Would you be okay with procedural generation if it meant endless, good content?

For the purpose of this post, let's consider procedural generation to include everything from environments to NPCs to missions, stories and loot. Also, AI generation is still just procedural generation. You are still just using algorithms.

A lot of people have mixed feelings about procedural generation. It hasn't always been very good, failing to come anywhere close to hand-built. From the terrible 7 Days to Die initial implementation of procedural map-making (it's gotten much better now) to early attempts at AI creating dialogue, a lot of people are highly skeptical of games that use procedural generation for much of anything.

But procedural map generation has improved dramatically through the years. Games like Warframe and Remnant have used ingenious combinations of human-made assets procedurally generated. And AI story-telling and dialogue won't get any worse, certainly. Like AI art, it's only going to improve. Several games have even procedurally generated creatures and at least one has randomly created humans. Meanwhile, random loot drops, for good or for bad, have been in practice for a long time.

So would you be in favor of a largely procedurally generated game if it were done well and meant endless play? For me, I find a game that I love, like Remnant 2, and I think that if they could have found a way to procedurally generate a continuation of the game that was faithful to the human-made portion that I would still be playing it today.

What do you say? Skepticism about whether procedural generation will ever reach a high enough quality is fine, but I think it is short-sighted. I think it's very nearly reaching that point right now, much less 5 to 10 years from now. Do you have ethical concerns about it? Surely more developers would be laid off if procedural generation became good enough. Thoughts?
 

Brian Boru

King of Munster
Moderator
Well since I already get endless good content via ProcGen, then I must answer a resounding "YES"!
Another way of saying that is I recently spent yet another enjoyable month playing Civ4 :)

However, ProcGen on its own isn't enough to make it 'endless'. Civ4 allows a huge amount of default customization, and then the mods I use flesh that out a lot—ie they also have a lot of customization within them—so I can make it the game I want to play 'endlessly'.

So the key for me is that devs provide the framework and tools for modders and players to make their own 'endless' gaming experience.
 
Well since I already get endless good content via ProcGen, then I must answer a resounding "YES"!
Another way of saying that is I recently spent yet another enjoyable month playing Civ4 :)

However, ProcGen on its own isn't enough to make it 'endless'. Civ4 allows a huge amount of default customization, and then the mods I use flesh that out a lot—ie they also have a lot of customization within them—so I can make it the game I want to play 'endlessly'.

So the key for me is that devs provide the framework and tools for modders and players to make their own 'endless' gaming experience.
I thought you'd moved to Civ6?
 
  • Like
Reactions: Pifanjr and Alm
Well since I already get endless good content via ProcGen, then I must answer a resounding "YES"!
Another way of saying that is I recently spent yet another enjoyable month playing Civ4 :)

However, ProcGen on its own isn't enough to make it 'endless'. Civ4 allows a huge amount of default customization, and then the mods I use flesh that out a lot—ie they also have a lot of customization within them—so I can make it the game I want to play 'endlessly'.

So the key for me is that devs provide the framework and tools for modders and players to make their own 'endless' gaming experience.
I would love to see a fantasy Total War that goes full-on procedural generation: maps, units and everything. And, MOST IMPORTANT, I don't want it balanced to the point of boredom. Maybe sometimes you get a faction where the units just aren't great, and you just have to figure out how to deal with it. Maybe you get a faction with nothing but cavalry units. Maybe you get all archers except for one great infantry unit, but you can only have so many at a time. Make the new and surprising a regular part of the experience. What I'm describing could really only be accomplished with proc-gen.
 
From a purely selfish viewpoint, I would enjoy procedurally generated games if there were no downsides. I don’t know if procedural generation is only good in single player games.

From an ethical viewpoint, imo, AI taking jobs sucks but it’s never going away and it will snowball as money is the driving force and not empathy.
This won't save a lot of jobs, but even with proc-gen, some people are better at getting what they want than others. Artists, for instance, are better at getting AI to generate better art. So I think there will always be room for humans in the process. But it's very scary right now. It almost makes me sick to my stomach to think of the potential loss of jobs, so I hope there are always going to be successful human made games, and I'm pretty sure there will be, even if it's just people working alongside AI to create something wonderful.
 

