooble doesn’t cooperate, we’ll partner with foreign firms.”

But the way he said it sounded uncomfortably like a threat.

“Of course, as a company you exist to pursue profit, and I respect that choice… but I do worry about what kind of repercussions such a decision might bring.”

The more they thought about it, the more it was clear—it was a threat.

America at that moment was boiling over under Tranton’s nationalist agenda.

If, in the midst of this, core AI technology were to flow overseas?

“Let me say once more, our only wish is for this technology to remain in American hands. And I sincerely hope Gooble feels the same.”

The message was unmistakable: if the partnership failed, all the blame would fall squarely on Gooble.

The frame “Because of profit-driven Gooble, American AI technology slipped overseas” was already taking shape in their minds, ready to spread across the entire nation.

And the man sitting before them was no stranger—he had already set public opinion ablaze multiple times on national TV.

“If we reject this…”

No matter how good a product they released, it would be meaningless.

Gooble would be branded overnight as a “traitorous company.”

Ha Si-heon, still wearing a gentle smile, delivered his final line.

“We fully intend to respect Gooble’s decision. The choice is yours.”

“…We’ll review it and get back to you.”

***

In the end, Gooble wrapped up the meeting with nothing more than a noncommittal answer.

Before rising from my seat, I added one last persuasive remark.

“This is a rare opportunity to become a national hero. I sincerely hope you’ll make the wise choice.”

The expressions on everyone’s faces as they left the conference room were far from bright, but I wasn’t particularly worried.

In truth, it wasn’t even something worth agonizing over.

Facing public outrage as traitors was far worse than accepting my proposal, which offered far more benefits.

I wasn’t entirely without conscience.

I knew full well how much time and resources had been poured into RL, and I had no intention of simply taking it for free.

For example, I’d make sure to frequently mention that Gooble was “working for the future of America’s AI.”

That way, the partnership would endure.

To apply AI to the development of a Castleman’s cure, I needed Gooble’s cooperation.

There were clear limitations that could never be overcome with LLM alone.

LLM could analyze massive amounts of medical data to identify the causes of disease or promising candidate substances for treatment.

But when it came to determining which among those candidates had the highest likelihood of success, LLM fell short.

That’s where RL came in.

To use a metaphor… if LLM drew the map, RL was what found the fastest and most efficient route across it.

In any case.

There was one last thing left to do.

As I climbed into the sedan, the driver asked, “Where to next, sir?”

“To Quantum Genome, please.”

Quantum Genome was one of the biotech companies I had invested in, a frontrunner in the field of spatial transcriptomics.

It was also the lab I had entrusted with decoding Milo’s biosamples.

I may have mentioned it before.

Milo’s biosample was like a vault.

Inside it lay hidden information about Castleman’s “madness switch.”

“So far, it’s remained locked…”

The truth was, I had waged an AI war for no other reason than to acquire the technology to open this vault.

There were three keys required to unlock it.

GPU, GNN, and Ignus.

The GPU had already been secured and delivered.

The GNN model, designed to interpret gene-to-gene interactions and spatial patterns, had been handed over last month.

And just last week…

The final key, Ignus, had been thoroughly tested—parallelized computation, optimized memory, compatibility with massive spatial transcriptomic data—before being passed along.

Its results would come soon.

But just as I was trying to suppress my impatience, a message arrived yesterday.

―It seems we’re ready to begin applying the technology to human samples.

It was the signal that everything was in place.

In other words…

The time had finally come to open that vault.