Since we started using the development assistance AI, the speed of game creation had noticeably changed.
I can only operate my computer for one hour a day, the same as before. That promise hasn’t changed even though we built new computers and created an AI.
What changed was the time after I left the chair.
“Check the collision 100,000 times. Change the angle and speed of contact little by little, and exclude the ones that bounce normally. Just separate out the ones that spin strangely or fly in unexpected directions.”
“Understood. For abnormal behavior, we’ll also record the conditions under which it occurred and the change in force just before.”
“That’s it, please. And for the terrain you’re creating, look for places you can’t move to once it’s completed. However, you can exclude steep slopes that were made so you can’t climb them from the beginning.”
“Conditions set. Both processes are expected to be completed by 3:14 a.m.”
“When it’s finished, just wait. I’ll see the results tomorrow morning.”
“Understood.”
And so I stepped away from the computer.
Previously, this was where my real work began.
After running a process, I’d wait until it finished, check the results one by one, and if there were any areas of concern, change the conditions and try again. There were days when I’d spend the whole hour just waiting instead of making something.
It’s different now.
Once I decide what to investigate, the AI takes care of the rest.
And the next morning, when I checked the results before getting ready for kindergarten, the AI had already compiled them into the form I needed.
“Completed 100,000 collision detections. There were 207 cases of behavior outside the specified normal range. Additionally, out of the 4,812 terrains generated, we detected 31 unmovable areas that could not be explained by normal gradients alone.”
“Can you tell me the order in which they should be corrected?”
“Priority has been calculated based on the impact on the surrounding area and the likelihood that players will encounter it. Displaying now.”
A list appeared on the screen.
It was divided not only into areas that actually needed to be fixed, but also into areas where the numbers were a little strange but wouldn’t be a problem as long as you were just playing.
Previously, just making that call would have taken a lot of time.
“…It’s convenient.”
Even though I’d made it myself, I couldn’t help but groan.
I decide, and the AI tries it. I receive the results the next day, look at them, and make my decision again.
The one-hour-a-day restriction itself hasn’t been loosened.
However, the “waiting” disappeared from that hour, and the things I was able to do increased many times over.
However, as I continued to use it, I began to see what the AI was good at and what it wasn’t.
◇
“Of these two, which one do you think we should keep?”
I lined up the two types of physical behavior I was prototyping and asked the AI.
Plan A had much lighter processing, and even when objects collided, they bounced back cleanly. Plan B increased the load by calculating the deformation at the moment of contact in a little more detail, but depending on how the collision occurred, the vehicle body might sink or lift on one side.
“Comparing only in terms of performance, we recommend Plan A. The average processing time is 11.7% shorter than Plan B, and the average frame rate under the same conditions is 6.2% higher.”
“What’s the load of Plan B?”
“It’s about 1.38 times that of Plan A. Considering only the current intended use, there’s no performance benefit to the additional processing.”
“Then I’ll go with Plan B.”
There was only a momentary pause before a reply came.
“To confirm: do you want to adopt Plan B, knowing that Plan A is superior in terms of performance?”
“Yeah. I’m choosing it knowing the numbers.”
“Then, to assist in making similar decisions in the future, I’ll record the reason for adopting Plan B.”
“Because it’s more interesting this way.”
“The standard for the rating ‘interesting’ is insufficient for reuse.”
“That again?”
Recently, this kind of exchange had been increasing.
I played the two videos again.
In Plan A, when an object collided, it bounced in the expected direction. The calculations were light, the behavior was stable, and there were no problems for the game.
In Plan B, the moment the corner of the vehicle’s body hit the ground, its posture might collapse slightly, and the reaction could cause the other side to lift up. Depending on the conditions, it could bounce in directions you’d never imagine.
If you only looked at efficiency, there were more reasons to choose Plan A.
Still, it was Plan B that made me want to watch it again.
“The moment it hits, it does something a little different from what you expected. When you see that, you think, ‘Whoa.'”
“Please explain the conditions that cause the ‘whoa’ reaction.”
“When I get a result that’s slightly different from what I expected.”
“Does this mean that the larger the difference from the prediction, the higher the evaluation?”
“No, if it’s too big, you won’t understand what happened. If you can’t see the connection between what you did and the result, I think it’ll just feel unreasonable.”
“Then, can you numerically specify the desired range of prediction error?”
“If I could turn it into numbers that neatly, I wouldn’t be having a hard time either.”
“With the current information, it’s difficult to re-evaluate based on the same criteria.”
“I guess so.”
