Imagine asking for a space-pirate RPG with sword combat, planet exploration and a crew that complains when you park the ship badly. It is a great idea. In your head, at least. The interesting part begins when someone puts the controller in your hands and you discover that the sword feels like a broom, the planets put you to sleep and the parking joke has already come up eight times.
That is why Ken Kutaragi’s vision for artificial intelligence deserves a conversation that goes beyond the generate button. If imagining an experience becomes much closer to playing it, what happens to the work of finding out whether it is any good?
What Kutaragi actually said
Known as the father of PlayStation, Kutaragi spoke to 4Gamer in an interview published on October 9, following a September 29 event at the University of Electro-Communications in Japan. He criticized the pace of progress in games compared with AI and imagined a form of entertainment in which fantasies and communication would become content.
One detail calls for care: Kutaragi cited a forecast for 2028–2029, when real-time AI video and music generation could reach speeds comparable to those of current consoles. The interview does not present this as a product announcement or a promise of complete games made to order. Nor does that passage identify the source of the estimate. Automaton covered the conversation on October 15.
Taking that thought as a starting point, our question is an editorial one: assuming these tools advance, how do we turn easier production into fun? The examples below are hypothetical scenarios for discussing design, not demonstrations of an announced technology.
The boss has appeared. Now make the fight work
Back to the pirates. A system capable of generating a six-armed alien captain would solve one visible part of the problem. Someone would still have to decide which attacks the captain uses, how it signals an incoming attack and how much time you have to react. If all the arms move at once, the spectacle might impress while the fight turns into a mess.
The initial request, “I want a difficult boss,” hides decisions that can lead to very different experiences. Difficult because it demands quick reflexes? Because positioning matters? Because it has an absurdly long health bar? The third option is a remarkably efficient way to turn a space battle into a day at the office.
That gap between the description and the experience is at the heart of the issue. A game needs to explain itself as it unfolds. The player should be able to see what they did, understand why they lost and recognize a chance to improve. We can debate each choice, but someone will have to make those choices and check that they work together.
Can a prompt capture the pleasure of pressing a button?
Then there is the familiar idea of game feel: the sensation of controlling something. Think of the impulse of a jump, a car slowing down or the brief interval between pulling the trigger and feeling the impact. These are details a player might never describe precisely, even though they notice immediately when something feels off.
“Make the combat feel heavier” sounds like a clear instruction until implementation begins. Should the animation slow down? Should the camera shake? Does the enemy need to react more? One change might strengthen the sense of force while making the controls less responsive. Another might work in an arena and get in the way in a narrow corridor.
A future tool could help with that tuning. The point is to require evaluation to keep pace with generation. Producing twenty versions of an attack would be useful if someone could compare them, keep the best and understand why it works. Without that process, we would have twenty different ways to get the same attack wrong.
The prototype could be the most exciting part
There is a real opportunity in this discussion. Imagine a creator with a good idea for a detective game in which every lie changes the architecture of the house. Before dreaming of fifty rooms, they need to find out whether one lie, one door and one consequence already make for an interesting bit of play.
If a tool reduces the effort needed to build that first experience, there will be more room to experiment. An odd idea could be tested before requiring a huge commitment. And when the result is bad, abandoning the solution and trying another would hurt less.
That use seems more promising than measuring success by the amount of content generated. A prototype is valuable because of the question it can answer. Did the door changing position make the player suspicious of the character? Or did they think they had found a bug? That difference could determine the entire future of the project.
More content also means more work deciding what to keep
It is tempting to imagine that an infinite library of possibilities would solve a shortage of creativity. But abundance also calls for editing. If creating a hundred missions becomes easy, someone will need to decide which ones deserve to be included, how they connect and why the player should want the next one.
In our hypothetical RPG, a short mission about the crew might be worth more than three extra planets. Perhaps the best decision would be to cut the hunger system, shrink the map or have that funny character appear less often. Creative direction includes standing behind those choices, even when the button to add more is right there.
It is also worth asking who will pay for the experience. Generating content during play would raise questions about processing, connectivity, availability and maintenance. The budget would depend on the solution adopted; there is no basis here for firm claims about savings or cost. The practical question remains: would the game stay accessible and playable under the conditions in which its audience actually plays?
Who looks after the game once the first surprise wears off?
A demonstration can win people over in a few minutes. A complete game needs to withstand a player trying to open the same door fifty times, returning to an area after ten hours or discovering a combination nobody had considered. The initial surprise covers only part of that relationship.
This becomes especially interesting when the content changes for each person. If a player encounters a mission that cannot be completed, someone will need to understand what happened in that particular case. If the solution is to generate another mission, will it preserve the earlier choices? Will the character keep remembering what matters? These are design questions, not details to leave until later.
The same goes for sharing the experience. Part of the fun of a challenge is telling someone how you got through it and hearing a different solution. In a personalized world, designers would need to think about what remains common to players. Perhaps other ways of sharing stories will emerge. That possibility is worth exploring without treating the social side of play as a problem that has already been solved.
Imagination deserves a good controller
Our bet is that the best version of this future will give more power to people who know how to observe, test and choose. Shortening the distance between idea and prototype would be excellent. So would being able to fix an attempt without starting everything over. The better the tool, the more interesting it becomes to ask for concrete results from those possibilities.
The standard remains within any player’s reach: did it make you want to try again? Did losing teach you something? Was the discovery worth the journey? If the answer comes with a smile and one more round, great. The space pirates can stay. But someone still needs to stop that sword from feeling like a broom.
