Star Citizen, Cyberpunk, and the future of video games

a reply to: SaturnFX

What is really going to be mindblowing is when AI starts kicking in and making levels and npcs.
Take technology today like the social brain of Replika (AI chatbot that is actually good)

Yes it will be amazing when AI can generate 3D assets simply by describing those assets or showing it some pictures. We already have some AI which can sort of do those things but it's not really at a commercial stage yet. But I have high hopes after seeing the latest AI's for generating 2D art. Even some of that 2D art can still be used in games, for things like posters on walls and other flat textures.

It would also be pretty cool to see AI chat bots used to power NPC's in games, but unfortunately I think that's a pretty long way away. There are two main reasons, the first is because those types of AI require a huge amount of memory in order to act intelligently. Right now you basically need a supercomputer to run the best models like GPT3 and LaMDA. The second problem is if companies can't fully predict the NPC behavior and dialog it could raise legal issues and prevent their game from getting a rating.

originally posted by: ChaoticOrder
a reply to: yuppa

I believe it's just over a decade old (10 years) and I highly doubt they will switch to UE5 considering how much time and money they have spent customizing their CryEngine/Lumberyard engine. Even when Star Citizen is fully released, the money wont dry up, if anything sales will increase. And it's not like they will stop selling ships when they release the full game. It will continue to make money like GTA5, although I hope they don't make it a grind to buy ships in game. I've been playing Star Citizen the last few weeks (hence my inspiration for this thread) and right now it isn't super hard to earn money in the game and buy the best ships. You can make a few million in a single day mining with the prospector.

I think Roberts sees this as his cash cow and is in no hurry to release it EVER. yandere simulator is in final stages now and that was a running gag about Star citizen would be out before that game. looks like the opposite is true.

a reply to: yuppa

I find it more funny that so many people expect a game like Star Citizen to be made at the same speed as any other game. And that's the entire point I'm making in this thread, if you're trying to make an almost photorealistic game which has a massive game world, then it's going to take an excruciatingly long time to finish building the game, unless we employ AI and other tactics to speed up the development. I don't want to wait until the 2030's until we get another Fallout game, but according to the development timeline provided by Todd Howard, it will be the 2030's before we get Fallout 5. And games are only going to take longer and longer to make as the demand for photorealistic graphics increases.

originally posted by: ChaoticOrder
a reply to: yuppa

I find it more funny that so many people expect a game like Star Citizen to be made at the same speed as any other game. And that's the entire point I'm making in this thread, if you're trying to make an almost photorealistic game which has a massive game world, then it's going to take an excruciatingly long time to finish building the game, unless we employ AI and other tactics to speed up the development. I don't want to wait until the 2030's until we get another Fallout game, but according to the development timeline provided by Todd Howard, it will be the 2030's before we get Fallout 5. And games are only going to take longer and longer to make as the demand for photorealistic graphics increases.

bethesda doesnt want to use a easier engine,and thats why it will take around 8 yrs.

Roberts kept switching engines, added more and more content to it. Squadron 42 is practically done and he wont release it. This isnt EVE online,he has to stop being such a perfectionist for a space sim shooter,that once played and beat people will drop.

a reply to: yuppa

Bethesda's engine and their Creation Kit is actually very easy to use once you get the hang of it. Even Bethesda designs their games and their DLC's using tools similar to the Creation Kit, which is why they usually come as esm files. The only thing holding Bethesda back is the size of their team and their budget. They aren't trying to make games which have photorealistic graphics, so it shouldn't take them so long to make a game.

The main issue is the size of the open world games they make rather than the quality of the games they make. If they just had more man power, especially by hiring modders who are already familiar with the Creation Engine, they could get a lot more done. I'm sure Bethesda has the budget for it, but they don't seem to be expanding quite as much as they should be. They should have at least two different teams who are dedicated to Fallout and The Elder Scrolls.

originally posted by: ChaoticOrder
We are approaching a point where video games will have the ability to render photorealistic scenes. This is especially true with newer game engines such as Unreal Engine 5, which can render an almost unlimited amount of detail. The real limitation is now on what the artists can create, and I would argue that we are already seeing this limitation. Many games use lower quality assets simply because that's all they can afford, but they can often use stylized art and innovative gameplay mechanics to make their game appealing regardless of that.

Now that game assets are so hard to create, they can take up most of the development budget, which often takes away from other important things. Many modern games have pretty graphics, but they are completely lacking in terms of gameplay depth because they didn't focus enough on the coding. We can see this with games such as Cyberpunk, it looks great but is lacking many game mechanics that we would expect from an open world game created by a AAA development company, despite the fact many older open world games have those features.

In the past it wasn't necessary to spend so much time on creating art because our computers were very limited, and that allowed developers to put much more time into the code. There are many old games which are still praised for their complicated yet bug free game mechanics, some mechanics which developers still struggle with. A big part of the problem is due to developers being limited by the game engine they are using and the coding tools it provides. Game engines make it easy to create a good looking scene, but creating good game mechanics requires skill.

