The Beginning#

AI growing pains. It really has become one of those topics people have talked about endlessly over the past few years

The reason I decided to talk about it only now is that a friend and I brought it up over lunch today, which reminded me of the whole thing again. And since the blog still didn’t have an article about it, I thought I might as well write one…

In the age of AI, I want to build up more of my own stock of thoughts. Maybe that will help me make better use of AI in the future, too

The Conversation#

My friend brought up a problem over lunch: he didn’t know how to make his long-term project fun

That has been bothering me lately as well. I keep tinkering with older projects without finding what makes them fun, while new projects and new ideas seem more interesting instead…

Then we started talking about the recent Spotlight Game Jam

He took part the old-fashioned way, and he was completely stunned by the games being generated with AI

How could someone make a game with several hours of content in just 21 days?

You can see traces of generative AI in the art and other parts, and as for the code, there is even less to say

Which is understandable. Even if there was only one person on the team, maybe they had an army of AIs costing hundreds of dollars a month working under the hood, so of course they could make something that impressive or something…

Then we talked about the internet as it is now

It feels as if more and more things can be built with AI. A lot of internet services are inseparable from code, after all, and code happens to be something AI is especially good at handling

A typical example is Opus 5.5. Recently, a pile of animations made straight out of Opus 5.5 have already reached a remarkably high standard

I really have no idea what the world will look like in a few years, when we graduate…

Making It Work Is Not the Same as Making It Fun#

So, back to the question that has been bothering me all along

How do you make a game that is actually fun?

I have been making games for more than a year now. During that time, I have learned quite a bit, made a few small games, and kept trying to develop my own long-term project

As AI has developed, the way I work has changed a lot too

In the past, when I ran into a feature I didn’t know how to implement, I had to search documentation, read tutorials, or study other people’s code

Now, most of the time, I just need to describe my needs clearly and let AI teach me while helping me implement them

AI really has made many things that used to be quite troublesome much easier

But as it became easier and easier to implement the features I imagined, I found that the game did not become more fun because of it.

Take the long-term project I am working on now

The character can move, dash, and attack, and also has a doppelganger and several time-related abilities

Some of these systems took quite a lot of time to design and implement

But when I actually put them together and played the game… something still feels wrong. The combat has a hard time creating that urge to keep fighting

There are so many features, yet it still feels boring

What is worse, sometimes I don’t even know where to start making changes

If I can’t even explain why my own game is not fun, how much can having AI implement more things really solve?

In the past, I might have had plenty of ideas in my head but lacked the technical ability to turn them into reality

Now my ability to implement things has improved, and AI is there to help me, but I have started struggling with the opposite question: what should I make in the first place?

If AI could operate a game autonomously in the future, understand player feedback, continuously adjust the gameplay, and even design mechanics more interesting than anything humans could make, wouldn’t that be nice…?

It would certainly help me make more interesting games

But…

If AI can solve even that problem effortlessly, how much will there be left for me, as a game developer, to explore for myself?

AI Growing Pains#

I think this may be part of the growing pains of AI

I have always liked AI, actually

From chatting, searching, and learning in the beginning to taking part in programming, creation, and development today, it has helped me in countless ways

You could even say that without AI, the development of many of my current works would probably be much slower

But because I use AI so often, I can also feel its progress more directly than most people

Things that once took a long time to learn can now sometimes be completed with a few sentences describing what I want

Works that once required a whole team can now be made independently by one person using a collection of tools

The barrier to creation has fallen, and ordinary people can try ideas that used to be out of reach

But for people who have already invested a great deal of time learning and creating, this change is not always easy to accept

You spend months learning a skill, only for a new model to appear not long afterward and do most of that work directly

Or you spend a long time finishing a game, only to see someone else use AI to make a larger work in a much shorter time

Even if the quality is not necessarily the same, it is hard not to compare the two

Especially when this progress does not happen once every few years, but changes visibly every few months, or even faster…

The feeling is genuinely strange

It is as if I keep running forward, while the track beneath my feet is running even faster than I am

Of course, there is another question here

If I use AI myself, why do I feel anxious when other people use AI?

Maybe my subscription plan isn’t expensive enough. I have never tried the strongest, most advanced models

I do try whatever I can use. For example, now that I subscribe to GPT, I often use it to support my studies and development. As for the expensive models I cannot afford for now, such as Claude’s higher-tier plans, I can only watch other people play with them (sad)

Maybe it is because I want AI to help me become stronger, while also worrying that it will make the abilities I have built up in the past less important

And as using AI becomes more common, the things everyone can make become increasingly similar

A complex feature or a beautiful animation might once have been enough to impress people

Now it is easy to see large numbers of works made with similar techniques

So a strange kind of competition appears

Whoever gets access to the newest technology first has a better chance of making something novel before everyone else catches up.

Things made before everyone else has time to catch up may attract a great deal of attention

For example, animations made with Opus 5.5 are already being made by top creators on Bilibili with nearly 800,000 followers

The frustrating part is that this really is hard to argue with

Before Opus 5.5 animations became popular, this kind of animation was something new that humans had never seen before

Naturally, people wanted to watch it

But as more people start using it, the same thing quickly becomes ordinary

Then we have to chase newer technology again to create a more distinctive experience

And the cycle continues…

Sometimes I even feel that, in the age of AI, chasing technological progress itself might be the most correct choice

But I still have games to make, articles to write, plenty of knowledge to learn, and other things that deserve my full attention

If I spend most of my time learning about the newest models, comparing tools, and researching how to improve my efficiency, what happens to everything else?

Even after I realize that, it is still difficult to stop paying attention

Because AI really is progressing

And I really don’t want to fall too far behind

What About the Future?#

After all that, I still have no idea what the world will look like in a few years

Maybe by the time we graduate, the technologies that seem astonishing today will have become completely ordinary

The game-design problems that feel so difficult now may genuinely be solvable by AI in the future

So how much value will the technologies I am learning now have later?

I don’t know

Maybe the answer is still the same old one: develop your ability to think, and learn how to judge and solve problems

As for the technologies themselves, I think they still have some value

At least for now, I still need to learn them in order to understand and create my own work better

But who knows what will happen later…

Even with access to similar tools, different people can still make completely different choices

And sometimes creation is not only about finishing a work

I like turning the ideas in my head into reality. I also like trying new technology and exploring things I have never encountered before

Even if AI can do most of that work in the future, I don’t think I will necessarily stop creating

That said, if AI could one day make the game I want faster and better than I can, I would probably still feel a little disappointed

Our generation happens to be living through a rather unusual period

Maybe this is the growing pain of AI’s transformation

The Best Investment Right Now#

That brings me back to the conclusion my friend and I reached at lunch

If AI is developing so quickly and the future is full of unpredictable changes, what should we invest in right now?

Learning a technology may soon be followed by a stronger AI

Chasing the newest tools may not keep us ahead forever either

So sometimes it feels as if the best investment in the AI age is to take care of your body, stay healthy, become stronger, and live long enough to use even stronger AI

Although there is something absurd about that conclusion too