Long term benefiting ideas for the future of GW

[BubZ.8150]

Hi, 

I been here around forever and I'd like to contritbute my ideas and thoughts, ideas for gw2 in these changing, fast developing technological advancements.

My Idea, wich will happen nonetheless, is, yes : A.I.

People love gw2 for its authentic way of playing a social MMO, quirky gameplay and races, animations and physics, yes this game is old but who cares? GW2 is still soo much fun and still actively developed. However, the game is more than 10 years old already, and you notice a huge gap of telling 'the story' from core story, to EoD. back then we only known what we had, so its was all okay, but lately some things, like textures, shading, could be heavily improved if you let an A.I. polish the textures, the whole engine, every polygon and animation polished with A.I.

With a few clicks on  a button you can enhance the game drasticly.. really.. with A.I. empowered graphics, noone would dislike tangled depths anymore, but gaze their eyes out to beautiful leyline energies, caves, waterfalls, etc. 

A.I can help create more naturally looking animations, and maybe even optimize the whole game for people not able to play it on max settings.
Hope we gw2 veterans can still live untill that day that gw2 (and many other older remastered games) get their rightful designated place in top tier MMO's.

What do you think? would A.I. be good? or a hassle. 

If you have an idea of your own. lets brainstorm together 😄

Love you gw2 community,

Veryl

 

[Khisanth.2948]

Ah yes ... magic. I wonder why no one has thought of using magic to improve the game before.

[Farohna.6247]

That's not why people dislike Tangled Depths....

[Ashantara.8731]

2 hours ago, Veryl.7861 said:

What do you think? would A.I. be good?

What "A.I." exactly are you referring to? I've never heard of a program that could magically exhance an existing MMO's maps the way you suggested.

53 minutes ago, Khisanth.2948 said:

Ah yes ... magic. I wonder why no one has thought of using magic to improve the game before.

Abracadabra! ✨

[Randulf.7614]

I can’t wait to see six fingered npcs everywhere..

OP, just no. AI is a fun tech, but it absolutely should not replace hard working artists who pore time and effort into their work to make something that has style and substance rather than soulless and copied.
 

Let’s never suggest this again.

[Rogue.8235]

A.I. is still at the point of massive unintended consequences that are truly unpredictable.  Automating something as complex as MMO software is as likely to break it as polish it, if not more.  

I understand the desire for continuous world polishing, though.  Unfortunately, A.I. is just not well-developed at this time.  Maybe in..... never mind hopefully never.  I don't want Cylons building my MMO.  Mostly because they'll nuke us before finishing development.  It's a small quibble, but the little things matter.

 

[Eekasqueak.7850]

No that would be bad actually. 

[TheLadyOfTheRings.9148]

Who would use this AI, ANET or the players? 

[Khisanth.2948]

It is amusing to see the polarized opinions about AI that all seems to be based on the mountain of the low effort content that has been produced.

AI will definitely have a positive(probably negative as well) impact on game development. Whether or not ANet will take advantage of it is unknown. I think they should but that is going to require hiring the necessary people and having the R&D budget to pay those people.

I blame all this on OP presenting the topic with the energy and enthusiasm of a cryptobro trying to sell you NFTs.

2 hours ago, TheLadyOfTheRings.9148 said:

Who would use this AI, ANET or the players? 

Depends on what you mean by "use". For example "optimize the whole game for people not able to play it on max settings" is more or less describing DLSS 2.0/FSR/XeSS. I think FSR might be usable already without integration by the developers if you have the right GPU, not sure about the other two. Although there isn't actually any sort of optimizing going on but optimization is a term that often gets misused(as it is here) anyway.

[Rogue.8235]

2 hours ago, Khisanth.2948 said:

AI will definitely have a positive(probably negative as well) impact on game development.

AI, to include machine learning, is seriously not at a point to be able to work complex systems, such as gaming software, without breaking it.  Do not confuse AI with scripts.  Things like ChatGPT or resume screening AI, has already produced some extremely negative results that are unintended, but still present.  For example, resume screening AI does adopt racial bias on its own (due to the human-created data set it is built upon).  The way resume-screening AI was developed was thought to be completely equitable.  Again, unintended consequences are a huge issue and it's often hard to identify how such behaviors were developed by the AI.

 

Again, AI is not currently at a place to be a reliable tool for development.

AI will definitely have a negative (probably positive) effect on game development, in its current state.  This will continue to be the case for the foreseeable future, barring any sudden, major breakthroughs.  To clarify, I am referring to actual machine learning, not a series of scripts with no possibility of self-growth.

