Unreal is safe now, but there are no guarantees under capitalism. A FOSS license does guarantee that enshittification won’t be a factor because it literally can’t become the exclusive property of some company with a greedy executive board. Unreal doesn’t have that protection, Godot does.
Could Godot be compromised some how? Sure. Can it be enshittified? Not really.
The thing is, this could change at any time. The problem with enshittification is that it spreads. A company that’s doing great work today could be bought out by corporate profiteers and leeched of its actual value at any point in the future. We’ve had plenty of companies that started out with a vision and a set of strong principles who’ve been reduced to predatory business practices that are bad for everyone. You can’t assume that because a company seems to have integrity now, that integrity will remain.
Remember Elon Musk 15 years ago? Wasn’t quite the same, was it?
To me, sitting in a position of getting started in game development, that makes me want to sink my time and effort into an engine that I know can’t be enshittified because it can’t be bought out. I want to know that in a few years I’m not going to completely scrub every asset and mechanic that I make for the engine because somebody’s pulled some Darth Vader shit.
Bottom of the lake is pretty cheap iirc.
I don’t really even trust Unreal until Unity takes a legal hit for this. What’s to stop Epic doing the same thing?
Considering how locked into an engine a project can get, why risk a corporate engine unless you absolutely have to?
The bucket is also increasingly full of bots and astroturfers. I think they’re doing a half decent job of hiding the impact when it comes to numbers, but the drop in quality sheds a little more light.
The real question is whether or not people will continue to use Unity. Apologies mean less than nothing in a case like this regardless of whether or not they’re sincere. This is a company that’s shown their cards. Why give them business when you can go elsewhere?
Personally, this has made me start looking more into Godot. I’ve got a project I’m going to be working on that I was going to do in Unreal, but this Unity stuff has made me skeptical of tying my creative output to any one company that can’t be easily replaced. Getting that wrapped up with a proprietary platform that comes with licensing that might change just seems like a bad idea now. Maybe Unreal is okay today, but what about down the road? Why start building into a system that there’s no guarantee won’t enshittify a few years down the road?
I’d like to get my major mechanical stuff squared away and develop a visual style and then tell more stories without reinventing the wheel every time. Once I’ve got my assets built on top of an engine, I’d rather add to it over time than arbitrarily scrap it every few years. Updating and refactoring is all well and good, but I’m not in it to code the same system over and over.
That makes Godot look pretty appealing, and any closed source corporate offering look pretty shady.
Why does the use of AI to modify art require justification?
We seem to have this general culture of people who don’t make things coming after those who do. Every decision of design, methodology, or artistic preference treated as though the creator has an obligation to please every single person who posts their opinions on the internet.
The reality is that this simply isn’t true. Art that spends all its energy fretting about whether people will like it ends up being some bland bullshit produced by committee. Art that allows itself to be what it is doesn’t need opinions and suggestions to flourish.
If the author of that article were remotely interested in their process or what the actual practical implications of using AI on a project are, they could have had something worth reading.
Instead they went into the interview looking to push a position and badgering without listening rather than making even a passing attempt at something resembling journalism. Because ultimately they don’t care about AI, or art, or games; they care about rage clicks.
They can be binding in the sense that they can govern the licensing or potentially ownership of submitted assets. So like, for example, a ToS could have a bunch of clauses that carry no legal obligation for you, but could also include a clause that grants the company licensing to use your likeness or things submitted to the server or interaction with it. The same way any ToS can license the use of your metadata for sale to 3rd parties.
That doesn’t have any particular legally binding requirements of you, but it can serve as a shield in the event of a lawsuit if, say, Facebook uses your profile photo in some advertising materials.
It can also be useful if you’re running a small project like an independent game server. Even if there’s literally no money in it, it can be helpful to clarify who owns what in the event of something like a false DMCA. If a developer who once was doing work with you suddenly decides to take their ball and go home, some sort of agreement that outlines your ownership or usage rights surrounding code submitted to your mod can protect you when they turn around and send Steam a DMCA.
But yeah, nobody’s going to get sued for using a service in a way that the ToS prohibits unless it’s already illegal, like theft.
Artists aren’t lawyers and don’t want to be. Except for the ones that are. But that isn’t most of us.
Artists make art. If you want to look for the people who like to make policy, look to the jackasses in suits who sit around having meetings about meetings all day to justify scalping the work made by actual artists. The same kinds of people who fund stories like this blatantly uninformed hit piece.
Fuck them and the horse they rode in on.
At some point the line will have to be discovered, because the use of AI for art isn’t going away. Suits can whine about it all they want. Art doesn’t really care.
Real world protest on a meaningful scale is extremely dependent on population density both of individual cities and of the country in general.
In the United States we’re extremely spread out, even if we have some urban areas with incredibly high population density. The result is a situation where wide-scale protest logistics really aren’t in our favor. Even if you mobilized all of New England’s protest-inclined leftists to organize around one city, it’s hardly the numbers needed to shut anything down. We can’t just all go outside and produce nation-wide mobs in our major cities.
Look at Occupy Wallstreet. They had pretty substantial support and there were busloads of people from all over the country going to New York. But in the end it wasn’t enough to create substantial change.
Strikes have been effective because they mobilize the actual workforce that’s looking for a result, and are inherently disruptive to the thing they’re trying to change. Their demands end up being addressed because they crank up the immediate urgency of addressing them.