It’s not that the AI is bad.
It can accurately compare processing time, and it can also determine success rate, average speed, and probability of failure. If enough player behavior data is collected, it’ll even be possible to compare which version people played longer, and where the majority of players gave up.
However, that’s a matter of measuring results that have already occurred.
When I line up two plans that no one has played yet and feel that
――This one seems more interesting.
It’s a little different from that feeling.
“Then I’ll ask you again. Which is more interesting, A or B?”
“It’s not possible to judge based on the currently set evaluation criteria alone. However, if you add the factors you want to use for evaluation, we can create a comparison model.”
“I guess that’s fine for now.”
If I could reduce fun to a complete mathematical formula, I’d hand over that formula and be done with it.
However, even with the same mistake, there are times when I can laugh and times when I get angry.
There are games that are more fun because they’re difficult, and there are also games that are more fun because you can play them without thinking.
There are times when you feel good because things worked as expected, and times when you laugh because you got a result you didn’t expect.
It seemed unnatural to lump them all into one number.
“Well, that’s fine.”
“What do you mean?”
“It’s okay that the AI can’t decide everything.”
“Do you wish to change my role?”
“No, it’s fine as it is now. I’ll leave the calculations and verification to you. I’ll think about what to keep in the end.”
“Understood.”
I marked Plan B as accepted.
We didn’t create the AI because we wanted to leave everything to machines.
I made it to show me things I couldn’t see on my own.
In that case, this division of labor suited it better.
◇
“Yuu, look!”
That day at kindergarten, Haru came over with his arms full of plastic building blocks of different colors.
“Did you make something else?”
“A car! It’s done now!”
“Show me.”
I looked at the thing he placed on the floor.
There were six round parts attached to the left and right sides of a square box. Two large boards were spread out on top, and a long, thin piece like a spear protruded from the front.
Up to that point, it was understandable as a car made by a child.
The problem was at the back.
Two dolls were stuck on with their heads down.
“…Is this really a car?”
“It’s a car!”
He answered without hesitation.
“Then why are there six tires?”
“The more there are, the faster it goes!”
“And what are these wings for?”
“For flying!”
“So it’s a flying car.”
“Yeah!”
Finally, I pointed at the dolls.
“So, what are these two doing?”
“Engines.”
“They’re people, though.”
“But they’re engines, right?”
If anything, he looked at me like he didn’t understand what I didn’t understand.
There didn’t seem to be any point in trying to get some logic out of this.
“Then try running it.”
“Let’s go!”
Haru placed the car on the floor and pushed it vigorously.
The six tires spun, and of course the wings on top were of no use.
The dolls stuck upside down didn’t generate any power either.
Still, the car kept moving forward, and as soon as its wings hit the leg of a chair, it spun to the side.
“Ah!”
The dolls all flew off, and one of the tires came off too.
Even so, it rolled a little on its remaining five wheels, and finally came to a stop on its side.
Haru and I looked down at the half-broken car.
Haru was the first to burst out laughing.
“That’s weird! It turned sideways!”
Hearing the laughter, the children nearby gathered around too.
“Let’s do it again!”
“Let’s add more wings. Since it just crashed, bigger ones are stronger!”
“Let’s put tires here too!”
“Over there, on the side?”
“Put them on the side too!”
What was supposed to be a car got more and more parts, and it became increasingly unrecognizable.
No one minded, though. If anything, the more the shape changed, the more fun they seemed to have.
“Yuu, you add one too.”
“Me too?”
“Yeah, this.”
What Haru handed me was a long, thin red part.
I didn’t know if it was some kind of fence or a decoration.
“Where should I put this?”
“Wherever you like.”
“Even in a place where it serves no purpose?”
Hearing this, Haru tilted his head slightly.
“Somewhere cool would be good.”
At those words, my hands stopped.
It’s not about whether it’s useful or not.
It’s not attached because it’s necessary.
It’s attached because it’s cool.
I remembered watching an anime about giant robots shortly after I was reincarnated in this world.
If you think of it as a weapon, it doesn’t need to be a giant humanoid.
There are many structures more stable than standing on two legs, and there’s almost no rational reason for it to swing a giant sword like the one a human wields.
Still, when I saw that machine holding a huge sword against the sunset, I thought, beyond all logic—
Cool.
“Yuu?”
“Ah, sorry.”
I looked at the red part in my hand again.
In terms of performance, it had almost no meaning no matter where I attached it.
On the contrary, it might make things harder to move depending on the location.
Still, I tried attaching it diagonally to the side of the car body.
