Development is now in the age of large-scale AI. Utilizing conversational AI technologies like ChatGPT and LLMs to reference code, or leveraging GitHub's Copilot functionality, which surpasses the intelligence features commonly used in the past...

It's a very convenient and interesting time.

In many aspects, this is a time of change and development that is expected to bring many improvements and advancements in terms of "productivity" enhancement.

So, I have also been making efforts to see if I can improve productivity for someone like me, a solo developer, by utilizing AI technologies that have been gradually becoming known.

Trying Game Development Using AI Technology with ChatGPT
ChatGPT is incredibly hot these days. I studied it and tried using it a few times, but... actually, I couldn't get a feel for where to use it. For an aging developer who can't keep up with the rapid pace of the times, new technology seems to bring both awe and a lot of worry. Then, I thought it might be a great partner for solving an issue I've been worrying about recently, so I decided to test it. It's about working together on "planning assistance" tasks
SOONSOON STUDIOsoonsoon

To sum it up

"It wasn't easy."

It is certainly highly productive in some aspects, but.. I currently judge that it is still far off. There are too many things to consider..

In the meantime, Unity, which I primarily use, has developed and released its own AI service directly within the editor.

I had heard that Unity was directly providing AI functionality, but I was surprised to see it launch so soon. Isn't that too fast?

The introduction video at the time of Unity Muse's announcement is as follows.

The features introduced on the homepage are as follows.

It seems that, just like conversing with an LLM, you can ask about code, write code, generate image resources within the Unity editor, and even create modeling such as terrain data.

It seems that not all features are open yet. You can experience it through interactive commands on the site below.

So, I decided to give it a try. There is also a video created on YouTube, so I recommend checking it out if you are curious. (Please subscribe and like if you appreciate it.)

Unity Muse | Ask anything
Accelerate creation with Muse Chat, Unity’s AI assistant based on large language models. Leverage Muse with Unity resources to speed up development.
Ask anything

First, I asked Muse directly what kind of functions it can perform.
It understands my clumsy questions perfectly. The response results do not appear immediately, but it seems to have a response speed comparable to ChatGPT 4.

I asked more specifically how to use Muse.

...

It seems to be based on an LLM model similar to ChatGPT, judging by the nonsense it's spouting.

Let me ask a more specific question to get a proper answer. However, there is no substance to it.

Assuming a certain level of familiarity with Unity, it answers questions about the software's detailed features quite well, acting somewhat like a manual.

I asked more specific questions about code used in Unity.
It seems to create these parts quite convincingly well.

I asked about some difficult requirements that require a bit of understanding of Unity.

And the answer is... quite basic, but it seems to be saying the right thing.
It doesn't seem to be the answer I wanted, but this might be due to my question itself not being specific.

I asked specifically how to do item 2 from the method Muse suggested. Since I am not a C# developer, I am not very familiar with how to use a Queue. (I think I've heard of it before.)

using System.Collections.Generic;using UnityEngine;public class ObjectPooler : MonoBehaviour{  public GameObject pooledObject;  public int poolSize;  private Queue objectPool;    void Start()  {      objectPool = new Queue();        for (int i = 0; i < poolSize; i++)      {          GameObject obj = Instantiate(pooledObject);          obj.SetActive(false);          objectPool.Enqueue(obj);      }  }    public GameObject GetPooledObject()  {      if (objectPool.Count > 0)      {          GameObject obj = objectPool.Dequeue();          obj.SetActive(true);          return obj;      }        return null;  }    public void ReturnPooledObject(GameObject obj)  {      obj.SetActive(false);      objectPool.Enqueue(obj);  }}

Oh.. it looks quite convincing.

Since I am not well-versed in C# syntax, I think I can actually study through this example code and learn how to use it.

Of course, it seems to be up to the user to try it out directly and verify that it works well.

That concludes a very brief introduction to Unity Muse.

For now, it seems to be a good tool as an interactive manual for users unfamiliar with Unity, and as a tool to reference code.

However, I think we need to check its effectiveness and accuracy by using it a bit more. As seen above, it can reference code, but... since there are already various solutions like GitHub Copilot, I'm not sure if we really need to use this one.

Although I still need to observe it a bit more, I think we need to maintain continuous interest and study as various development methodologies are rapidly changing through the utilization of AI technology.