Interested in learning what's next for the gaming industry? Join gaming executives to discuss emerging parts of the industry this October at GamesBeat Summit Next. Register today.
In the First Trip demo of Otherside, Yuga and Improbable were able to bring 4,500 players together at once in a 3D world. What was remarkable about that was that the players enjoyed full physics effects for their characters, and they could speak with each other using 3D audio and hear all the players at once.
This demo is the kind of thing that makes some people think there is a future in nonfungible tokens (NFTs), which use the blockchain to authenticate digital items such as Bored Ape characters.
There are plenty of skeptics out there, and Herman Narula, CEO of Improbable, has facec them for years. Improbable’s big promises of massive worlds haven’t come true. The company tried acquiring its own game studio (Midwinter) and tried launching its own battle royale game, Scavengers, but the game failed and Improbable sold the game studio off to Behaviour Interactive.
Now Narula claims that the new version of the company’s operating system is ready and it’s powering Otherside, and people can just look at it to see that Improbable can now do what it promised.
Narula is also working on his own book about the metaverse, dubbed Virtual Society, about what he envisions the universe of virtual worlds will be like in the future. We’re expecting Narula to speak at our GamesBeat Summit Next 2022 event on October 25-26 in San Francisco.
I talked to Narula about this and listened to him talk about the video of the event as we watched it.
Here’s an edited transcript of our interview.
Herman Narula: Everything here has full physics. There are collisions in this environment. We made the collisions cushier so that people could crowd without blocking each other off, but everything has full physics. You can slide into the walls, bounce off things. Everyone can see everyone. There are 4,500 people here.
There is no rendering bottleneck. We invented our own rendering solution that can handle tens of thousands of customized characters. This is a production test of that system. There’s also no voice cutoff. You can hear voices from anyone on the server. We do sound mixing on the back end so that sound falls off or moves forward. The sniper problem is solved in this as well. You can zoom in and out. You can see a fidelity change as you do that. That’s all working as well.
GamesBeat: How do you make out the voices when you have a lot of people right next to you? How is it at determining which voices you hear?
Narula: Other systems have to only let you listen to people who are within, say, 10 feet. What we do, you can listen to absolutely everybody at once, all 15,000 people at max, but the system will give you realistic loudness based on how far away they are. The big thing is that many other people have achieved scale by compromising on something — physics, gameplay, or fidelity. They have to use the same operations per second budget to create a different experience. Because we can handle those billion messages per second, we don’t need to compromise.
GamesBeat: Is it a very different technology than what Midwinter used?
Narula: Yes, it’s a whole other generation. Our old technology, SpatialOS, could not roll at density. It was hard to use. It was robust. A couple of VR games have launched on it, but it left a lot to be desired. It could only handle about a million operations per second.
The new technology, Morpheus, we’ve been testing it all last year, and now it’s ready for prime time. It has modern benefits. It can handle scale, but it’s also a lot cheaper to run, dramatically cheaper. It also works well with streaming technologies. You see us using streaming to make it easy to jump into the world. It also has machine learning-based bandwidth optimization. It learns how people move and optimizes bandwidth after each time you play. It becomes more efficient based on that data.
GamesBeat: What’s behind making it cheaper to run?
Narula: Just dramatic efficiencies in the way the system is architected. It costs less than $100 an hour to support 15,000 to 20,000 people simultaneously in the same world. That’s so cheap that we can run it for free, basically, for events.
GamesBeat: Is there a recommended number of people to have in a space? Do you have a ceiling?
Narula: We doubled it last week when we moved to a billion operations. Our current system handles 15,000, but we feel pretty confident that we can move to 30,000 or 40,000 now that we’ve doubled the number of messages. There are bottlenecks that we’ve identified. I think 100,000 is probably the endgame of what this architecture can handle.
To go beyond that number of concurrent people in one spot would require more. But only the top three or four games in the world have that many CCUs. That’s in the entire game world, not just one spot. I think 30,000 to 40,000 should put us in a pretty good place to cover every single use case that we’ve actually seen.
GamesBeat: How sophisticated is each one of these characters? Is there a polygon count to sum that up?
Narula: The best way to think about it is there are no polygon limitations. The visuals are rendered on the client. The limitation is in how many things can change at one moment in time. When you’re in EVE Online and moving around in a spaceship, very little information is actually shared with the server. All that’s really changing is the position of that object. When you’re in this environment, again, there’s very little interactivity by design, so you can handle the scale.