Frindis

Dominar of The Hynerian Empire
Moderator
I don't have a problem with that type of content, but I would not think it would be endless of good content. Humans are not intelligent enough for that thus AI so far suffers in that area. If that does happen, then we will have a large discussion about whether the AI would be able to gather enough information about the human genome to become vastly superior in those areas or perhaps not, I dunno. I guess my point is that we decide creativity until we don't anymore. :grin:
 
Last edited:
I usually get bored of games well before I reach their end/seen all of their content, so whether the content is endless doesn't really matter. The content being more varied because of procgen is more important and might get me to play a good game longer, though if I'm done with the gameplay loop it doesn't really matter how varied the content is.

I have no ethical concerns about its impact on jobs. Jobs have been lost to automation before and it will happen again. If the people who lose their jobs to AI get into (financial) trouble because of it that indicates there's a problem with the society they live in, not with the use of AI.
 
Last edited:
I usually get bored of games well before I reach their end/seen all of their content, so whether the content is endless doesn't really matter. The content being more varied because of procgen is more important and might get me to play a good game longer, though if I'm done with the gameplay loop it doesn't really matter how varied the content is.

I have no ethical concerns about its impact on jobs. Jobs have been lost to automation before and it will happen again. If the people who lose their jobs to AI get into (financial) trouble because of it that indicates there's a problem with the society they live, not with the use of AI.
History is filled with tech advances that killed jobs. As I said before, the story of John Henry comes from the 19th century, so we've been concerned about it for a long time. I do think, however, that AI presents the problem on a scale never before seen, and adapting to it may be a longer, more painful process.
 
History is filled with tech advances that killed jobs. As I said before, the story of John Henry comes from the 19th century, so we've been concerned about it for a long time. I do think, however, that AI presents the problem on a scale never before seen, and adapting to it may be a longer, more painful process.

I doubt AI will be on the scale of the industrial revolution, at least as its impact on jobs is concerned.
 
  • Like
Reactions: ZedClampet

Brian Boru

King of Munster
Moderator
I thought you'd moved to Civ6?

Oh yeah, did. Plenty of fun there—especially how they made the Barbs a real PITA—will dip in again. But fundamental problem is devs seem to have confused complexity and strategy, ie add in all sorts of unique requirements for everything to get its best value.

AI taking jobs sucks

I bet in 10 years AI will have created far more jobs. Remember how computers were going to put clerks and secretaries out of a job? The world works that way—eliminate buggy whip and carriage building jobs, create 10x as many jobs in auto industry.

If the people who lose their jobs to AI get into (financial) trouble because of it that indicates there's a problem with the society they live in, not with the use of AI.

Second the well-said! Advanced societies plan ahead and cater for transition hardships.

I doubt AI will be on the scale of the industrial revolution

I guess we'll see if AI has a similar off-setting aspect to it

I expect AI will have a similar transformative effect as the IR.

Civilization mostly started when we discovered agriculture, which enabled cities, which vastly improved the scope of human pursuits and the standard of living.

Industrial Revolution got the ~90% who were working the land into far more productive pursuits which vastly improved the standard of living.

AI, when combined in coming decades with machine learning, robotics—and maybe quantum computing and nuclear fusion, nice but not necessary—Internet of Things in industry, the Cloud, battery advances, new materials, etc etc… it will make a colossal difference, and if managed properly—ie minimizing control by greed—can bring an era of plenty where people don't have to do the daily grind to survive in some comfort.

"Work is for horses and machines", and there's no need it has to be coupled with income. Humans who choose to can then concentrate on the huge challenges coming up, humans who want to create, can, while those who want to play games all the time can do that.
 
Yes, obviously. Who wouldn't?

it depends what type of game it is really as to how much work that would require. Scale of play space is a key. if its a platformer its way easier than say a game like NMS. Games making up the map out of tiles on the fly have existed for years. Diablo 2 used to do it... but i assume everything has to be made on the fly?

Too many people still have old crappy PC now to push procedural generation fully yet. It all requires processing power that only a small percentage of users have. Release game, need 4090 to get 60 fps at 1080p... um... sounds like Cities Skylines 2.

We don't have true AI now... just a lot of people trying to get ahead of the wave. We have smarter search engines. And photoshop has worked out how to remove the user from the process. I wish that was free, pretty sure even I with no artistic skill could think of a sentence to write.
 

