How does Source 2 not make it onto this "Market Map"? It's mostly used by Valve and a smaller number of studios than something like Unreal, but it's viable none the less. I suspect it has a much broader user reach than several of the entries on the Open Source end of this map.
It's also missing Panda3D, Armoury 3D, and Construct, which are all used in very well known indie games. While including GDevelop which is a teaching tool for children not actually meant to create commercial software, and Dreams which isn't a game engine but a sandbox game for the PlayStation 4 for creating interactive fiction. And meanwhile Yahaha isn't even a game engine, it's Web3 software designed for Decentraland and Horizon Worlds, which are so dead the corpse doesn't even smell anymore.
AFAIK there's no public sdk for it, only a single third party game is in development and it's by the developers of garry's mod (one of the biggest third party source "1" games).
I'd consider it still proprietary, at that point you'd want to add Snowdrop (ubisoft), RAGE (Rockstar), REDengine (CDPR), Frostbite (EA), etc.
a16z doesn't know anything about video games, basically. A bunch of stuff widely used isn't on this map at all and instead there are buzzword startups on the map that will never be used to ship a video game to customers.
A "next gen" game engine would let me open it up and describe the type of game I was making and then plop me into a scene with all of the basics set up via non-desctructive, procedurally-generated content that I could edit and refactor in a real-time environment and then one-click-export to common platforms.
For example, I could type "a open world 3d game", or "a pong-like 2d game", or "a 2d platformer" or "an fps", and each of those prompts would generate a default scene, with whatever other detail I wanted to add to the prompt. Then I would be dropped into a beautiful landscape - to take the first example - which I could then use procedural tools to start building. A paintbrush lets me paint a city outline which is then procedural cut up into city streets and sprouts buildings. I click on one of the buildings and switch it's property from "realism" to "toon", with the 'on all related meshes' option checked, and the city flips from something out of GTA to something out of The Simpsons. I finish the design to my current liking (it can always be edited later, in real time), so I click the "add populous" button and it randomly generates some folk for me. I click on one of them and start giving him a backstory and quests. Etc, etc, etc. I have an idea, I get all of the boilerplate nonsense done for me, and then I start crafting. That's the dream.
All of that is entirely do-able right now, today, with zero AI necessary (yay, wave function collapse!). And yes, AI could make that first bit pretty easy, so I'm not against it alltogether. Just saying that its usefulness, after that, is far more in the "generate other assets based on the ones I've already created" category, rather than the "drop in a character and autoconfigure my physics" arena.
This strange idea that I would ever want an AI to be creative, instead of the very natural idea that I want AI specifically to do the tedious, complicated, error-prone bullshit that is inherent with how we interact with computers, in the most auditable, and precisely-editable way is just so exhausting. Let it go, guys; nobody wants AI generated creative work. We want AI to make creative work so easy that anyone can do it.
The thing is, "AI art" and the like aren't meant for creative people, they're meant for those who consider the act of creativity itself to be "tedious, complicated, error-prone bullshit." People who can't even comprehend it being fun and meaningful in and of itself, but who can only see the inefficiency of a human process which needs to be optimized and automated away so that a commercially viable product can be delivered to market as quickly and efficiently as possible. So that they can "be a writer" without knowing how to write, or "be an artist" without knowing how to draw. "Make a game" without being able to write a line of code. And most importantly at all, not have to pay writers, artists or coders a dime.
This is why AI, for all of its potential good (and I do believe that potential exists, no technology has an inherent moral or ethical axis) can only inevitably be a cancer on society, because those developing it, putting up the funding and pushing the research, want the ends without the means.
I agree with you, I also looked into existing games that provide some level and quest editor and except Bethesda games there is almost nothing. So today if I have an idea and want to experiment I either have to mod Skryim/Falout or use RPG Maker.