Many of the old classics didn't use a game engine, they wrote their own engine or they used a bare bones engine which they would highly customize for their own needs. This gave them a great deal of control and allowed them to implement unique game mechanics that would be a pain to replicate using only the tools provided by a game engine. A truly great game requires that adequate resources are allocated to both art and programming. An open game world can look photorealistic, but it wont feel very immersive until we bring the world alive with clever programming.

Personally, I feel that Star Citizen is one of the very few games to achieve this feat without sacrificing graphics quality or game mechanics. The graphics are almost photorealistic and to give you an idea of how complex Star Citizen is, almost every key on a full size keyboard has a function. For those unaware, Star Citizen is an online space simulation game, there is also a single player narrative-driven version of Star Citizen called Squadron 42, both of which are still in development. Star Citizen is currently playable and it has been making great progress recently.

It has massive planets covered in massive cities, and the quality of the graphics really makes it feel like you are living in a future civilization. They are obviously using a game engine but they have heavily customized the graphics pipeline to support the mind-boggling scale of the world they are trying to render. But it's not just the graphics which set Star Citizen apart, it's the complexity of the game mechanics which also draw me in. They are aiming for realistic physics, for example planets have different gravity, ships have momentum and fly in a plausible way, etc.

You could be flying at a fraction of the speed of light, and get out of your seat and walk around the ship, and some how you wont be flung into space, even when you get lag from a slow internet connection. Now obviously Star Citizen has its fair share of bugs, which is one of the main criticisms I hear, but it's to be expected with a game so complicated and so massive. Star Citizen and Squadron 42 have been in development for about a decade now, but they are still considered to be in a fairly early stage of development, and it's not simply due to a lack of funding.

GTA5 is the only game which had more funding than Star Citizen, the real issue here is the sheer scope of Star Citizen and the quality they are aiming for. It's not something that can be made quickly even with a large team, and this isn't just a problem which is impacting Star Citizen. As I said, many modern games require immense resources dedicated to art assets, because people expect the latest games to make use of the latest hardware. But now our hardware is starting to outpace our ability fully utilize it, and forcing us to cut resources in other areas.

As much as I hate to say it, the the solution to this problem is more AI generated art. The latest AI models such as DALL-E 2 can generate art which is comparable to an expert artist, but it can do it much faster. Another important part of the solution is photogrammetry, which involves scanning real world objects and turning them into 3D models we can use in a game world. I remember the first game I played which made use of photogrammetry was The Vanishing of Ethan Carter, and it's still one of the most beautiful games I have played to this day.

There's just something about using real world objects which makes a video game feel way more immersive, like your brain is much more willing to accept what it is seeing as being real. Even though you might not notice any obvious differences between a hand-crafted rock and a rock scanned in from the real world, your brain will notice the subtle details. That applies to not only the shape but also the textures which are applied to the mesh, it's much easier for your brain to accept when the textures are essentially photos of real objects.

Star Citizen gets a lot of criticism for the slow development, but I do fully understand why it's taken so long to build, the quality they are putting into it is unprecedented. It always amazes me when I watch their old trailers and see how far ahead of the curve they were compared to everyone else. I'm probably more hyped for Squadron 42 than any other game, even though it might not be released before I'm an old man, if it ever gets released. That's why it's so crucial for their team and other development studios to make use of technology to speed things up.

The future of gaming looks very bright if we can learn the important lessons taught to us by older games, and improve on that with modern technology. Star Citizen gives us a glimpse into what this future could look like when you merge amazing graphics and immersive gameplay. That's why I'm so excited for Squadron 42, it will be Star Citizen but with a story line, character development, and none of the online networking issues. If the elevators work reliably then I'll be happy. I'll end with this epic trailer for Squadron 42 which uses in-game footage:

Nice read, I don't play modern games so it's nice to hear the reflections from someone who does keep up.

I don't know if I agree about the need for future AI-created graphics, but maybe.

There's definitely good sense in different games re-using basic assets and game engines, instead of every single game being created from the ground up.

...Squadron 42 looks amazing but I'm wondering if it will ever be released... or if it's even possible. The scope seems impossible! But if they manage some form of online set-up, it will change the world.

Also for whoever is interested in Squadron 42, I'm not trying to be funny but you guys should probably be playing around with Google Earth. Not a joke. There aren't space-ships but Google Earth is essentially the world-building that's described for the concept of Squadron42, Google Earth already has the entire planet mapped out, for the user to fly around and "land" wherever you want. The street-view is built from photographs and lets you essentially walk around the area, anywhere on Earth.

Plus you can fly to the moon and Mars, and fly around the entire thing.

I never hear about people playing with Google Earth like this, but I've been enjoying it.

Haven't played Star Citizen yet, but played No Man's Sky a lot back in 2018. So far the biggest differences of the two is that Star Citizen seems to have a lot more procedurally-generated cities I think? And overall a better plot maybe? But I haven't seen any unique wildlife on it, though. Not like No Man's Sky, anyway.