Improbable is attacking the metaverse’s networking problem

Herman Narula, CEO of Improbable, is trying to solve the networking problem of the metaverse, the universe of virtual worlds that are all interconnected, like in novels such as Snow Crash and Ready Player One.

His London-based company started creating internet infrastructure and software that was built for multiplayer gaming and massive virtual worlds.

The company’s Project Morpheus software, which Narula says will be able to support 10,000 players in the same space. Currently, games like Fortnite and Call of Duty: Warzone can support 100 or 150 players, respectively, as they battle each other with automatic weapons. This capability was why Improbable was able to raised $500 million from SoftBank and more money from others as it readied its software.

Narula is aware of the sniper problem in the metaverse. Kim Libreri, chief technology officer at Epic Games, brought this up to me, explaining how most games separate players by location into grids. They provide responsive networking for actions that take place between characters in a grid, but not between grids. But a sniper sitting on a mountain perch has the ability to see far beyond one grid. That means that sniper has to be instantly networked with another player, who might be a target who is instantaneously trying to get away.

Sniper vision

Narula said he is trying to solve this problem. Improbable started off very ambitious trying to widen what is possible in a typical virtual world.

“You have all these limitations, the number of players you can have, the number of interactions they can have,” he said. “The history of multiplayer games has been the history of avoiding mistakes and not really being able to solve them because of existing technological approaches. So our mission was always to blow open the doors as possible, and make ever more meaning for ever more useful virtual worlds.”

The company evolved into a full-service provider. Improbable acquired developer Midwinter Entertainment in 2019, and to shore up its tech, Improbable acquired backend firm Zeuz in early 2020. Midwinter Entertainment created the game Scavengers to show off the ability to put thousands of players in a game.

Now, dozens of game companies are using the technology around the world. Narula said that a game like Fortnite might handle 10,000 operations per second. But Improbable designed Morpheus to handle more than 300 million operations per second and it uses machine learning to optimize rendering. With that kind of performance, Improbable has expanded outside gaming and it is being used by the British government to simulate large bots.

“I’m incredibly excited about this. It’s been 10 years in the making,” Narula said. “Last year, I think you caught some of the events with thousands of players that we were able to run. It can support about 10,000 people, all with high fidelity and interaction quality. Everyone can rush up together, interact together with physics, shoot and fight, and even be rendered on screen and have all that information.”

With the Kpop star AleXa’s concert, she was able to interact with thousands of fans at the same time in the same space, thanks to Morpheus.

“The more you desire live interaction between lots of things in a world where there’s a lot of information being exchanged, the harder it is,” he said.

Trade-offs?

Improbable also staged a huge battle with thousands of Scavengers fighting together in one space.

I asked if there was a trade-off between how many people can be in a simulation and how responsive or realistic that simulation can be.

“That trade-off exists at a theoretical level, but existing architectures are nowhere near hitting the boundaries of what’s possible,” Narula said. “So for example, with our platform, we’re able to support so many thousands of people and so many hundreds of millions of updates a second because we’ve architected from the ground all the way for a decade just solve those problems. We haven’t focused on building networking in the traditional way.”

“Sniper rifles are really hard in most games because I plan only for connecting to a certain number of other objects, as long as it’s aware of a bubble of reality,” he said. “If I just zoom in and zoom out a great distance, suddenly giving me information about new things really quickly, as soon as you think about it. In the real world, when you glance over a crowd or look around the world, you actually do have a limited amount of attention, your brain adjusts what things It looks at with high fidelity. What improbable was done with Morpheus is we’ve adopted the same approach.”

In terms of processing power, it is extremely cheap and very fast to switch attention with something like a sniper rifle, zooming in and out while looking at thousands of unique zombies. There are bottlenecks in the client and on the backend, so Improbable had to invest a new rendering engine that allows the creation of lots of characters on the screen in great detail while optimizing the bandwidth used.

Becoming profitable

Narula said that the company will actually be breakeven or profitable this year, which is a big milestone. He said the demand for the company’s services is growing.

Improbable polled 2,000 gamers and 800 game developers across the United Kingdom and the U.S. to explore what the future of the metaverse looks like.

The majority of respondents think it will take between one and five years for the metaverse to come to fruition, though older developers are slightly more sceptical (45% versus 38%).

