General feedback about my experience in creating a world in Meta Horizon

[MResearch]

Hey everyone,

My team and I worked on Crafter vs Dodgers and submitted it last week to the Meta Horizon Creator Competition.

• You can find our game here: https://horizon.meta.com/world/1176811227228729
• Our competition entry here: https://devpost.com/software/gravity-craft

This was our first time using and developing for Meta Horizon so I've got a few feedback to share in case that's useful.

PROS

• Networking was a cinch to use. Networking isn't our strong suit but everything worked out of the box as expected so I was surprised by that. It allowed us to create a social party game rather quickly. This is by far the strongest value IMO that Meta Horizon editor brings, which is great considering how important social is to Meta. Good work here.
• Gen AI was introduced to us midway through our development. We had purchased our assets before this and it was a big task to switch everything over, but we did it because Gen AI looked simply better. It was excellent at prototyping things and getting a general idea for space (i.e. where things should go). This is a groundbreaking tool and I expect most creators to be using this feature in the near future. It's that good!
• Player controller was basically plug-and-play for what we wanted to build. No more fiddling and developing custom code. The controller was simple which was perfect for the simple game we were developing. 
• Similarly, the Standalone Editor was just powerful enough for us to do what we wanted in our game. Save for one major feature (will mention that further below), we were equipped with everything we needed. It's a refreshing difference compared to Unity where we've had to develop a lot of things or buy third-party assets to do what we wanted.
• The performance was surprisingly good on mobile. I don't know if this is because we've spent 10 years developing VR/AR games and apps but we were surprised by how far we could push performance even with complex Gen AI models.
• Finally, separating the Play button and the World Simulator button was brilliant. I wish Unity had that feature. With it, we were able to skip all of the game logic by not having to hit Play and instead level design with physics on via the World Sim button. This allowed us to very quickly iterate on our physics-based game.

SUGGESTIONS

• Multi-select and change common properties in the selected objects like in Unity would help save a lot of time. I found myself having to click way too much on a daily basis due to the editor not having this. This would be my #1 request.
• Re-texturing a Gen AI 3D model would be really helpful since some of the textures don't turn out too nice. Even if we have the perfect 3D model generated, if we don't get the texture results we wanted, we were forced to generate a new 3D model from scratch. This wastes our time and it costs more for Meta so I think this feature is important to implement.
• Naming the downloaded Gen AI 3D models with the same name as what's in the editor would be nice. Right now, it's just a long string of characters. Also, including the file format in the name is great too, especially for new creators. If I hadn't attend one of the Gen AI workshops, I wouldn't know that I needed to add the file extension at the end of the file name.
• It was a bit more difficult than I'd like to control the styling of Gen AI. I've been using AI for several years now and so I'm familiar with prompt engineering. What could really help creators use Gen AI more effectively is negative prompts. This will allow creators to filter out unwanted themes, subjects, etc, instead of generating the same thing over and over again.
• A lot of documentation is provided but some of the pages could be organized better. For example, the community tutorials helped us immensely with our work but the content there could be broken apart and spread across their relevant pages instead of being collated under a single page.
• OBJECT POOLING. This is my #2 request. Given how the engine works, there's a strong need for a customizable built-in object pooling system. We had to create our own which was time consuming to do. I'd imagine most creators that are building worlds where new objects need to be spawned have/will develop their own pooling system. Might as well have this built-in to save everyone time. Thinking about this some more, Meta Horizon engineers could take a look at the most popular Unity assets and see if they're worth building out in their editor since many games need similar systems.
• Finally, this was another time waster for me: Quickly dragging in two different objects into the world will create two of the SAME object. This is more of a bug but slowed me down considerably given I had to test and retest many Gen AI 3D models in the scene by dragging them into the world.

Hope this helps! Honestly, we were all very surprised by how mature and easy to use the editor was. I'm glad the Creator Competition existed because it was an incentive that opened our eyes to what Meta have been building.