New World Notes

When Will We Get AI-Powered Virtual World NPCs? Seasoned MMO Game Designers Weigh the Possibilities & Perils

Screencap of generative agent simulation -- playable here

Last week I wrote about the game developer who incorporated ChatGPT into his NPC-to-player conversations in his virtual world; now comes an experiment with "generative agents" -- AI-powered NPCs, who talk with each other:

In the experiment, the researchers placed 25 generative agents within a virtual world resembling a sandbox video game, similar to The Sims. Each agent was assigned a unique background and participated in a two-day simulation. The study observed various remarkable behaviors:

One agent planned a party, informed several friends who then invited others, and collectively they coordinated the event.

Another agent decided to run for mayor, sparking organic discussions about their campaign and political stances among the community. Different agents held varying opinions on this candidate.

Some agents retained memories with human-like embellishments, occasionally imagining additional details or interpreting events with their own perspective.

Developed by AI researchers at Stanford (read their academic paper here), some enthusiasts are effusive about the implications, declaring, "Entire virtual worlds could be populated with these agents, producing emergent behaviors."

Definitely eventually, on some level -- but the practical applications are still pretty challenging:

"First," veteran game designer Damion Schubert tells me, "we're at the very infancy of this kind of technology. Both the technology and game designers' capacity to work with it and shape it are going to be very different even six months from now than it is now.

"Second, every game is a different design, and what you need NPCs to do in every game is very different. Chat GPT today could probably make NPCs that spewed out random dialogue and light worldbuilding facts, but it's a longer walk to, for example, have NPCs be at a Bioware Companion-level of intelligence. But there's a lot of room between those extremes!"

Damion, as it happens, was lead designer of Bioware's Star Wars: The Old Republic MMO.

I've been excited about GPT-powered NPCs for quite awhile, but thinking about actual multi-user virtual worlds I actually play and enjoy, not to mention single-player sandbox worlds like Skyrim, it's difficult to imagine autonomous agents wandering around on their own devices without it becoming a vexing traffic jam. Like, what happens when the town blacksmith who's supposed to upgrade your broadsword decides that he'd rather become a master of interpretive dance?

Damion again:

Since TikTok Can Turn Users Into Avatarized Versions of Themselves, Will the Proteus Effect Kick In?

, 

As a general rule, when people create a virtual world avatar for themselves, most tend to opt for a conventionally idealized version of their offline selves  -- younger, more symmetrical, in better shape, and so on. Basically, hotter.

TikTok, everyone's favorite social media app controlled by the Chinese Communist Party, has taken that to the next inevitable stage with Bold Glamour, a new machine language-driven augmented reality filter which basically avatarizes a person's real appearance in real time: 

The Bold Glamour filter, now used over 16 million times since its release last month, contours your cheekbone and jawline in a sharp but subtle line. It also highlights the tip of your nose, the area under your eyebrows, and the apples of your cheeks. In addition, it lifts your eyebrows, applies a shimmer to your eyelids, and gives you thick, long, black eyelashes. It has, as the name implies, a glamorous effect. The aesthetic itself is impressive. However, the really amazing thing is how well it functions. The filter doesn’t glitch when your face moves or if something like a waving hand interrupts the visual field, as filters usually do. 

There's obvious concerns that this is yet another social media app that will only worsen users' self-esteem, especially for vulnerable young women and girls.

At the same time, I'm also curious how Bold Glamour might relate to the Proteus Effect. That's the phenomena discovered by Stanford researchers who found that when people embody an attractive avatar in a virtual world, they themselves exhibit more confidence in real life:

Lensa Scraping Artists' Work Without Their Consent Into Its "AI" Images -- Even the Artists' Signatures!

The Lensa app has gone viral in recent weeks, with much excitement over its new AI-driven “magic avatars” feature.

Small problem: It's not exactly magic nor purely artificial intelligence. Instead, to create these avatars, the app is apparently scraping up artists' images without their consent. The image appropriation is so blatant in many cases, the Lensa-generated images even include the original artist's signature. (See example above.)

I learned this via professional illustrator Lauryn Ipusm, who's been posting examples of this on her Twitter in recent days.

 "I think they didn’t think artists would stand up for themselves because we don’t [have] industry labels the way the music industry does," Lauren tells me. She says that, because Stable Diffusion, the AI company which provides Lensa's neural network, is very careful about how the platform samples and trains from recorded music.

