More thoughts about games on the web

A few month ago, I wrote an article featuring my thoughts about games on the web, and why I think it's necessary to create a dedicated game engine for the web.

A few things happened since then :

My game engine, Fibbo improved a lot and will soon turn 1 year old (first commit was on the 29th of April 2024¹).
levelsio recently popularized a new coding trend, called "Vibe coding" (basically using LLMs to generate code) which enabled web developers to create games pretty easily².

I think it's time to write a follow-up to my previous article, and share my thoughts about the future of games on the web.

Vibe coding

The recent "Vibe coding" trend emphasizes a few important points :

Games on the web are cool

As I said in my previous article, if the web is the best platform for communicating with your friends, it's not so far from being the best platform for playing with them too.

It cleary shows that even if these games can be very basic and ugly-looking, they benefit from the fact that they are on the web, and thus easily accessible.

LLMs struggle with game development

The reason all these web developers are able to create games so easily using AI is because AI knows the technologies implied in the process.

I've seen people asking why they're not using traditional game engines like Unity or Godot, and the answer is simple : LLMs don't know these engines, and they don't know how to use them.

LLMs are trained on whatever data they can find on the web, so their efficiency is limited to what's available on the web. And what's available on the web is mostly web technologies.

But when it comes to games, the best technologies are proprietary, and there are not so many resources available. How many games are open-source ? Not so many.

If you don't know what the metaverse is, it's a concept that was popularized by Facebook (now Meta) in 2021. The idea is to create a virtual world where people can interact with each other and with digital objects in a more immersive way.

The metaverse is often associated with virtual reality (VR) and augmented reality (AR), but it can also be seen as a more general concept of a shared digital space, focused on social interactions.

Meta's Metaverse ended up not being a big success, but the concept of a metaverse isn't dead.

For example, Roblox can be considered a form of metaverse. It allows users to create and share their own games and experiences, together, in the same world. The fact that players can reuse their avatars across different games is also a big part of the metaverse concept.

Inspired by these ideas, levelsio initiated the concept of "vibeverse", which defines a protocol to travel across different games by integrating portals in each game.

People then started to wonder if a standardized -and open- technology could be used to create an open metaverse. And I'd like to see that too !

The problem is sharing an avatar across different games is not that easy. Roblox can do it because it has a centralized system and technology. But the web is decentralized and has tons of different technologies.

That's one of the reasons why I started working on Fibbo : creating a common technology for web games will make possible new standards protocols.

Open source games

Games being made with web technologies, by web developers, will very likely increase the number of open-source games. And that's a good thing, because in the massive amount of games that will be created, some will actually be good.

This will not only increase AI's habilities to create games, but also improve the general quality of learning resources available on the web.

Moding

One very cool aspect of open-source games is that they can be modded easily. Modding has always been a big part of the gaming culture, and it's a great way to learn how to create games.

Some popular games actually comes from mods, like Counter-Strike, which was a mod of Half-Life, or Dota, which was a mod of Warcraft III.

And some games still have a big community of modders, like Minecraft, Skyrim, or GTA V.

E-sport

You were probably not expecting me to talk about e-sport in this article, but I think it is quite possible that open-source games will be the future of e-sport.

First, modding being easier will make games easier to cast, for example by adding a missing spectator mode or tracking data more easily.

Second, it means games won't be proprietary, and thus competitions won't be controlled by a single entity. It will be easier to create competitions, and it will be easier to create a competitive scene. For example, League of Legends' competitive scene is controlled by Riot Games, and third-party tournaments are not easy to organize. But if games were open-source, any entity could create any competition, and it would be possible to have a competitive scene without any private entity interfering with it.

Lastly, it means we can change the rules at any time. Don't like the current meta ? Change it. Don't like the current map ? Change it. If your changes are good, they will be adopted by the community, and a new way to play the game will emerge.

The same way traditional sports do not belong to a single entity, e-sport should not. As far as I know, you can play football with your friends without FIFA's permission or supervision.

Why should it be different for e-sport ? Right now, because games are very expensive to make and thus proprietary, but making games open and way easier to create will change that.

AI's limits

One of the reason I started working on Fibbo was because I wanted to enable web developers to create games easily. It's interesting to see that the solution to this problem right now is using AI, but I think it has its limits.

The reason AI is heavily used here is because underlying technologies are not accessible enough. These web games uses libraries like Three.js for 3D rendering, which isn't that easy to use for the average web developer.

But then they need physics, so they use libraries like Cannon.js or Rapier, which needs to be integrated with Three.js correctly, and then they need a way to manage sounds and music but the HTML Audio Api sucks and available libraries are not that great,...

Not even talking about multiplayer games, by handling networking, through WebRTC or WebSockets.

In the end, the AI not only spends a lot of time reinventing the wheel by creating its own engine, but depending on your code, the AI's context window might not be enough to keep working on your game.

If we want good and reliable web games, we need a good web game engine.

Who's gonna make it ?

I don't know !

I must say creating the all-in-one game engine I'm dreaming about is a pretty insane journey.

If you're interested in my bet on this, feel free to give a star to Fibbo on GitHub, and maybe even start using it for your games !

Thanks for reading.

Le futur des jeux vidéo, c'est leur accessibilité.