 

[Gehenna.3625]

Don't believe the AI hype. It can be useful for certain things but it's nowhere near as powerful as you give it credit for.

[Astralporing.1957]

AI is a good tool for humans to use during development process of some stuff. It's just a tool (and one of many) though. It's not a magic wand OP seems to think it is. The one thing it can currently be used for in graphics is upscaling textures or cinematics in old games for modding or for remaster versions. Notice, though, that it's still a  lengthy process that requires both human supervision and some work after that to fix the inevitable errors that would have crept in. Notice also, that while updating some of the older textures might be nice, it's hardly a major problem in GW2 - most of the updates would be in adding technologies that have been developed and added into engine throughout the years to the older areas, and that's not something that can be so easily automated.

 

[Khisanth.2948]

21 minutes ago, Rogue.8235 said:

AI, to include machine learning, is seriously not at a point to be able to work complex systems, such as gaming software, without breaking it.  Do not confuse AI with scripts.  Things like ChatGPT or resume screening AI, has already produced some extremely negative results that are unintended, but still present.  For example, resume screening AI does adopt racial bias on its own (due to the human-created data set it is built upon).  The way resume-screening AI was developed was thought to be completely equitable.  Again, unintended consequences are a huge issue and it's often hard to identify how such behaviors were developed by the AI.

 

Again, AI is not currently at a place to be a reliable tool for development.

AI will definitely have a negative (probably positive) effect on game development, in its current state.  This will continue to be the case for the foreseeable future, barring any sudden, major breakthroughs.  To clarify, I am referring to actual machine learning, not a series of scripts with no possibility of self-growth.

 

If you are referring to the crazy magical thinking OP is doing then sure. We are definitely not there yet and I am doubtful we will ever get there but that doesn't mean ML can't be used to improve development.

Even the hand issue someone was making fun(some human artists have this issue too ...) already has a solution  https://www.youtube.com/watch?v=ptEZQrKgHAg

 This is already 3 years old but the part at 4:00 reminded me of all the complaints about armor clipping. I am not say that specifically would eliminate clipping but at the very least the technique can be applied. I am not in that field so I don't know about the claims but since it is a peer reviewed presentation at SIGGRAPH I assume it should be mostly true. In what world would a tool can than reduce something that takes around 40 hours down to 5 minutes not an improvement? It also reminded me of the amount of animation reuse for EoD. Maybe they wouldn't have had to do that if they had a tool that provided similar reduction in time required to make those animations.

 

[Rogue.8235]

7 hours ago, Khisanth.2948 said:

If you are referring to the crazy magical thinking OP is doing then sure. We are definitely not there yet and I am doubtful we will ever get there but that doesn't mean ML can't be used to improve development.

Even the hand issue someone was making fun(some human artists have this issue too ...) already has a solution  https://www.youtube.com/watch?v=ptEZQrKgHAg

 This is already 3 years old but the part at 4:00 reminded me of all the complaints about armor clipping. I am not say that specifically would eliminate clipping but at the very least the technique can be applied. I am not in that field so I don't know about the claims but since it is a peer reviewed presentation at SIGGRAPH I assume it should be mostly true. In what world would a tool can than reduce something that takes around 40 hours down to 5 minutes not an improvement? It also reminded me of the amount of animation reuse for EoD. Maybe they wouldn't have had to do that if they had a tool that provided similar reduction in time required to make those animations.

 

 

Those appear to be complex algorithms incapable of adapting and growing on its own.  I didn't see any part where the software actually changes and updates itself, on its own, without user input.  The systems I was referring to actually do that.  They are created and given a database to work with, then the system, on its own, self-updates and modifies its own behaviors, without any input from any human.  That is why it is so difficult to troubleshoot when unintended behaviors manifest, the engineers have to retrace the path tthe system took in its self-modification to identify what happened.

 

[Vayne.8563]

9 hours ago, Gehenna.3625 said:

Don't believe the AI hype. It can be useful for certain things but it's nowhere near as powerful as you give it credit for.

That's funny, I remember people saying the exact same thing...about computers. lol

[Zephire.8049]

AI is no substitute for a human's training and skill. Using AI to help upscale textures wouldn't be bad, but animations would need to be done by human (or at least the key frames, which is how it's already done) lest they break or enter uncanny valley territory, something humans can sense and steer away from while working on it but AI would have no way to sense what would be natural movement and what by all accounts should be fine but human brains pick out minutia that make it very wrong.

Then the process would have to happen all over again with hope the AI gets it right this time.