"The fact that they do so with their music model shows they are well aware of copyright (I mean it’s a basic concept, anyone who isn’t a little kid is aware of copyright), and that it’s not something that was too complicated to implement."

Lauryn's findings were noted yesterday by NBC News, but so far, Lensa's corporate owner has only offered an incomplete reply:

Artist Uses ChatGPT AI to Create Text-Based Adventure Game -- Then Feeds That Text Into Midjourney AI to Create Images for Game

Game industry artist Martin Nebelong recently started experimenting with ChatGPT, and discovered he was able to quickly create an entire "choose your adventure"-style adventure game through the AI text generator. Taking that a step further, he then took the ChatGPT text and fed it into Midjourney, the AI image generator, which then created illustrations to accompany the game.  (Watch above.)

"I was very surprised by how quickly the chat bot was in on the idea," Martin tells me. "To begin with, it wanted to write the whole story including all branches in one reply. But when I asked it to serve them to me as a story bit with a few branching options, to wait for my selection and only then to move on in the story, it got that an started over with the correct format. When it wanted to wrap up the story faster than I wanted, I just asked it to do a more detailed story and it did so.

"I guess the most impressive part to me was when I asked it to describe the characters in high detail for the prompts I wanted to use in Midjourney. It gave very detailed descriptions of all the characters including their clothing, their look, personality, names etc." (See an example below.)

But he was also impressed at how ChatGPT added descriptive flourishes he wasn't expecting:

Meta Employee Survey: 82% of Staffers Believe More Could Have Been Done to Prevent Current Layoffs

We saw Twitter employees' opinion of the company in the Elon Musk era last week, now here's a survey of Meta employees taken as the social media giant started laying off thousands of its own staff. Courtesy of Blind, the anonymous messaging app, this survey reflects the opinion of 1,179 verified Meta professionals in the U.S., and was taken between November 10 to 11, with a reported 4% margin of error.

The most striking answer is above: 82% of Meta staffers believe more could have been done last year to avoid the layoffs rolling out now. The way the question is phrased, this probably reflects at least some Meta employees blaming the company for spending many billions on making its metaverse platform. We saw a similar sentiment in anonymous Blind comments by Meta employees recently reported by Business Insider:

"The Metaverse will be our slow death," one user, who called themselves a senior software developer, posted on Wednesday. They added: "Mark Zuckerberg will single-handedly kill a company with the meta-verse." 

Zuckerberg apologized to staff for the need to cut 11,000 jobs, admitting that he "got this wrong".

In fairness, Meta's recent woes are more directly attributable to Apple's recent policy shifts which hurt advertising revenue collected via Facebook's iPhone, but Meta spending some $36 billion on metaverse-related technology is still a massive opportunity cost. 

Related to that, compare and contrast Meta employees' opinion of the company, versus Twitter employees:

BREAKING: 89% of Twitter Employees Believe Twitter Won't Succeed Under Elon Musk; 95% Say Recent Layoffs Not Done With Empathy -- Blind Survey

Just got some pretty amazing new data from Blind, the anonymous online community app for employees: In a survey of Twitter staff using the app conducted from November 7 to 10, a full 89% say they don't believe Twitter will succeed as the social network company's new CEO.  (See image above.) 

Based on incessant reports of major layoffs and resignations by top execs, I expected a large-ish percentage of Twitter's staff to have a negative outlook on Musk's leadership. But 9 in 10, nearly the entire company? This also suggest Musk's reported all-hands meetings and internal pep talks at Twitter have not helped morale -- nor has his general reputation as a tech genius capable of pulling off seemingly impossible tasks. (Historically, admired tech executives are able to build overall morale after a layoff.) 

Elon Musk's Twitter takeover has helped helped one online community, at least: The Twitter employees using Blind. 

"The increase in usage from Twitter employees has been incredible," Blind's Rick Chen tells me. "They now check Blind an average of four times a day compared to an average of two times a day six months ago (i.e., before news broke about Elon having bought a stake in the company). Twitter users are almost exclusively posting and reading content in the private company channel."

In fact, over three in four Twitter employees are now active Blind users. 