Because we’ve blown out the budget limits of operations per second, we can have animation states that look different. We can have characters that are interacting in the far distance. We can have people who are speaking. I’ll show you an example that I like. This is an example of our rendering. People talk about the rendering bottleneck. This is more than 10,000 completely different characters being rendered at great distance with different animation states and different visualizations. This is because we use machine learning to learn how the characters are represented. We use a combination of that and a new hardware acceleration algorithm we wrote to be able to handle this at great distance.
This ran flawlessly last weekend with 4,500 people. It runs weekly in tests with large numbers of people that are more private. We use it for our calls. It’s basically good to go. We’re not developing technology right now. We’ve met many times over the years, and it’s always been, “Coming soon, coming soon, coming next month, coming next year.” I’m happy that we’re not in that position anymore. This is finished, completed production technology. What we’re working on now, other than just making it even bigger at scale – which is cool, but unnecessary – is how we can build open interoperability standards for objects. How do we create some of the backend services to make commerce and things easy to use for all of the worlds on the network?
GamesBeat: If you’re shooting for the metaverse, are you developing software here that can run on anything yet, on future generations? Or are you already focused on what can actually run?
Narula: That’s why the graphics are cool, but we’re being cautious on the poly count and things. Not because we’re limited by what M2 can do. You can see even higher fidelity demos. But we really want to make sure the experiences run on phones as clients. I take the opposite view from Facebook with VR. I want to reach the maximum number of people in the most casual context possible. I want to focus on crowds and large events where there’s so much value to so many people.
I think that if the metaverse just tries to build better video games, but with crypto, it will fail. As you have seen, it takes a long time to make fun games. But if we focus on experiences that are more differentiated, like we saw on Saturday, with actors, we’re in a blue ocean. We’re not competing with hardcore games. We’re competing with television and events. You can build hardcore games too. You saw us do the battle with a thousand AIs fighting, but I don’t think that will be the first use case. The use case will be social.
GamesBeat: Would it be relatively easy for somebody to do something where you team up with a group of 10 friends and go to a concert, something like that?
Narula: Absolutely. In fact, we ran a concert last year with a K-pop star called AleXa. Making that really easy is important. The new thing that you saw on Saturday was that you didn’t even need to download a client. We used a streaming solution we integrated quite well into M2. People can just click a link. You send someone a link on WhatsApp, they click it, and they’re in the concert. They’ll show up standing next to you, because of the link-sharing system that we’re developing.
GamesBeat: On the Bored Ape side, can you talk about getting into contact with them and how that worked?
Narula: Since we started demonstrating the technology last year at really big scale, we were surprised by the sheer volume of people that reached out. A lot of them were not from gaming. You know, Improbable built Fall Guys on the back end. We work with all these game studios and publishers. But none of them care about web3. They’re happy with our services, but they don’t want web3 and they don’t necessarily want these metaverse experiences yet. The people that want that are web3 companies like Yuga. They’re sporting leagues and fashion brands and music labels. They have a very different view of the kind of experiences that will be entertaining for their customers.
We were inundated by different people after the demos. Guy Oseary, who’s the talent agent for a lot of celebrities and sits on the board of Yuga Labs, he saw this and introduced us to the founders. I fell in love with the founders. I don’t know if you saw some of the tweets on the weekend with our free day and stuff, but they got in it. We were at a point where it took a few months to really dig deep together and understand we could collaborate. It’s a huge investment and risk from both sides. But we’re working on maybe the most high-profile project in all of web3. If we fail, it will look really bad. I’m glad Saturday went well, because I think we’ve convinced the community we can deliver something amazing.
Now they’ve offered to help a lot with the M2 network as well. Otherside is not only a stand-alone metaverse, but it’s also part of the network. Yuga is helping us with that. Later, when we launch a token around it, it’s going to be quite cool to have their support.
GamesBeat: What else is landing soon for you?
Narula: We’ll be announcing more partners soon. They’ll be pretty cool. All the partners that were lined up are non-gaming. They’re people in sports or music or other areas. Announcing experiences with those partners and how they relate to Otherside is going to be a lot of fun. Honestly, I didn’t expect things to go in this direction. I’m surprised by what has turned out to be — there are a lot of theoretical conversations about the metaverse. I was arguing with Matthew Ball on Twitter. He was saying, “Are large crowds entertaining or fun?” I was saying, “We have large crowds. We can see them. They’re fun. People are in them and enjoying themselves.”
