voxelmetaverse VS AI-basketball-analysis

Minecraft may actually be closer to the metaverse than commonly believed:

Since the introduction of BungeeCord in 2012 (and then Waterfall and Paracord), it has become increasingly popular to link together multiple servers to act as a gateway to different virtual worlds on different servers. Most of the top popular "servers" are in actuality multiple servers joined together, somewhat decentralized, though still centrally managed.

Even before Bungee, the reign of Bukkit (2010-2014) introduced a plugin API system allowing for managing multiple worlds. To this day the "Multiverse" plugin remains among the top plugins. The multiverse, not the metaverse, but a related concept.

It wouldn't be too far of a leap to link together unrelated Minecraft servers.

Regarding "you can certainly not expand the protocol -southerntofu" - the Minecraft protocol is commonly expanded by modders. In fact, it is specifically designed to be expanded, since the introduction of Plugin Channels in Minecraft 1.1: https://wiki.vg/Plugin_channels. Forge modders frequently enhance the protocol to support new functionality far beyond what was possible in the original game.

"new media forms and new mediums of access (web, mobile, PCs, AR, etc.)" vs "you can only run Minecraft where Microsoft distributes it (unless you crack/RE it)" - granted, but there are multiple unofficial efforts to develop new ways to access Minecraft servers, including through the web. My humble attempt at building such a client: https://github.com/voxel/voxelmetaverse

Not coincidentally, I called it "Voxel Metaverse", thinking along the same lines as you were, and had high aspirations. It did not pan out, though we had some cool features including connecting to Minecraft servers, embedding web page content in a 3D space (including interactivity with voxel-webview, still working in the demo: https://voxel.github.io/voxelmetaverse/) and I wrote a retrospective about its successes and failures earlier this year: https://medium.com/@deathcap1/6-years-after-6-months-of-voxe... but it showed a lot of promise in what could be done to build a decentralized distributed malleable virtual world. Voxels are particularly attractive in my opinion due to the ease of content creation.

Other more recent efforts to build web-based Minecraft clients include https://github.com/PrismarineJS/prismarine-web-client and https://www.spigotmc.org/resources/websandboxmc.39415/, both are currently quite limited, but its only a matter of time/effort to complete the implementation and not a fundamental technology limitation. There are dozens of unofficial Minecraft-compatible clients, in various degrees of completeness: https://wiki.vg/Client_List

Vivecraft started in 2013 to allow a VR experience in Minecraft, and there is now an official Minecraft VR port though Vivecraft still has its fans. There's official mobile and console clients (Bedrock Edition), and although not officially interoperable with PC servers, there are also 3rd party solutions to bridge the two, including Dragonet DragonProxy and GeyserMC.

Will Minecraft blaze the way forward into what becomes The Metaverse? Honestly, maybe not. Mojang may not see the same potential in Minecraft as I do, but I feel the modding community is onto something developing projects on the edges of a hypothetical Minecraft Metaverse. If it isn't Minecraft itself, I am convinced a similar game will play a fundamental role in the development of what we come to know as the metaverse.

Voxelmetaverse, a voxel game platform based on voxel.js

Live demo: http://voxel.github.io/voxelmetaverse/

Source code: https://github.com/voxel/voxelmetaverse

I had big ideas for it but didn't get too far, after several years of working on it. Just published a retrospective today: https://medium.com/@deathcap1/6-years-after-6-months-of-voxe... - it could be developed much, much further.

Source code: https://github.com/chonyy/AI-basketball-analysis

I built an app to visualize and analyze basketball shots and shooting pose with machine learning.

https://github.com/chonyy/AI-basketball-analysis

The result is pretty nice. However, the only problem is the slow inference speed. I'm now refactoring the project structure and changing the model to a much faster YOLO model.

Thanks for asking! This is not a noob question.

I would say that the similar workflow could be applied to any ball-related sports. The object detection and the tracking algorithm is basically the same. Then, you could add any sport-specific feature!

For example, I have used a similar method to build AI Basketball Analysis.

https://github.com/chonyy/AI-basketball-analysis