Zloth

Community Contributor
Oh good, Zed is making topics again!
For the purpose of this post, let's consider procedural generation to include everything from environments to NPCs to missions, stories and loot. Also, AI generation is still just procedural generation. You are still just using algorithms.
That's a lot! No Man's Sky is the closest that I've played. Daggerfall and Dwarf Fortress have a lot, too.
What do you say? Skepticism about whether procedural generation will ever reach a high enough quality is fine, but I think it is short-sighted. I think it's very nearly reaching that point right now, much less 5 to 10 years from now. Do you have ethical concerns about it? Surely more developers would be laid off if procedural generation became good enough. Thoughts?
Laid off? Well, some would, but somebody has to deal with all that AI and procedural stuff. There will probably still be some sort of main story that people are going to want to tell - and that might well require hand crafted levels at the end of the game. Tutorials will be extra hard, too, so they might stick to hand crafting there as well.

Anyway, I'm fine with it.

The big problem I see with this kind of stuff right today is that it can't deal with the rest of the game world. Any story made will be entirely its own thing because the procedural generation doesn't know about the other procedurally generated stories, thus they can't play off each other. You'll end up with a game that feels more like a grab bag of short stories rather than a novel.

I wonder if that's what Dwarf Fortress is trying to fix by starting its generation way back at the dawn of time?
 
The big problem I see with this kind of stuff right today is that it can't deal with the rest of the game world. Any story made will be entirely its own thing because the procedural generation doesn't know about the other procedurally generated stories, thus they can't play off each other.
That won't be a problem for long if it's even a problem today.. Any big proc-gen project now would almost certainly involve AI, which would track all the previously created stories.
 
For the purpose of this post, let's consider procedural generation to include everything from environments to NPCs to missions, stories and loot. Also, AI generation is still just procedural generation. You are still just using algorithms.

A lot of people have mixed feelings about procedural generation. It hasn't always been very good, failing to come anywhere close to hand-built. From the terrible 7 Days to Die initial implementation of procedural map-making (it's gotten much better now) to early attempts at AI creating dialogue, a lot of people are highly skeptical of games that use procedural generation for much of anything.

But procedural map generation has improved dramatically through the years. Games like Warframe and Remnant have used ingenious combinations of human-made assets procedurally generated. And AI story-telling and dialogue won't get any worse, certainly. Like AI art, it's only going to improve. Several games have even procedurally generated creatures and at least one has randomly created humans. Meanwhile, random loot drops, for good or for bad, have been in practice for a long time.

So would you be in favor of a largely procedurally generated game if it were done well and meant endless play? For me, I find a game that I love, like Remnant 2, and I think that if they could have found a way to procedurally generate a continuation of the game that was faithful to the human-made portion that I would still be playing it today.

What do you say? Skepticism about whether procedural generation will ever reach a high enough quality is fine, but I think it is short-sighted. I think it's very nearly reaching that point right now, much less 5 to 10 years from now. Do you have ethical concerns about it? Surely more developers would be laid off if procedural generation became good enough. Thoughts?
Kind of a loaded question in the thread title ;)

Proc gen is really fun in games like Against the Storm or Dead Cells. In those cases surely it has to be monitored and refined by humans to make sure you dont end up with levels that are impossible to complete or parts that are just broken. Its randomized puzzle pieces, but those pieces are crafted by humans who cleverly put them together and edit the pieces that go into the 'AI' mixer.

If we are talking about a game generated by an AI entirely from concept onwards from having a load of other games fed into it its complicated. Its going to depend on what you look for from games, or individual games. Am I interested in a big Sony style narrative game made entirely by AI based off of other games that makes comments on the human condition? Not in the slightest, I wouldnt artistically value an AI book or movie either. Could an AI generated co-op survival game be fun to play, with friends or whatever? Sure.

As a tool for human developers AI generated dialogue and might be fine in places to pad things out, but an entirely AI generated story concept is only acceptible if the story is totally superfluous anyway. To a lot of people who play games, thats not the case.
 
bashing a stack of different genres together isn't as easy as some people say... it requires lots of testing to make sure it works properly. More systems you put together, more unexpected results.

The big problem I see with this kind of stuff right today is that it can't deal with the rest of the game world. Any story made will be entirely its own thing because the procedural generation doesn't know about the other procedurally generated stories, thus they can't play off each other. You'll end up with a game that feels more like a grab bag of short stories rather than a novel.
the simple answer is the generator has to do it all. Not just the game engine but all the story as well. If/when we have super human AI that can simulate the game before its made, then we will be in a better spot. Or dead, depends on your outlook :)
 
Kind of a loaded question in the thread title ;)

Proc gen is really fun in games like Against the Storm or Dead Cells. In those cases surely it has to be monitored and refined by humans to make sure you dont end up with levels that are impossible to complete or parts that are just broken. Its randomized puzzle pieces, but those pieces are crafted by humans who cleverly put them together and edit the pieces that go into the 'AI' mixer.