“Maybe here.”
“Oh! That’s good!”
Haru happily leaned forward.
“Let’s put one on the other side too! It’s even cooler with two!”
I started to think about it structurally too, whether it should be symmetrical, but decided against it.
That’s not what matters right now.
“Good. Let’s add another one.”
Once it was attached to the other side, Haru and his friends immediately started looking for the next part.
◇
Even after I got home, I couldn’t get those building blocks out of my head.
The children weren’t playing with a finished car.
The time spent building the car was itself the play.
Attach something and try moving it. If something doesn’t go as planned, instead of dismissing it as a failure, laugh at how it broke and then add another part.
They didn’t even have a rule that it’s not a car if it doesn’t have four wheels.
It could have six tires, it could have one tire, it could have wings.
No one stopped playing even when someone insisted a doll was an engine.
I opened my notebook and wrote at the top.
――Make the making itself the play.
Below that, I continued writing down my thoughts.
Children can have fun just by attaching parts in shapes they like and making them move.
If adults want to play deeply, they should be able to think about the center of gravity, friction, output, air resistance, and even the strength of the material.
“You can keep the entrance simple and make it deep.”
In 《Pettan Forest》, we reduced the number of controls so that Haru could play alone.
Move, jump, and stick.
This time too, you don’t need to remember much at first.
However, this time, the player will be able to choose the “things to attach” themselves.
“Grab parts and attach them wherever you like. Once you’re done, run it… I think that’s easy to understand.”
At the very least, you can play with just that.
On the other hand, those who want to go deeper can think as much as they like.
Even if they’re playing the same game, it doesn’t matter if Haru and his father see completely different things.
“I can make this.”
The necessary mechanisms were connecting in my head.
◇
The first thing I prepared was a square frame, tires, and a simple power unit.
First, I attached four tires in the natural positions at the front, back, left, and right and connected the power.
When I ran it, the vehicle on the screen moved across the grassland without any problems.
“It’s normal.”
“Stable running has been confirmed.”
“Okay, let’s use only one tire this time.”
I attached a single wheel at a location slightly offset from the center of the frame.
When I connected the same power and ran it, it managed to maintain its posture for a few seconds, but eventually its center of gravity collapsed and it fell sideways.
However, because the single wheel kept rotating and dragging the fallen frame, it continued to move along the ground in a strange arc.
I couldn’t help but laugh.
“Unable to maintain running posture. For stability, we recommend adding training wheels.”
“You don’t have to fix it this time. Let’s leave it as is.”
“Understood.”
If you look at it as a proper car, it’s a failure.
However, if it’s interesting to watch, there’s no need to delete it.
Next, I increased the power.
When I attached one rocket, it accelerated rapidly, and when I attached two rockets, it went even faster. When I increased the number to four, the front wheels began to lift slightly when accelerating.
“Then what happens if we add ten of them?”
“With the current car body structure, there is a very high possibility that joints will break during acceleration.”
“I know. That’s why I want to see it.”
I started it up as it was.
The moment the ten rockets burst into flames all at once, the car body jumped forward with tremendous force.
However, the joints couldn’t withstand the load, and the front wheels came off first, followed by the rear wheels being blown off. Only the remaining frame spun in the air, slammed into the ground, bounced, and finally crashed into the wall and came to a stop.
“This is terrible.”
“The recommended number of rockets for the current structure is two. If ten units are to be installed, the frame and joints will need to be reinforced.”
“Then let’s make reinforcement parts that can withstand ten units.”
“For normal driving purposes, a two-rocket configuration is lighter and more efficient.”
“If you want to run normally, two are fine. But there may be people who want to attach ten rockets.”
“At this time, we cannot see any rational advantage in choosing ten.”
That’s when I remembered Haru at kindergarten.
“Some people think it’s cool to have a lot of things on.”
“Coolness cannot be compared using current evaluation standards.”
“That’s fine. Even if you don’t understand all of it, just calculate the necessary strength and that’ll be enough.”
“Understood.”
I laughed a little.
At times like this, it was actually easier for me when there were things the AI didn’t understand.
◇
From there, we added parts one by one.
Simply attaching wings didn’t make it float into the sky. The force they received changed depending on speed, angle, and area, and without sufficient speed they had little meaning.
However, if a child who doesn’t understand such logic thinks, “Maybe if I add wings, it’ll fly,” and the car body lifts just a little bit, that alone is a discovery.
The propeller was designed to generate force in the direction it was pointed. If you point it forward, it becomes propulsion, and if you point it upward, it tries to lift the car body. If the number and output differ between the left and right sides, it naturally won’t go in a straight line.
