Spatial Computing

Thoughts on “GDC in VR”

The conference world has been thrown for a loop with the wide-reaching recommendations for COVID-19 discouraging or prohibiting large gatherings. As the games and VR industry raced towards March, the month where both GDC and SXSW were scheduled to overlap, cancellations have meant that almost everyone has had a change in plans

Naturally, VR Twitter began to weigh the pros and cons of “GDC in VR”, hoping to embrace the growing medium for a chance to recreate the experience of meeting with friends and colleagues in a space that was guaranteed to be virus-free. The perspectives were split

Surprised no #VR startup is launching a #GDC in VR.— Adam Draper 🛸 🌊 (@AdamDraper) February 26, 2020

After everyone has RTed someone talking about having GDC in VR people are taking steps to make it happen. I am on boaaaard— The Mulmolorian (@mulmbot) February 28, 2020

Everyone who is saying we should just do “GDC in VR” has clearly never attended an event in VR before.— Andrew Eiche (@buddingmonkey) February 27, 2020

Over the past three years, I’ve worked on social VR pretty much exclusively, starting at High Fidelity and now at Mozilla working on Hubs. As someone who has created a lot of social VR content and given numerous conference talks in  both VR and at physical events, I have a lot of opinions of my own on the topic

So, readers, let’s buckle up and get to it – in today’s post, I’m going to share my take on “GDC in VR”, specifically, the technical challenges of building out large-scale conferencing systems and the social considerations of replicating an event like GDC in social VR.

Building this system today would be difficult at best, impossible at worst

Right off the bat, let’s take a look at one of the ultimate challenges facing a large-scale conferencing system that could run in VR at the magnitude of GDC: technical feasibility. While many an impressive demo has been created to showcase thousands of humanoid agents running around a space, and techniques exist to cull out crowds at distances to reduce performance, the fact of the matter remains that there is no demonstrable proof that any platform has successfully existed to pull off VR-controlled crowds of over 29,000 people.

High Fidelity raised $79 million to try and do this, and ultimately shuttered their VR platform earlier this year. At max, some ~423 people joined the same domain during a stress test in 2018, which was an impressive feat requiring bringing down (and spinning up) servers constantly to support the load. Many avatars seemed to update below frame rate, which made me feel a bit like I was a character in a stop-motion film. The audio servers would drop off entirely at points. Our content failed to work, hitting loads that we simply couldn’t test for under simulated circumstances. The goal of 1000 people in a virtual space was never reached. (To be clear, this is absolutely nothing against the developers – High Fidelity had an excellent engineering group trying to tackle these challenges in an impossible time frame.)

But, technical challenges can generally be solved given enough computer power and engineering time – so let’s say that a team of the best rendering and network engineers took to the problem with an unlimited budget and managed to build out a technical demo that allowed the entirety of some 29,000 GDC attendees and speakers to share a virtual space. What would the design of that look like?

In practice, wide-scale virtual events are hard

Even if you had a completely robust, highly scalable platform with the world’s most intuitive UI and it magically works for every platform in the world with no bugs and no network issues, the logistics are still hard.

To be flaw-proof, such a system would need to integrate with existing applications – email, calendars, chat platforms. It would need to do so without draining out the battery of your phone, it would need to be resilient if a network dropped, your speakers would need to have a working knowledge of how to solve microphone and audio issues, and you’d have to have a team at the ready to support this through for all of your talks or virtual workshops. There are a ton of failure points here and this is not a reasonable expectation with technology today.

On the bright side, with virtual events, you’d save money on catering!

The internet is not socially ready for this kind of event

You know how we say that one dog year is equal to seven human years? That framework seems to work for the internet too, but in reverse. The internet is now almost 31 years old, which means that it’s a delightful 4.5 year old in Internet Years. You know what 4.5 year olds do?

They ignore social norms. They poke people, seek out boundaries, butt into conversations, ask a ton of questions when someone is trying to talk, and lack the cognitive development that is required to go out in public alone left without a guardian or rules.

Gross generalization aside, there’s still truth to the underlying point: the internet is full of subcultures that push boundaries and platforms that have been around for a decade are still trying to figure out how to reign in bad actors and illegal behavior (we could talk for a while about Section 230 and how that feeds into all this, but that’s for another time).

Let’s look at general GDC for a minute. You know that you’re in the United States of America, specifically, in San Francisco, so you’re under a specific jurisdiction with a set of rules that apply to you, and a framework as to who is going to enforce them, how they will be enforced, how they will be applied and judged, and what you can do to appeal your judgement. Finally, at the top level, there’s the rules that Moscone Center and GDC set themselves, which might be a loosely or strictly enforced code of conduct.

Now, let’s consider our virtual GDC-like conference with 29,000 people meeting. The jurisdiction of such an event is now up in the air – do you judge based on where you’re connecting from? Where the servers are? Where the company is based? There isn’t a physical body or identity to directly tie to a single virtual account that may be acting up, you have limited capacity to enforce simulated harm, there aren’t yet global laws that recognize virtual rights – the list goes on. We’ve already seen research demonstrating some of the vile things that people can do in shared social spaces, and while we can’t avoid progress for the sake of avoiding the internet “dark side”, we certainly haven’t been able to demonstrate a more socially-sound system at scale yet on this front.

Finally, let’s just be up front about this – not everyone enjoys what it feels like to be surrounded by 28,999 others in 700,000 square feet of space. Proper environment design could make this more manageable, but at the end of the day, there is a significant number of people who simply don’t like what physical conferences require from them.

Geez, Liv, you really hate social VR for conferences, don’t you?

Actually, on the contrary,I’m really heckin’ excited about social VR for conferences!The difference is that I’m not referring to jamming thousands or tens of thousands of digital bodies together in one room for the sake of replicating the sense of being in the auditorium during a keynote presentation. I’m excited for what comes after that.

I’m excited for when I can get into a room and watch a keynote with my closest friends with my cat on my lap. I’m excited for platforms that make it easy for people to pop into a private room for a follow-up chat, where they don’t get lost in navigation menus. I’m excited to move beyond PowerPoint and Google Slides and explore what next-generation talks look like when they can spawn out 3D objects into the audience.

I’m excited for times when the audience can collectively take notes in a shared room and save it as an artifact for the future to revisit. I’m excited to move from a speaker -> audience model to something more democratic and equal, where the default format becomes birds of a feather over a 45 minute lecture. I’m excited when speakers can create their own pop-up spaces to follow up with the most engaged members of the audience when their time slot is over, or to do away with time slots altogether.

One of the most memorable events that I have ever been to was an “unconference”, an event where 75 people working in tech got together and created their own agenda the morning of the conference, hosted their own sessions, and built a space that was open to being truly creative, vulnerable, and collaborative. Conferences today largely reproduce our least favorite parts of school – but VR can do so much more than that, and that’s what we should aim for when we talk about “GDC in VR”.

I want the conference where I can shift the entire environment to make a point. We did this to a small degree during some of the talks we hosted at High Fidelity and it was one of my favorite “future of conference” elements. I want the conference where people in my session can jump in or out of existence depending on what else they’re doing, without disrupting others. I want the conference where people can preview what I’m talking about and listen in before deciding to attend my session, or the conference where they can jump between conference talks unobtrusively.

But beyond all of that, I want the conferences that more people can attend. I want the conferences that I can attend from my phone when I have downtime on a trip on the train. I want a conference where I can spontaneously decide to give a spin-off talk from a session that I just watched. I want more conferences where attendance isn’t dependent on visas or financial capabilities; conferences that don’t require me to output 1.08 metric tons of CO2 to get there.