"President, we've been studying the issue of NPC freedom in the ga for quite a while now."
"And?"
"Our conclusion is this: with current technology, trying to cram this much content into the ga will make it unbearably bloated. Even setting aside whether consoles can even run it, the sheer scale could cause the ga to break under its own weight. We already have so many bugs identified internally that we can't count them all, and the more content we add, the more those bugs multiply exponentially."
"I see..."
In the Fourth Developnt Division, Takayuki stared intently at the logic patterns being developed for civilian NPCs.
In Red Dead Redemption 2, when the protagonist enters a certain area, all NPCs in that area begin to activate, following their own logic routines.
Soone might be staggering around puking after a night of drinking, while a warehouse worker starts hauling sacks of goods to a waiting wagon as he prepares for an early shift.
It all looked incredibly lifelike—like a real, breathing world.
But even that had its limits.
...
In Red Dead Redemption 2, NPC behavioral logic typically lasted three to five minutes at most. After that, NPCs would begin looping, repeating their earlier behaviors like clockwork.
A walking pedestrian would just circle the sa route endlessly, and no dynamic or random events would occur.
This, in a way, was the ga's way of reminding players: this is still just a video ga. There are boundaries. It can't be truly real.
But truthfully, Takayuki really wanted to create a living, breathing cyberpunk world.
It just... seed nearly impossible.
Every ga he had developed before had sothing to emulate or build on.
After all, the original world had over 40–50 years of comrcial ga developnt history, with decades of technical iteration and refinent, and a thoroughly mature frawork.
Even with that, the pinnacle of imrsive worlds was still capped at titles like Red Dead Redemption 2 or The Legend of Zelda: Breath of the Wild.
Those two had already pushed the industry to its limits—so much so that even four or five years after release, very few gas ca close in terms of world interactivity or technical execution.
In the original world, ga innovation had actually plateaued. For nearly half a decade, there was very little true progress—no explosive new chanics, no giant tech breakthroughs. The golden age of creativity had started to fade.
Most developers were now just polishing existing systems and making small-scale innovations.
As for major breakthroughs in tech? Those had largely stalled.
Now, Takayuki had hit that very sa wall.
In this world, video ga developnt had surged ahead for nearly 20 years—and it had progressed even faster than in the original world. Its current level of technical maturity was roughly equivalent to the old world's peak.
To push beyond that now, Takayuki would have to forge a completely new path.
What he wanted now was for every single NPC in Cyberpunk 2077 to have a fully defined logic system.
He wanted these background characters to act like real human beings—not mindless shells. He didn't want soulless, reactive puppets. He wanted believable routines.
Of course, he knew it was a bit of an exaggeration.
He wasn't expecting fully sentient AI. That was still science fiction.
But he did believe his idea could be achieved—at least in part.
To put it simply, he wanted NPC scripting to beco far more intricate and layered.
But the downside was massive.
The ga would beco bloated—losing the elegant "small and beautiful" feel of tightly designed titles.
Worse still, bloated gas are like skyscrapers stacked too high: at a certain point, the foundation gives way and the entire structure collapses under its own weight. There are no materials strong enough to sustain infinite height.
The sa is true for ga code.
The more complex a ga becos, the higher the risk of bugs and instability.
Eventually, the codebase becos so heavy that the ga itself risks being unplayable.
Even the advanced Unreal Engine of this world couldn't magically solve such issues. These were not engine problems—they were limits of software engineering itself.
To go further, Takayuki needed a new path forward.
His eyes were fixed on the screen.
There, an NPC was going through their simulated "daily life" in a partially constructed Night City.
That day included: taking a cab to a restaurant, eating a al, watching TV for a bit, then heading to a ripperdoc's clinic to repair a damaged cybernetic implant. Once the new implant was installed, the NPC wandered the streets.
By chance, he stumbled upon a gang shootout with city security. He panicked and fled. Upon reaching a safer zone, he stood still, breathing heavily, occasionally glancing back toward the direction of the fight.
Finally convinced the danger had passed, he flagged down a cab, told the driver his address, and even asked him to avoid the area where the gunfight had taken place.
He got ho, changed clothes, and collapsed into bed.
That... was one full day in the life of a background NPC.
It looked rich and compelling—like soone who actually lived in Night City.
This single routine, including scripting and animation data, took up over 300MB.
And this was for one NPC.
Takayuki wanted hundreds, even thousands of different NPCs, each with their own unique routines and behaviors.
But honestly? It was starting to seem far too complex.
Would players really care about such details?
Would anyone spend an hour following a lifeless background character just to watch them go about their day?
And worse, this 300MB only covered a single in-ga day—which equated to about one hour of real-ti gaplay.
After that hour passed, the NPC would reset and repeat the exact sa routine.
Programmatically, this was the upper limit. Even the one-day simulation was the result of careful balancing—any longer would be overkill, even in Takayuki's eyes.
Still... it left him unsatisfied.
Maybe it wasn't logic.
Maybe it was just a lingering regret.
The original Cyberpunk 2077 had ant sothing to him—and this was his way of redeeming it.
Reviews
All reviews (0)