"We now have more than 6,100 users from Twitter on Blind, or more than 80% of all Twitter employees," says Chen. "About a quarter of our Twitter users signed up in the last six months (between March and October 2022)." [Update, November 14: As of last Friday, Chen tells me, "There are now more than 7,100 verified Twitter professionals on Blind or about 95% of the pre-layoff employee count!"]

This survey also asked Blind employees how fellow staffers were treated in the recent layoffs ordered by Musk:

Best Laptops For Second Life: Recent MacBook Air Model Runs SL Surprisingly Well (With Some Graphics Settings Tweaked)

Thanks to reader "dkronfeld" for sending this photo which effectively demonstrates a pretty surprising point he made in a recent open forum: With the right settings, Second Life runs pretty damn well on a recent model MacBook Air! Pictured above is his 2020 MacBook Air with an M1 processor, 16GB RAM, and the 8 core GPU. (This one here, with a starting price of $999.)

Emphasis on "right settings", though, as dkronfeld explains:

Best Laptops For Running Second Life: ASUS - ROG Zephyrus Recommended for Great Graphics & Power, MacBook Air for Lightweight Access

Above: Whiskey's laptop running SL  -- photo courtesy Whiskey Monday

If you're looking for a new PC gamer laptop that runs Second Life well, acclaimed metaverse artist Whiskey Monday recommends the ASUS - ROG Zephyrus 16":

"This is my new laptop and it's light and ridiculously powerful in Second Life," she writes in Comments. "I can run Second Life and Photoshop at the same time and while the fans are loud, it runs like a champ."

It's currently selling for under $2000 on both Amazon and Best Buy so definitely worth a look. As for the loud fans, that's a frequently reported side effect of the Firestorm viewer grunting and groaning to process all those high end graphics. (For what it's worth, my Dell G5 15 gamer laptop runs the official SL browser pretty well without heavy fan action, but then again, I generally don't run it on maximum graphic settings.)

Above: High-res image created on Whiskey's Zephyrus laptop (via her Flickr)

"Light" by the way, might be a relative term, because (as reader Mondy notes), the Zephyrus is just over 4 pounds. That plus the AC adapter may get you in 5-7 pound range. Which might not seem heavy, but if you're planning on taking it many places, I'd definitely recommend a good backpack that distributes that weight well.

Speaking of weight, longtime reader "dkronfeld" recommends (somewhat surprisingly) the MacBook Air, which weighs in at under 3 pounds. However, that comes with a graphics hit and settings to adjust:

Run Second Life on Steam Deck? This Redditor Says It's Possible Through Firestorm & Gets Good Performance

SL Redditor /RealEvilind recently posted this tantalizing image, showing Second Life running on Steam Deck, the handheld videogame console from Valve, and they report it gets pretty great performance:

This is Firestorm running on the Steam Deck pretty well. It pulls off 30 - 40 FPS in a normal crowded area, on high, with ambient occlusion on. It runs on Linux, and it's very easy to install. All I had to do is customize the controller, so that I could move and manipulate the camera. The device's trackpads can be set as mouse with click and right click.

We finally have a pretty cheap, truly portable solution for Second Life.

Maybe so! If you have a Steam Deck and the patience to set it up to run SL. As I'm not a Steam Deck owner myself, I'm talking with RealEvilind now, and will hopefully get a video to show their demo in action. 

As for the setup, here's their notes on that:

Study: Online Video Meetings Impair Creativity -- Opportunity for Virtual World-Based Conferencing?

The title of the study published in Nature is "Virtual communication curbs creative idea generation", but it's specifically referring to Zoom-style online video calls:

 In a laboratory study and a field experiment across five countries (in Europe, the Middle East and South Asia), we show that videoconferencing inhibits the production of creative ideas... Specifically, using eye-gaze and recall measures, as well as latent semantic analysis, we demonstrate that videoconferencing hampers idea generation because it focuses communicators on a screen, which prompts a narrower cognitive focus. 

This tracks with my personal experience. I do countless editorial/brainstorming calls on Zoom and Skype video, and to make real substantial creative progress, find myself putting a browser window in front of the video feed, so I don't get distracted by people looking at me (looking at them). Or only having my camera on for the Hello and Goodbye portions of the call.

This strikes me as an opportunity for virtual world/metaverse platform-based meetings where typically, there is no direct eye contact, and the visual cues are immersive 3D graphics: