Here's How Making Realistic Mesh Humans SL's Main Default Avatar Type Has Hurt Second Life

Image by Susanne Drechsler

Since the flat avatars of Project Mesh Revolution have provoked quite a lot of buzz -- but also, a fair amount of anger from SLers who were expecting a standard mesh body -- here's an apropos excerpt from Making a Metaverse That Matters.

While many people in the virtual world industry buy the myth that ultra-realistic human avatars should be the end goal, my reporting about what they did to Second Life suggests otherwise. Here's the excerpt:

Myth: Photo-realistic human avatars and world graphics are the Metaverse’s end goal

This is a favorite myth among metaverse advocates who consult or work for companies producing 3D graphics cards, 3D development software, and 3D engines.

There is actually no proven relationship whatsoever between the popularity of a metaverse platform and photo-realistic graphics. Despite this, a wealth of metaverse startups and platforms frequently announce new plans to roll out ever more realistic environmental graphics, and ever more eerily human-like avatars.

The very most popular platforms, Minecraft and Roblox, are intentionally low-fi, immersive through their physics and responsiveness. Their whimsical avatars are similarly abstract. 

Why this is likely relates to their core user base: People in their teens and pre-teens, who are often still uncomfortable and unsure about their own real life identity and appearance. This seems even more acute for teen girls and young women, still negotiating the social expectations and judgements around their real life presentation; presenting them with a lifelike avatar to customize is effectively asking them to take even more social expectations and judgements. 

We also have something of an opposite proof point.

At launch, Second Life avatars were human by default but not realistic. The internal prim creation tools encouraged the construction of avatar attachments (robot helmets, furry tails, etc.), which led to a wide variety of avatar types and environments to explore.  

The arrival of mesh in Second Life in 2010 -- high resolution 3D files created in offline software and then uploaded into the virtual world -- greatly changed this dynamic: 

Nick Yee has already spoken in Chapter 9 about the community moderation issues that realistic human avatars engender: Preference them in your virtual world, and all the hidden and not so hidden prejudices of our offline world come along with them.

Thanks to mesh and other graphics enhancements, Second Life avatars and environments now look as detailed and as vivid as those from top AAA games. (For those lucky enough to own a powerful PC.) However, this rise in visual quality has contributed little to actual user growth. 

But by enabling ultra-realistic avatars, especially through mesh-based body attachments, mesh quickly altered the world’s culture. The Second Life web-based Marketplace accelerated this trend, since content creators now had incentive to create mesh avatar enhancements compatible with the most popular mesh bodies. 

Within years, the virtual world’s economy came to be dominated by ultra-realistic avatars; the overall creative culture changed, accompanied by the rise of environments most suited to them -- glamorous beachside homes and nightclubs, beautiful locales that resembled real life tourist destinations and locations for the latest reality TV show.

As Second Life’s economy snowballed around quality mesh items, so did its culture. While avatar fashion and virtual housekeeping were always a crucial part of the virtual world, the creative tools also attracted a cohort of creators and tinkerers more interested in using the platform as a multi-user game development space and all purpose sandbox space. 

By and large, however, tinkerers of this type faded in prominence within the larger community, overwhelmed as it was by new fashion releases and shopping extravaganzas. (They still exist, but are less prominent in the community.)  

Realistic human avatars also shape how the rest of the world takes shape:

“Once users are presented with a believably human template,” as researcher Nick Yee explains, "You want chairs and furniture and cars where your bodies can sit in and drive around, and you need large virtual closets to put all your virtual clothing and people are building these beautiful cantilevered houses by the beach side because that's what people do in the real world.”

While Linden Lab may have hoped Second Life users would define reality according to their wildest imagination, the realistic human avatars shaped how much of the world would evolve:

And with them came all the social problems typically associated with wealthy beach enclaves in the real world. 

“[That's] where the racism and the sexism comes from," as Yee puts it. “Because when the avatars are sufficiently human to make human assessments upon, our inherent human biases come clawing into the digital world… It's almost unavoidable, because once you have bodies that are anywhere near realistic, people feel the need to dress up their bodies, and to look cooler than the next person. And suddenly you have this whole economy based around selling bodies and hair and body parts.” 

To be clear, Second Life is overall an endlessly creative community that for the most part is highly positive and supportive. And the SL mesh community is incredibly creative and largely positive. At the same time, Linden Lab making attractive humans the default was (and is) a conduit for much player-to-player abuse.

I began noticing that as an embedded journalist very early on. Female avatars which were poorly customized often got roundly ridiculed by many in the user community; people who adjusted their avatar to look fat or disproportionate (because in a virtual world, why not) were jeered at and harassed. 

Worse still, avatars which weren’t white regularly received racist comments and trolling. I've written about a woman, white in real life, who customized her avatar to look like an attractive Black woman -- and was instantly hit with racist jibes, even by people she thought were her friends.

In the same way that Nick Yee's early Stanford studies suggest that users unconsciously bring the unwritten rules of eye contact and social distance with them in these virtual worlds, "a lot of racial norms, gender norms, and sexual harassment follows us in,” he tells me. “We shouldn't be surprised by it now.”  

Much of these insights, by the way, are featured in his seminal 2014 book, The Proteus Paradox: How Online Games and Virtual Worlds Change Us―And How They Don't. 

A metaverse platform is inherently a virtual world, but as Yee’s book title suggests, any assumption that a virtual world will offer a complete escape from our daily prejudices and cruelties is sorely shortsighted.

What do you think? Agree/disagree? 

Originally published on my Patreon.

Comments

[Adeon]

No metaverse platform that lets users write their own shaders for their models opt to go for photorealism by majority.

Only platforms that dictate what shaders will be used on your models and give you no choice in the matter funnel towards photorealism.

[Wagner James Au]

Yeah good point, in the book I contrast this with VRChat, which originally marketed itself with realistic human avatars, but let the community go wild with its own anime/furry/robot/etc. etc. shaders.

Guess which one has 10 million users and which one hasn't grown since the addition of mesh?

[Ghost Bird]

Agree. This drove me away from SL. The focus on photorealistic bodies turned the culture into a gross mirror of exactly what should be left in RL: misogyny, racism, classism, sexualisation and, ironically, materialism. Abstract, whimsical avatars and spaces obviously aren't free from bad behaviour, but I am convinced stylisation helps and photorealism hurts when it comes to socialising and community-building in virtual spaces.

[Lex4art]

My chaotic 2 cents on the subject:
1)Photorealism, if it's understood and implemented as "lets try to achieve real life look and feel", is quite boring. Grab a phone, took a random shot on your street, look on the resulting photo: it's absolutely real & highly detailed ... and boring.
2)Photorealism, if it's understood and implemented in virtual world as a standard, with path-tracing level of rendering quality (like in recent HL2 RTX demo), is something different: it provides awesome basis to create all kinds of stuff: from realistic to cartoon, from blocky/polygonal to silky smooth life-like - a freedom to choose style because it will look good, backed by path-tracer render and stuff. Personal art skills & limits will become a problem, but AI seems to be fitting to simply enhance creative input, making it's technically perfect & detailed. Personally I tried Suno AI to turn 40 minutes of my lyrics writing efforts into the song and results were interesting, here is one of them: https://suno.com/song/b0552249-817d-4c74-bf45-1d01db7add5f?sh=DSO7G38cO3AFGpQm

3)Top-notch graphics is only one of the pillars to support virtual world - not enough, rest of the pillars should be in place and also very well made: social rules for smooth interactions, economy, UI for building and doing stuff, gaming areas - and all that should be delivered to the each person in real time, with low cost and ping... some compromises are in order, but smart ones.
4)And on top of those pillars there is a need for core concept that gives a meaning to the people in this virtual world. Why to be there and do stuff? What's the point? Maybe earning for living is enough to keep things going, so there is a need to connect metaverse with real world, making it essential in real world business. Another great challenge on the path from virtual world to the first true metaverse...
5)SL's meshes were able to provide artists more freedom & quality tools to create - both realistic and non-realistic avatars. But overall quality of SL render is quite low - it's just conceptually very old, barely reaching visual quality of games in 2005-2006, so adding high-detailed meshes or PBR there is like adding high-detailed meshes into 2006 year game without improving rest of it; a patch. Better than nothing, though. I'm not sure that consequences of having meshes were the only reason for SL to turn into "beach luxury party condo" something, meshes just a tool to make stuff better - demand for better stuff was only in that luxury domain, for some reason. Maybe this is simply the very last bastion of SL economy & activity? Scary.

[Luther Weymann]

Before my current mesh avatar, and way back in early 2004 I looked like Ruth most of the time and today my Alt which was that early avatar still has a classic body and with tweaking over the years, except for the hands and nose, you'd never know it was classic. I was okay looking like Ruth and with classic bodies. My first SL wife was classic body and she was a beauty. From 2008 up until about 2016 I made and sold many thousands of great design, but all prim, residential and commercial buildings and by 2016 that business was mostly dead from mesh buildings having a more real life appearance. Today I'm wearing a mesh body that is 9 years old that you never see in SL anymore and its very low lag. There are a thousands of non-beach luxury humble people in SL just enjoying what they have without any desire to upgrade. If you hang with the elite of SL and the SL bloggers you might get the impression that SL is a place of cool kids and clicks of families, but its not, it still hundreds of thousands of single individuals with old SL prim buildings with old stuff bought in world and though Magic Boxes and Xstreet and those people are happy with what they have.

[Gan Starling]

As 18-year member Aplonis Ember in SL, the most stiking avatar in memory was from my earliest days. A female cyborg memaid with hydraulic cylinder actuated tail who moved swimmingly about in the attitude of a seahorse.

[WinterRose]

I am a bit curious as to the numbers when it comes to people who design their avatars and how close they adhere to their RL selves. Without any numbers or data to back it up, I'd imagine by and large, most people tend to emulate themselves racially, if not sexually. Being someone who does robots, drones, dolls and the like, I tend to do several versions of any one avatar so I can customize or blend in with the tastes of any companions or playmates. But I'm far from standard in that regard.

I'm also neurodivergent, so my first instinct is to mask a bit. But I'd be interested in the data as to how much people tend to emulate their actual races or genders or even species when it comes to games/environments like these. Is the trend more to emulate one's self, or to customize and try to be what one would want to be?

[Old SL Guy]

Man in my 18 years of SL I always frowned upon "weird" content, I thought getting everyone in line was the way to go, and too much weird was what was scaring away potential new users. But hell, I didn't expect the "normal side" of SL's appearance to mutate into such a toxic elitist thing. This is so messed up if true.

[Val]

As another 18 year older in SLer my summary of SL post mesh is 'Nowadays we are way prettier but much less diverse and way way less interesting". It's an interesting idea that when avatars weren't conventionally RL realistic and beautiful, many more people seemed to say 'Well, heck, I'd rather be a dragon or a mermaid' but given the choice they now say 'Heck, I'd rather be gorgeous'.