If you attach a float, it becomes easier to stay afloat when you enter the water. However, if you attach a lot to only one side, the car body will lean, and it adds extra weight on land.
I also made legs.
Posture control was difficult with two legs, but relatively stable with four. Adding eight legs made the movement more complicated, and it looked more like a giant insect than a car.
“This is good.”
“Compared to the four-legged configuration, movement efficiency for the same power is reduced.”
“That’s not what I’m talking about. I just like the way it looks.”
“Recorded.”
You can also attach both tires and legs. You can also attach tires to the tips of the wings. You can also place a float next to the propeller.
I didn’t decide on any “wrong combinations” in advance.
Tires rotate and transmit power to the ground, wings catch the air, and floats get buoyancy from the water. After giving each part its role, I made it so a joint would break if a large force was applied to it.
The rest was up to the player to decide how to combine them.
“Prepare the causes and let physics give you the results.”
It was an idea I’d used in 《LITTLE FRONTIER》 and 《FRONTIER SAGA》.
This time, I’d use it to make vehicles.
◇
Naturally, the heaviest part of this game was the physics calculations.
Even when moving a single vehicle, it was necessary to consider the mass and center of gravity of each part, friction with the ground, air resistance, and buoyancy when entering water. Furthermore, if there were wings, it was necessary to calculate the lift, track where the force generated by the rockets and propellers was transmitted, and even cause breakage if the load concentrated on the joints.
Furthermore, broken parts don’t just disappear.
A tire that’s blown off can hit other parts and cause further damage from the impact.
With the laptop I used before, it wouldn’t have been practical unless I significantly reduced the accuracy somewhere.
It’s different now.
I have a computer for development, and an AI that runs verification even while I’m sleeping.
“Please specify the test conditions.”
“First, make a million different car bodies by changing the positions and numbers of parts. The tires don’t have to be symmetrical, and I want a mix of power sources, from one to ten.”
“Understood.”
“Don’t use just one type of place to run them. Use paved roads, sand, mud, rocks, ice, and roads made from wooden planks. Also test different speeds in each case.”
“To what extent do you want to document the damaged structures?”
“Don’t just record the broken part, but also the place where the force was concentrated right before it. Also, don’t throw away the ones that didn’t run properly. I want you to separate them by how they failed.”
“Conditions set. Estimated completion time is 4:37 a.m.”
“Thanks. Just wait when you’re done.”
“Starting processing.”
And so I stepped away from the computer.
Today’s hour was over.
◇
The next morning, when I checked the test results, the AI had sorted the million types of vehicles, including the reasons why they couldn’t run properly.
Some were unable to maintain their posture and fell, while others broke midway because their frames couldn’t withstand the acceleration. There were structures that didn’t have enough power to climb hills, structures that sent power to the tires but just slid on the ground, and even designs that ended up struggling in place because the wheels didn’t reach the ground properly in the first place.
Among them, there were 826 cases that the AI couldn’t explain well using the existing failure classifications alone.
“Show me only those 826. In descending order of distance traveled.”
“Displaying.”
The first thing that came up was a car body with three tires.
The problem was that all three were biased to the left. Even though there were no tires on the right side, the only power unit was attached at the right rear.
“Who would make this?”
“This is a structure created within the specified random generation conditions.”
“Ah, I specified that.”
I played the driving video.
As expected, it tilted to the left from the start.
It didn’t advance more than a few meters before rolling over.
Normally, the run would have ended there as a failure.
However, as the right rear power source continued to pull the fallen vehicle, the left tire that hit the ground bounced up.
The tires that touched the ground again rotated and lifted the car body once more.
It bounced as it fell, moved forward a little, and jumped again.
Repeating this at a regular cycle, it moved forward like a strange creature.
“…What is this?”
“Compared to a normal four-wheeled structure, the movement efficiency is 13.4%. However, continuous changes in position occur even though the running posture is not maintained.”
“Save it.”
“It’s an inefficient structure. Do you want to save it?”
“Yeah. That’s exactly why I want to keep it.”
“Understood.”
The next structure displayed was even stranger.
It had two wings, and the left and right wheels were different sizes. The propeller was also off-center, so just by looking at it, you couldn’t believe it would go straight.
When I started it up, the car body began to veer significantly, as expected.
The unbalanced thrust of the propeller added to that, and its posture became more and more unstable.
In that state, one wheel ran onto a rock.
The car body bounced.