When we were rehearsing the demo, we were thinking, “Maybe this is too silly. These jokes are kind of failing.” But when it’s in the context of thousands of people, the crowd loved it. They laughed it up. They thought it was amazing. There was so much Twitter chatter about it. I feel like this is a new format. When the Nintendo Wii happened and Nintendo went in that direction, a lot of people thought it would fail, that they were going for a market that still wouldn’t buy games. This is similar. The Web3 metaverse experience, a lot of hardcore game developers are going to turn up their noses at it. Would it be fun to have a battle royale game with 10,000 people? Probably not. But that’s not the point. It’s a different experience. It’s serving a different purpose for people.
GamesBeat: How is your progress on your own metaverse book? Are you still working on that?
Narula: It’s done. It’s launching in October.
GamesBeat: What direction would you say you went compared to some of the other books we’re seeing?
Narula: I wanted to go quite deep–it’s surprising because I’m probably one of the few people writing on this topic who’s actually building the technology. My book is not at all about the technology. I ended up writing about the history, philosophy and economics a lot more. What is it that makes a metaverse valuable?
One conclusion I came to is that the definition of metaverse a lot of people are promoting is very unhelpful. If you think the metaverse is a big collection of 3D worlds that are linked together, a lot of us gamers say, “We’ve had that for decades. What’s different? Why is this interesting?” But if you think of the metaverse instead as a network of meaningful objects — characters, events — that especially permeates real-world systems of meaning like music, culture, fashion and so on, suddenly it makes a lot more sense. This is a way of supercharging our culture. This is a way of giving people more ways of experiencing and connecting with the most important communities and people that they already value.
I can see a world where a famous sports star is hanging out with thousands of their fans, hyping them up before a game. Or where a musician has a really personal relationship with her fanbase, going from concert to concert, almost gamifying the experience of being together. These are things you can’t do today. Thinking about the metaverse only in terms of the game industry is very limiting.
I’d also say that a lot of game companies are rightly wary of the metaverse. Sharing value between worlds isn’t very helpful if you’re a game developer. Why would you want to connect World of Warcraft and Halo? That doesn’t sound like a good combination. Maybe it does to some people. But what makes a lot more sense is to connect Otherside and other web3 projects together. They have more of a relationship.
GamesBeat: For this particular one, is that a revenue generator for you?
Narula: Yes. Our revenue growth has significantly increased as a result of our metaverse work. I see a not too distant point in the future where Improbable is profitable with this growth.
GamesBeat: The deeper I got into talking about the real-time metaverse, the more discouraged I became. Does this make you more optimistic that we’re not that far away?
Narula: I don’t see technology barriers to achieving the stuff that —you won’t have to take my word for it. You can see it working on the screen. You can see people’s reactions to being there. You can see what they’re saying. These aren’t demos or paid testers or users that are part of some automated system. These are all real human beings, 4,500 real human beings who paid a lot of money to be there, and who are part of that experience. I don’t think there’s a technical barrier or limit. The barrier now is great content, really cool experiences that people get excited about. That’s the next mission: signing up and partnering with as many people as we can to create good experiences.
GamesBeat: Do you see more experiences like this, more big games or more concerts coming this year?
Narula: I think we’ll see a lot more events this year, even with Otherside, just on their road map, and probably some other surprises as well. I like events because you can do them fast. You can build them quickly, test them easily and run them efficiently. You don’t have to make a retentive, fun game. You just have to make an experience that’s compelling for a while. This can potentially be very lucrative as well, for all the participants. You can imagine the ticket prices for concerts. I think events will be a big focus.
More persistent game-like experiences are also in the works. They’ll come later. As you know, we raised a pretty big fund earlier this year for the M2 network. We’re quite interested in incentivizing content. But we’re not there yet. We want to focus right now on delivering Otherside and some of these other key things. If there’s one message in this conversation, it’s that we’ve done it. Here it is. It’s working with real people. Don’t take our word for it. Look at their reactions. This is maybe the first time we’ve spoken where I can confidently say that. It only took 10 years!
GamesBeat: It must be interesting for you to have conversations with people who were skeptical before.
Narula: It’s really weird. Since Saturday, now I get people emailing and texting and calling. They want to talk. They want to meet. They’ve seen it and they want to engage. Some of the stuff like voice, people didn’t think that was possible. It’s been great to prove everyone wrong, and to see physics at that scale as well.
GamesBeat's creed when covering the game industry is "where passion meets business." What does this mean? We want to tell you how the news matters to you -- not just as a decision-maker at a game studio, but also as a fan of games. Whether you read our articles, listen to our podcasts, or watch our videos, GamesBeat will help you learn about the industry and enjoy engaging with it. Discover our Briefings.