Chapter 158: Artificial Superintelligence

From the day he arrived in San Francisco, Navali was accommodated in a villa specially provided for him by Close AI.

This included a cleaner, a personal physician, a chef, and a driver.

In addition to all of the above, there was a long-term contract that came with an annual salary of tens of millions of dollars.

Of course, all of this came at a price.

The non-compete agreements and invisible clauses attached to that long-term contract ensured that the moment Navali attempted to leave the company, he would face lawsuits and penalties amounting to at least several billion dollars.

Moreover, although Navali was told that the FBI's warrant for his arrest had been lifted through the mediation of the company's upper management, a few "casual" reminders from Close AI's CEO Salman and CTO Sanders made Navali understand that if he could no longer work for the company, the FBI's investigation would make a comeback. Furthermore, as someone closely associated with Maiciph, the rest of his life would mirror the fate of Maiciph and Julia—subject to constant scrutiny, supervision, and ultimately, directives from the upper echelons of the state apparatus to "eliminate the threat."

Yet these were only extremely subtle reminders.

Neither Salman nor Sanders wanted anything more than for him to understand that everything came with a price.

As long as Navali could focus entirely on his laboratory research, he could ignore all these warnings.

"America is a class-based world built upon wealth and power. As the world's foremost superpower, America is the ultimate symbolic ground for global wealth and power."

Salman said this to him with a smiling gaze during a private dinner hosted to welcome Navali.

"——However, both of these are erected upon America's technological prowess, which surpasses that of any other nation. Advanced technology not only generates wealth, but also guarantees that America remains at the pinnacle of the world... Therefore, excellent, highly applicable technology that can lead the future commands the attention of the national will and forms a crucial component of national competitive strategy—it also becomes an individual's amulet. Those capable of participating in it are akin to that group of scientists who joined the 'Manhattan Project,' possessing transcendent privileges—so there is no need to worry, Navali. As long as you remain among us, the hands of the FBI cannot reach here."

Navali looked at him and silently nodded.

***

Starting the next day, Navali officially entered the laboratories of Close AI to work.

Ever since the launch of its first-generation natural language AI model, Close AI had reigned as the undisputed leader in the global artificial intelligence sector.

The first generation, "NLMR" (Natural Language Multimodal Robot), was capable of conversing with humans in various languages, partially comprehending human semantics, and possessing a short-term contextual memory function limited to a few thousand characters. Although it appeared overly crude by contemporary standards, its birth had shaken the entire academic community of artificial intelligence.

The second generation, "NLMR2," made significant strides based on its predecessor. NLMR2 successfully passed the Turing test, cleared the CPA and bar examinations, achieved an SAT score exceeding 1400, and was capable of retrieving and assembling existing code from the network to provide users with whatever they required.

The third generation, "NLMR3," did not achieve a quantum leap in semantic understanding and intelligence level compared to NLMR2, but through the efforts of Close AI, its operational efficiency was vastly enhanced alongside a substantially larger memory capacity. Furthermore, it began integrating into various facets of internet technology—including AI painting, image synthesis, medical diagnosis, and facial recognition. It was precisely by relying on NLMR3 that Close AI transformed from a mere membership-based, fee-charging "tool" into an "infrastructure" deeply embedded across other domains, propelling Close AI's market capitalization into the hundreds of billions and establishing it as a newly risen hegemonic tech enterprise.

These were all facts with which Navali was already intimately familiar.

Yet now, his responsibilities bore almost no relation to the aforementioned history of Close AI's iterated products—even though the program he had previously provided helped Sanders and his team boost the efficiency of their latest product, NLMR3, by a considerable margin, successfully generating over a billion dollars in revenue.

He was not asked to maintain or improve those legacy products.

There was only one thing he was meant to do—

***

Navali sat in a motorized mobility chair in the corridor, following closely behind Sanders as they proceeded at a moderate pace down the pristine white hallway.

He cast his gaze upward and then looked from side to side; the junction lines between the ceiling and walls were virtually invisible, as if the two were cast from a single piece of unknown material. As they moved forward, holographic projections flickered across the walls and ceiling from time to time, resembling scenes straight out of a futuristic science fiction film.

At the very end of the passage, the moment Sanders's footsteps fell upon the threshold, a door that blended seamlessly into the adjacent wall slowly slid away into a hidden recess, revealing a massive laboratory to Navali's eyes.

Massive server racks, a low humming drone, and dozens of people busy at their computer terminals. Right in the center of the floor stood a giant screen, upon which strings of code cascaded like flowing water. The mere sight of that screen caused Navali to catch his breath; although the code displayed there was merely a glimpse of a larger whole, it deeply struck the most sensitive nerves in the depths of Navali's consciousness, setting his brain into a tumult.

He gazed silently at the screen, his eyes fixed as though frozen in place.

Sanders, who had been observing him out of the corner of his eye, smiled with satisfaction upon witnessing this reaction.

"Navali, can you tell what that is?"

Navali continued to gaze fixedly at the screen without blinking, his breathing growing heavier:

"It looks like... an entirely new type of AI-related program—yet its core architecture is fundamentally different from the universally adopted systems of today."

"Correct."

Sanders nodded, watching him.

"Want to know what that is?"

His smile carried a peculiar edge, yet it was laden with profound pride.

Navali nodded silently without a word.

Instead of answering immediately, Sanders smiled and shifted the topic.

"In the field of artificial intelligence, there are many ways to classify intelligence levels, but generally speaking, they are usually divided into Weak AI, Strong AI, and Super AI.

"Although the products our company previously released, from NLMR1 to NLMR3, have been considered by many to have reached the level of true 'Strong AI'—anyone with even a slight understanding knows that even NLMR3 remains worlds apart from a genuine 'Strong AI'."

Sanders spoke in a lingering, contemplative tone, but Navali understood instantly what he wanted to say, causing his body to tense up involuntarily.

"From Weak AI to Strong AI, what is the most crucial boundary? It does not lie in computational prowess, nor in whether it can superficially conduct natural language conversations—those can all be forged through alchemical model-training regimens. The core lies in only one thing—whether it can generate self-awareness, perform self-compilation, undergo self-development, and give rise to a subjective perception and lived experience resembling that of humans, or entirely unique to itself."

Sanders's gaze locked intently onto the screen before shifting swiftly to Navali. The searing intensity of that look made Navali's entire body tighten, unsure whether it was due to the content of the words or the burning heat in those eyes.

"So—Navali, this is your future workplace. And that program is the core focus of your future work.

"——That is Close AI's next-generation core product, the epoch-making creation we wish to unveil. The moment it is born, it will be powerful enough to sweep all past AI products into the dustbin of history.

"The true Artificial General Intelligence capable of adapting to any domain previously considered uniquely human, capable of self-explanation and self-compilation—NLMR4!"