At that moment, enough wind hit its wings and it floated diagonally upward.
“Oh.”
It only flew for a moment.
It crashed immediately.
However, because of the angle at which it fell, the car body bounced again, and the wings caught the wind again.
Little by little, it moved forward, and you couldn’t tell whether it was running on the ground, jumping, or flying.
“A short period of flight occurs repeatedly between the running and jumping states. Movement efficiency is low compared to stable flight structures.”
“But this is more interesting to watch than the regular airplane I saw earlier.”
“That evaluation cannot be derived from performance values.”
“I know. That’s why I’m watching.”
I saved that footage too.
The AI didn’t bring it to me because it thought it was interesting.
It tried a million different types and only picked it up because it didn’t fit into the usual classifications.
But that was enough for me to find a structure I’d never have thought of on my own.
The AI can test in large numbers.
It can organize the results and find differences.
And if there’s something that interests me, I can keep it.
“It really is a division of roles.”
“What do you mean?”
“You don’t have to decide what I find interesting.”
“Understood. We’ll maintain the current evaluation method.”
◇
Once the play I was creating took shape, I needed a place to run the creations.
You can have fun just assembling things freely.
However, for first-time players, it’s easier to get moving if they have a destination to head for.
So, the first course was an empty grassland.
The goal could be seen in the distance from the starting point.
All you have to do is build your favorite vehicle and get there.
On the next course, I placed a river. No bridge.
You can build a fast car and use its momentum to jump over the water, or you can attach lots of floats and cross the surface. You can use wings to cross over from above, or if you attach tires to a sufficiently heavy body, you can drive along the riverbed while submerged.
“Driving along the riverbed is definitely inefficient, but it looks interesting.”
“Speed is significantly reduced due to water resistance.”
“It’d be nice if it could still reach the goal.”
Next, I built a steep mountain path.
There are ways to climb using large tires, and there are ways to use a multi-legged structure to grip the ground. You can use wings to get over it, or you can point rockets downward and climb the slope more like bouncing than flying.
Even if you bump into the cliff many times, losing parts each time, and end up rolling to the top with just the frame, if you reach the goal, you’ve succeeded.
Narrow tires tend to sink in sand, and the difference in friction on snow and mud affects the design.
If it’s a canyon, you can choose whether to jump over it or go down to the bottom.
A long-distance course that requires a detour across water, land, and finally the sky also seemed interesting.
However, there’s no correct blueprint for any course.
“The players themselves are the designers, so it’d be a waste if I decided the answer for them.”
I wrote down my thoughts in my notebook.
Run what you’ve created, and if it doesn’t work as expected, think about the cause and rebuild. After repeating this process and reaching the goal, some people may want to see if they can go faster with the same structure next time.
“Let’s make it possible to measure times as well.”
You can compare on the same courses.
However, competing doesn’t have to be about speed alone.
You could play by arriving with as few parts as possible, or by completing the course without using much fuel. There are people who’d aim for a design that doesn’t break anything along the way, and people who’d be interested in reaching the goal without using tires, or conversely, trying to do everything with just rockets. I’m sure players will find their own ways to play.
“You can play the same course as many times as you want by changing the conditions.”
It would also be interesting to regularly issue conditions from the official website.
Everyone builds in the same place with the same constraints.
As a result, the fastest vehicles will emerge, and there will also be vehicles that run surprisingly well with few parts.
And I’m sure there’ll be some where I won’t understand what they were thinking when they made them, even if it’s explained to me.
“Would you set something as an independent evaluation item for when it’s unclear what they were thinking in creating it?”
“You don’t have to. People who see it can just laugh at that part.”
“Understood.”
◇
I didn’t agonize much over the title.
The main focus is on making and running.
I wrote it out in the alphabet in my notebook.
《CRAFT RUNNERS》
“Yeah. This is fine.”
There’s no main character in this game.
There’s no hero chasing after his father’s killer, and there’s no Demon King carrying a hundred years’ worth of history on his shoulders. There was no need for someone to walk through the forest in place of the player, like the round, white protagonist of 《Pettan Forest》.
There are parts and courses.
And there are just the players thinking about what to make.
Even so, as long as I don’t decide on the answer, things I never imagined will definitely come out.
That’s more interesting.
◇
A few days later.
Haru came to visit my house.
He tried to run into the living room before taking off his shoes at the entrance, but his mother stopped him, and then he came back over to me.
“Yuu, what are you doing?”
“I’m making a new game. It’s still in the middle.”
“I want to see it!”