“This is incredibly enlightening,” he said.

In that way, it feels like the the movement toward virtual worlds is inevitable, Narula said. Some people favor centralized or decentralized services, but most are looking forward to a metaverse.The demand for massive social experiences like concerts or big virtual worlds is fueling the growth of Morpheus. I asked Narula if gaming would lead the way to the metaverse, or something like music concerts. (He has been thinking about this as he has a book coming out on the metaverse).Narula thinks the metaverse will be rich not only with games but places where people are going to want to hang out, if only to have status or connections with like-minded people. The activities can be transactional, or they can be focused on value exchanges. These spaces will become conduits for new types of jobs, like social media creators.“I think gaming is one really important activity in the metaverse, but hardly the only one,” he said. “And I think they’ll probably be a lot of people who live and work in the metaverse over the next decade.”

Metaverse vision

Narula thinks the metaverse will be a huge extension of society.

“I think societies have always constructed other realities that are important to us, like sports where one country’s team beats another and that is like a polite way of fighting a war.

When there is an important match, it creates a real transfer of tangible value from one world into the other world.

“Getting lost in all the networking and rendering, we forget what we’re actually doing here,” he said. “We’re building an embodied other reality that people can transfer back and forth between. So when you think of it that way, I look at the metaverse as this massive democratization of these other worlds that previously only a few, like famous athletes, and everybody can gain value. Everyone can extend. The term I use is the multiversal style, the idea of yourself. Now you have more realities, more identities, more fun and more relationships, more things to explore, and to see this kind of expansion of society, this expansion of this opportunity space for everybody. That’s the promise of the metaverse.”

He said the magic will be in finding the experiences, finding the content, finding the tangible ways people can take the things they’ve bought on the blockchain and bring that together as a system for value.

He thinks that the existing web has led to the creation of a few centralized platforms. He thinks there is potential for a shake-up where the people, the masses of content creators who create the value, get to keep more of that for themselves.

“We can move to a world where the actual creators, the modders, the inventors, the communities of gamers — they own a lot more of the value of the games and platforms,” Narula said. “I think that’s going to make a much, much better world. It requires a change of how we do these things, where the power is, and I think all the decentralized platforms have failed us a little bit.”

But he doesn’t want to exchange one monopoly for another. Improbable will happily provide services to clients as they try to create these new entities on the internet. It may be more convenient to have centralization and it may be easier to build applications. But the possibility of decentralization will keep everybody more honest, he said. The winning organization will be the one that creates the ecosystem that attracts the most creators.

“I see industry alliances, and standard, stepping through creating interoperability, these are all great and high-minded concepts. But the basic problem is actually how do we let the creators own what they make,” he said. “That’s the main thing we need to solve through in the cooperation. I worry that maybe the industry is so used to dictating the plans, so used” to giving content. When the shift happens, the creators will take money from the bottom line of the centralized companies that take most of the profit right now, he said.

Narula said it will be hard to create enough content for the metaverse, and we’ll have to use AI and machine learning to do that, he said. Asked when people will start spending as much of their work and play time as they have in the physical world, then that is when the metaverse can start, and that will happen sooner rather than later. But to get to the most futuristic visions of the metaverse will take a long time, he said.

“Within five years, I think we will see major shifts happening,” he said.

As for connecting different worlds, he said it is more likely we will create systems that encourage content creation with interoperability built into them to a content creator can move from world to world to world. You would have to get both World of Warcraft and Fortnite to rewrite their software to make such things happen. That’s a lot of work.

I asked him what is the end game, or the metagame of the metaverse.

“We know our minds,” he said. “We know we can interface with them. So what’s the endgame of the metaverse? It’s a little scary to think about. So little bit fraught with science fiction comparisons. But it seems clear to me that you’re going to adopt a completely rational position, where we’re going to end up connecting our minds to other realities. So we better build other realities that we want to spend our time in.”He added, “I think of governance. How much of these future outcomes look like if it replacing one company with another company. We all have the same story, different boss. That’s not interesting for me. I’m thinking about user equity and democratic ownership.I want genuine enfranchisement. That’s very hard for a lot of the companies. That’s something I think we should we should find a way to do as an industry.”