If we are talking about a game generated by an AI entirely from concept onwards from having a load of other games fed into it its complicated. Its going to depend on what you look for from games, or individual games. Am I interested in a big Sony style narrative game made entirely by AI based off of other games that makes comments on the human condition? Not in the slightest, I wouldnt artistically value an AI book or movie either. Could an AI generated co-op survival game be fun to play, with friends or whatever? Sure.

As a tool for human developers AI generated dialogue and might be fine in places to pad things out, but an entirely AI generated story concept is only acceptible if the story is totally superfluous anyway. To a lot of people who play games, thats not the case.
I didn't mean no human involvement, but I do disagree about AI stories. I think they can and will be good, at least as good as most game stories are today, which is not that great.
 
I didn't mean no human involvement, but I do disagree about AI stories. I think they can and will be good, at least as good as most game stories are today, which is not that great.
I agree on stories. They're all based on what humans have already written, so a good AI should be able to create a decent enough story.
We'll have to disagree then. I do agree most games dont have great stories, so AI there could be used if the story is useless. But theres a lot more subtlety to story telling in games that hasnt been explored yet, environmental details as well as dialogue and exposition and the game mechanics themselves. I dont think AI can manage that level, and it also has no intent which is the most important part.
 

Brian Boru

King of Munster
Moderator
Too many people still have old crappy PC now to push procedural generation fully yet. It all requires processing power that only a small percentage of users have

I expect that problem will be solved by the network, not individual users' hardware. Iow the hardware will be in the cloud and WiFi #6 or whatever will stream it to everyone everywhere FTL.

Oh good, Zed is making topics again!

Did you stock up during the respite? I got lots of extra sleep and protein shakes, even got a bit of work done—so I feel ready again.

Any story made will be entirely its own thing

ProcGen's greatest contribution, escape from the spoon-fed approach which is well covered by books, movies, TV.

at least as good as most game stories are today, which is not that great.

Give it time, when VNs get their time in the sun it will attract better writers. Walking simulators have potential to tell good stories too, there are already a few decent YA ones I'm told. Beyond that tho, fish out of water once gameplay enters the picture… so to speak :D

AI … has no intent

I don't follow. How does that happen?

No plot, or theme, or setting, or character dev, or what?
 
  • Like
Reactions: Zloth and Pifanjr
not individual users' hardware. Iow the hardware will be in the cloud and WiFi #6 or whatever will stream it to everyone everywhere FTL.
I don't want to rely on a server to be able to play a game. Enough of that already. There are still some games you can play without being attached, though number is dropping. Wouldn't matter if games worked on launch and still didn't get server problems

Also, that just moves problem. It still relies on people having fast internet so instead of just the people who can't afford a better PC, it could be that entire countries can't play as they not advanced enough. I know for a fact that Australia wouldn't be fast enough. Cities might be but rural is not great. I don't know what ping is like on Starnet.

Much prefer to invest the money at my end on hardware than to pay more for internet access that may not always work due to factors outside of my control. Cheaper in long run as its a one off cost, not monthly.

I don't see my life's purpose is to keep buying things. I wasn't born a consumer.
 

Brian Boru

King of Munster
Moderator
I wasn't born a consumer

Oh dear, no nappies as a baby. No wonder you… uh, never mind.

I don't want to rely on a server

I'm not talking about today's situation, rather about the "combined in coming decades" techs I mention in post #13.

Whenever that happens, fast internet will be like fast electricity is today—I'm surprised it hasn't been designated an utility already. Do you have your own electricity generator?

There will of course be places which won't have it, just as there are huge swathes of the planet without reliable electricity or drinking water today. But the tech should be there to make it happen, any absence will be down to how we manage ourselves as a species.

that just moves problem

No, it vastly reduces the problem. There is an incredible amount of waste currently with consumers and businesses having all that processing power in their own premises. Just like everyone having their own electricity generator, digging their own water well, growing their own food and transporting their trash to the dump every week.

That's what I mean with "the tech should be there to make it happen", massive savings which can be spent on deploying the tech—unless the greedy manage to control it for themselves.

Cheaper in long run

No, nowhere close—massively dearer.
 

TRENDING THREADS

Latest posts