We need that. Targeted action that is actually effective on a wide scale. I don’t really know precisely what that looks like, but I think a good first step is the kind of action we see around remote work. People literally will quit over it and go elsewhere. Companies have done so much to leech profits out of their employment models that they’ve reduced any incentive toward long term loyalty.
They really don’t have much bargaining power right now, so people can dictate the terms a bit more. Individual decisions regarding how and when companies get to buy your labor or your creative output, as well as which companies you support financially, may well be closer to an actual path through the muck.
Outside agitation can also be incredibly useful. An outside actor in contact with people inside a corporate entity have an opportunity to speak out without the chance, legal or otherwise, of employer retaliation. Again, targeted and largely individual action.
Whatever the solution is, I suspect it’ll look a lot more like that if it’s going to challenge the status quo in the United States.
AI art of any reasonable quality still requires significant human input. I don’t just mean prompt engineering, I mean actually having an artist using more traditional techniques to make adjustments or provide a base for the AI work. The output of raw AI art on its own can be impressive at times, but it’s not consistent enough to maintain a style for any sizeable piece of work.
If you want to be able to create a bunch of assets that look like they were designed for the same project with AI, somebody still needs to do some art.
What AI does do, though, is give those artists the ability to exponentially increase their productivity independently, with no particular need for the sort of labor-hour organization that a corporation provides.
It should be telling that the corporate media spin on this is to attack it and to publicize voices that criticize it, but never those that express nuance. That’s because it terrifies every useless corporate lackey who understands its actual potential to empower independent artists of all kinds.
You’d be making a mistake there. AI elements can’t be copyrighted, but human-created elements can. There’s also a line somewhere at which point AI generation is used as a tool to enhance hand-made art rather than to generate entire pieces wholesale.
Like, let’s look at this Soul Token for my Planescape themed Conan Exiles server (still in development).
https://cdn.discordapp.com/attachments/1097400802764664843/1110453997413867560/image.png
I went into GIMP, drew a simple skull based on a design I found on google image search, slapped it on a very simple little circle, and popped it into NightCafe for some detail work. The end result is something I composed myself, with the most significant visual elements created by hand and spiced up a bit essentially using a big complicated filter. The result saved me hours and gave me one of many little in-game items to mod into my server that I never would have had the resources to produce in bulk otherwise as an independent developer.
Who owns it?
Well, I drew the skull after training myself on google image search data, but presumably my hand drawing of a fairly generic object still belongs to me. I drew the circle that makes up the coin itself, but NightCafe added some nicely lit metallic coloring, gave it a border, and turned my little skull into a gem. This, of course, requiring some prompt engineering and iteration on my part.
So is adding a texture and a little border detail enough to interfere with my ownership? Should it be? If I didn’t hand-create enough of the work to constitute ownership, surely there’s some point at which a vanishingly small amount of AI detail being added to the art doesn’t eliminate the independent creation of the art itself. If I were to paint an elaborate landscape by hand and then AI generate a border for it, surely that border shouldn’t eliminate the legitimacy of my contributions.
At some point, the difference between the use of AI and the use of a filter in an image editor becomes essentially non-existent. Yes, an AI can create a lot more from scratch, but in practical terms it’s much easier to get it started with a bit of traditional art than it is to spend hours engineering prompts trying to get rid of weird extra eyeballs and spaghetti fingers.
I’d love to see a more elaborate discussion on this topic, but so far all we get is some form of ‘AI bad!’ and then some artists dropping a little bit of nuance without it really seeming to go anywhere.
This technology has the potential to elevate independent artists to the sort of productivity that only corporations, with their inherent inspiration-killing bureaucracy, could previously achieve. That’s a good thing.
People who haven’t used this tech really have it backward. This enables indie artists to create stuff on their own without corporate oversight. This interview was an opportunity to explore that, but they decided to follow the corporate line of attacking this actually successful four person studio instead of asking about what makes them tick with any actual interest.
What a terrible interview. The interviewer literally repeatedly asks questions that they’ve already answered and shows pretty clearly that they haven’t bothered actually researching or trying AI art technology. They certainly seemed to have read plenty of articles about how bad AI is, but didn’t even bother scratching the surface of how it’s actually used.
It reads like a hit piece coming from someone who only reads what comes up in their feed.
That’s definitely a fair take, but it doesn’t really eliminate the issue. I know Beehaw offered to pay to have some of this stuff done and were turned down. With the project being as substantial as it is, maybe it wouldn’t be a bad idea to accept some of that money and hire a third set of eyes.
Lemmy needs proper moderation tools. Currently if someone posts an image you have to take like a dozen steps in a command line to delete it, and it doesn’t delete from federated instances when removed by moderators.
So let’s say you access lemmy.world through lemmy.film. Anything federated to lemmy.film through lemmy.world that’s deleted due to content moderation will still appear on lemmy.film until it’s deleted manually.
Couple this with an influx of malicious users and you’ve got a nasty situation on your hands that isn’t sustainable for instance hosts while putting them at significant legal risk.
On the user experience end of things, people seem determined to flood the all feed with bot spam and it seems as though the more hostile tide of reddit users and plants sponsored by bad actors have showed up to play.
Lemmy’s devs need to give some indication that they can even be bothered to glance in the direction of producing more robust moderation and federation tools if it’s going to keep its momentum. User-level instance blocking would help a lot too.
Literally just button remapping support for my MX Ergo.
And for the fool who always comes into these threads to tell me again that I must not have tried in several years, I tried last month. Talked to the Solaar dev, tried to reach out to Logitech, literally nothing to be done.