AI shines best when it's used to augment a person's skill or with a set, defined limited scope. Things like scaling textures and merging textures seamlessly, not creating animations from scratch or touching the engine, because despite the name, AI isn't smart—it's coding that does something it's told and needs human oversight. And once it gets into generating graphics, that gets a lot more ethically and legally dubious.

And keep in mind some of what the game has (low res textures, low polygons, simplified animations, etc.) isn't because they couldn't be done better but because it would be too resource intensive—for both developers and customers' computers—to manage. It's a choice, not a lack of ability, when it comes to using low res or few polies for something that most people don't notice (e.g. the infamous snake model in Team Fortress 2 that caused major issues for some players because it had an obscene amount of polies despite being an ambient creature most wouldn't have noticed except for the issues it caused) or when accounting for it being an MMO and having to render multiple people at once wearing and doing countless things versus a single player game where everything can be accounted for thus textures/models/animations can be more detailed.

[Astralporing.1957]

44 minutes ago, Vayne.8563 said:

That's funny, I remember people saying the exact same thing...about computers. lol

And if you read the deeper explanations for it from that time, they're still as true today as they were before.

[Vayne.8563]

1 hour ago, Astralporing.1957 said:

And if you read the deeper explanations for it from that time, they're still as true today as they were before.

AI is going to change the world on the same level computers have changed the world already. There's computers in our cars, our phones, we do our banking, our shopping, meet our spouses (I did anyway) online.  The computer, the web and AI are all changing the way we do things.

That said it's not going to happen tomorrow. We're still in our first baby steps of AI, but I don't think anyone should underestimate would could be.

[zengara.8301]

just for reference, I think OP talking about the new Fortnite feature that was presented, where you could turn the game completely different, into a cool MMORPG looking game (Asmongold) It looked cool, and super easy, but I really do not know how real it is, and how much of it is "movie magic" Like actually pre-recorded stuff

EDIT: Unreal Engine 5.2 Looks super cool, and makes making game insanely easy....If what they say is true. But apparently the new fortnite feature can be used right now? Not a fortnite player

Edited March 24, 2023 by zengara.8301

[Astralporing.1957]

1 hour ago, zengara.8301 said:

just for reference, I think OP talking about the new Fortnite feature that was presented, where you could turn the game completely different, into a cool MMORPG looking game (Asmongold) It looked cool, and super easy, but I really do not know how real it is, and how much of it is "movie magic" Like actually pre-recorded stuff

EDIT: Unreal Engine 5.2 Looks super cool, and makes making game insanely easy....If what they say is true. But apparently the new fortnite feature can be used right now? Not a fortnite player

From what i can gather (from googling, as i do not play this game either) the fortnite feature you speak of is an effect of an engine upgrade. Notice, that engine upgrade is not trivial even when you follow an update path withing the same engine family (UE in case of Fortnite), and that most of the time just upgrading does not mean the stuff you've had done for the old engine version will suddenly start using new features. Most of the time you'd have to do a rework for that old stuff.

You can already see this in GW2, btw. Since original launch the game went through series of incremental updates to its engine done quietly in the background. That's why the EoD looks significantly better than core did 10 years ago. That saying, core still looks the same nowadays as it did then, because while all those engine updates allowed Anet to introduce new graphics of better quality, it does not automatically upgrade old ones. Someone would have to redo them still.

In short - no, there's no magic button (called AI) that can make everything better when pushed. We're still talking about massive amount of work, most of it being done by humans, not machine learning-based software.

[Gehenna.3625]

8 hours ago, Vayne.8563 said:

That's funny, I remember people saying the exact same thing...about computers. lol

I know AI will get there, but it's not as far as companies want you to believe. That's all I'm saying. We had our first PC in 1986 and it was really cool, but it has progressed a LOT in the years since. AI will no doubt make even larger and faster strides. But all I'm saying is that what companies want you to believe about AI is not what AI can really do.

Point in case...look at the AI chatbots like GPT and the one Microsoft implemented. People already found ways to break them, just by talking to them for a while. So no it doesn't live up to the hype, but I'm sure at some point they will.

[Vayne.8563]

6 hours ago, Gehenna.3625 said:

I know AI will get there, but it's not as far as companies want you to believe. That's all I'm saying. We had our first PC in 1986 and it was really cool, but it has progressed a LOT in the years since. AI will no doubt make even larger and faster strides. But all I'm saying is that what companies want you to believe about AI is not what AI can really do.

Point in case...look at the AI chatbots like GPT and the one Microsoft implemented. People already found ways to break them, just by talking to them for a while. So no it doesn't live up to the hype, but I'm sure at some point they will.

Yep, I agree, with you.  It's just easy to dismiss stuff cause it's not there yet. My bad, I didn't get where you were coming from.