Haru looked at the screen and immediately tilted his head.
“Is this a car?”
“It can be a car too.”
The vehicle left over from testing had three tires, two wings, and a propeller. I’d also added four legs to see how it behaved, so I didn’t even know what to call it.
“That’s weird!”
“I’m trying out different things.”
“Does it move?”
“It moves. Do you want to try it?”
“Yeah!”
I handed him the controls.
When Haru started it, the tires spun and the four legs kicked the ground.
At the same time, the propeller began to rotate.
Because multiple forces were acting in different directions, the vehicle began to run diagonally sideways instead of forward at a furious pace.
“Wow!”
It climbed onto a slope and jumped.
At that moment, its wings caught the wind and it floated slightly into the air.
“It flew!”
The next moment, it slammed into the ground, and one of the tires came off from the impact.
Two of the four legs broke.
The propeller was still spinning, so the nearly wrecked car body slid sideways down the slope and fell to the bottom.
When it finally stopped, I looked at the screen.
“It’s a complete failure.”
I heard Haru’s laughter beside me.
“What is this! Let’s do it again!”
“One more time?”
“I want it to be faster than it is now!”
He didn’t seem worried at all about it breaking.
If anything, he was enjoying the way it broke.
“Can you make these wings bigger?”
“I can.”
“I want to add more tires!”
“You can do that too.”
“Do you have rockets?”
“There are, though.”
“Can you add a lot?”
“You can add as many as you like.”
Haru’s face suddenly brightened.
“Then let’s put them all on!”
“Everything? Tires, wings, rockets?”
“Yeah! It’s stronger when it’s full!”
His logic hadn’t changed since he was making that block car at kindergarten.
Considering efficiency, there’s no point in designing like that.
If the number of parts increases, it gets heavier and the center of gravity collapses. If you increase the number of wings and tires without thinking, the resistance increases, and if you attach a large number of rockets, the joints won’t be able to withstand it.
There was a high probability that it wouldn’t run properly.
“Okay. Add as many as you like.”
“Yay!”
Haru grabbed the parts.
He added one tire and attached another one right above it.
It seemed the left and right wings alone weren’t enough, so he started attaching them diagonally to the top of the car body.
Rockets were lined up at the back.
“One, two, three, four… I’m adding more!”
“I think it’s fast enough already.”
“Make it faster!”
“Then it’ll break if you leave it as is, so it might be a good idea to reinforce it.”
“The hard ones!”
“Yeah. These.”
When I brought up the reinforcement parts, Haru began sticking them all around the car body as he pleased.
The shape became even more unclear.
In the middle of that, the AI’s analysis results appeared at the edge of the screen.
“The current structure is significantly less efficient when compared to a standard four-wheeled structure. Also, because the center of gravity is biased towards the rear, the front may lift when accelerating.”
“I know.”
“If the 12 rockets currently mounted are used simultaneously, the probability that a joint will break is 87.2%.”
“That’s pretty high.”
“Yuu, can I attach two more rockets?”
“Okay.”
Haru happily added the thirteenth and fourteenth.
Naturally, the displayed probability of breakage rose even further.
“We recommend changing the structure.”
“No need to change it this time.”
“Would you like to record the reason for maintaining the current design for future decisions?”
I looked beside me.
Haru was thinking seriously while slightly adjusting the angle of the fourteenth rocket.
“The reason… the person making it thinks it’s more interesting.”
“‘Interesting’ is not quantified by current evaluation standards.”
“Yeah. That’s fine.”
“We cannot confirm any superiority in terms of performance.”
“I know that too.”
Haru finally decided where to put it and looked at the screen with satisfaction.
“Done! It’s definitely cool!”
Is it a car, a plane, or something else?
I didn’t even know anymore.
However, Haru seemed happy.
“That’s right. Much cooler than before.”
“Coolness cannot be judged using current evaluation standards.”
“You don’t have to judge it that far.”
The calculations can be left to it.
It can also try out a million different ways.
It can tell in an instant that this design has a probability of failure of over 87.2%.
However, even after looking at the numbers,
――I still want to add all fourteen.
There’s no need to force the reason for thinking that way into numbers.
“Haru.”
“What?”
“Now that it’s done, let’s run it.”
“Yeah!”
I pressed the start button.
Fourteen rockets burst into flames all at once.
We don’t use ads!
If you enjoyed our translation, please consider buying us a cup…
Thanks!!!
Become a Patron!
We don’t use ads!
If you enjoyed our translation, please consider buying us a cup…
Thanks!!!
Become a Patron!