I tried to make something in Unity, it is such a waste of time, watching tutorials you see magic like "you import this model from X but notice it looks shit, that is because you need to go in this menus here and tick this checkbox here". Or I had the issue where I tried to make a character run, the C# code worked fine it moved the character around as I pressed the keys, but the run animation did not played even if I done all the steps correctly and the animation was 100% working, there was no fcking error,warning or hint on what is going wrong - just a waste of time for a solo person to try create a small thing (I do not want to sell any game so I am not a threat for the pro devs and I had not intention to put it on Steam either so please don't send me hate comments)
honestly, a non visual artist game dev might be intetested in generative ai for art.. this is not at all a weird thought. theres lots of creators of gameplay systems and cool.ideas who dont have the hands.to male beautiful creative stuff. it would help them unleash their creative game ideas without the burden of making visual art. i like it honestly.. 'make me a nice wooden crate', 'make a city in style x'. why not?
> 'make me a nice wooden crate', 'make a city in style x'. why not?
For the first part, you can literally just buy game assets like this, if not find them for free.
For the second part, game environments are designed. You can't just "make a city in style X" and drop a player model somewhere and expect it to be fun.
And overall, the more you trust the AI, the less of your own creative vision you have control over. Either the AI is creating everything, looking as generic and mediocre as all AI art does no actual sign of unique creative vision on your part, or else you have to put in effort, in which case you don't need the AI.
This reads like a TED talk full of hand-wavy marketing-friendly vagueness and ambiguity. May as well just say next-gen engines will be faster, better-looking, and cheaper to work with, leave it at that, and let history prove you uncannily prophetic.
Yeah, it's from Andreessen Horowitz. This stuff exists for software managers to point to when they make purchasing decisions in B2B. Doesn't serve any practical purpose.
It is weird seeing my hobbyist game dev and professional enterprise software world collide like this. Articles like this made sense when Facebook was willing to put up billions for VR services. But that was a bust, and now you have tons of consultants spinning their wheels. There is no reason any game studio or publisher needs to be told what game engine is good, I don't even think they care as long as they can make a good game. It's not a huge organization that you are trying to rally around some kind of decision.
And who is going to pay for this? Is a16z gonna fund a few game engine startups? The skills needed to build a game engine seem relatively niche and a lot of the talent has already been captured by the big players.
I suspect that the most felt impact of AI tooling will be an increase in shovelware generic games. I also suspect that curation will become even more important as developers try to shove as much AI crap as possible in every aspect of the game with little consideration for how fun or entertaining it'll make the overall experience.
There are game engine startups out there but I suspect veterans are going to stay far away from a16z money because they look clueless.
An industry vet I know recently complained to me specifically about a16z, unprompted, after seeing some of their PR like this. He recently raised 8 figures for his games startup.
AI assisted procedural tools will just open the floodgates for even more shovelware. Creating good games will still remain a challenging human endeavour.
as the article seems to imply roght at the start,it would be so interesting to marry game enignes with asset creation tools. those could use AI, or not, but i feel a big issue now which is hardly addressed is the plethora of tools used outside of the game editor of an engone like unreal, which makes creating games a mess of.tools, difficult workflows and a mess of export and import shit to sift through. i wish there was a one tool to make the entire game kind of thing. generative ai or just integration of blender and texture tools etc. into ue5 fully. running these tools.side by side requires such an extreme computer...
I honestly thought that’s what this was. I’m sure they will fund them especially since the a16z games seems to have sunk a lot of money in crypto junk so they need to do the “AI-pivot”. At least an engine could be a viable business
Engines being a viable business is kind of questionable in 2024 and on. There are lots of free or very inexpensive options already out there with incredible documentation, tooling and asset store components available, so any new startup has to compete with those on cost and capability simultaneously.
If your goal is to build a viable business instead of just scam your investors, you probably either have to try and compete with Roblox by creating an ecosystem where you siphon revenue off of your customers and their customers simultaneously, or sell actual product (sort of like how Epic bankrolls their stuff using revenue from their games) alongside tools.
What does a "new game engine" solve for anyone other than investors? Why not improve on existing technology by adding new features, improving workflow, or enabling users to target new platforms?
https://developer.valvesoftware.com/wiki/Source
Perhaps it doesn't qualify by some arbitrary "next generation" status, but Unreal does?