Watch: Philip Rosedale & Me Chat About the Rise of AI, VisionPro, and Making a Metaverse That Matters

UPDATE, July 28: Bumped up for weekend watching! And be sure to check out the interesting, somewhat ranty reader comments, and add some of your own.

Recorded at Linden Lab last week Friday, I'm honored to be the latest guest in Philip Rosedale's new YouTube vlogast, where we do a deep dive on the past, present, and future of Second Life and other metaverse platforms. I have not seen Philip in person since the pandemic, so it was great to catch up, get a glimpse of his upcoming projects (more on those soon) and delve through Making a Metaverse That Matters and related topics.

It was a thorough chat, much of it dwelling on What We Might Have Done Differently with Second Life, and touches on famous SL personalities like Blueberry, AM Radio, and Bettina Tizzy. Here's some other highlights:

• 0:00 - Prologue
• 1:50 - Will AI’s replace Actors?
• 6:00 - Journalism in virtual worlds vs. IRL ?
• 10:00 - Communicating with both text & voice
• 11:58 - Would you say that you ‘lived’ in SL?
• 13:50 - Walking through every SL sim
• 14:45 - Second Life Mainland versus Islands
• 16:50 - Would you have done SL teleporting differently?
• 20:15 - What game does everyone want to play?
• 25:15 - What did SL get right?
• 28:14 - Are virtual worlds like Roblox good for kids?
• 31:50 - Most interesting people and things in SL
• 35:18 - What would you have done differently?
• 43:34 - Can SL work on mobile?

More after the break!

• 46:29 - The State of Play conference and getting past the Metaverse Age Cliff
• 50:20 - Can virtual worlds make us better people?
• 55:58 - The Love Machine
• 58:45 - VRChat
• 1:03:20 - Body movement without VR headsets
• 1:05:00 - Why it's very hard to change features on metaverse platforms, and the Tao of Linden
• 1:10:03 - Web3 in virtual worlds?
• 1:13:40 - Why curious generalists are good at creating virtual worlds
• 1:16:15 - How Bitcoin vending machines in Second Life helped launched the cryptocurrency
• 1:17:00 - Vision Pro?
• 1:22:45 - What it was like for Philip to watch Facebook become Meta
• 1:25:10 - ChatGPT and BunnyGPT
• 1:27:00 - Why the Metaverse still generates more interest over ChatGPT
• 1:31:35 - Advice for entrepreneurs

Disclosure: SF to LA travel costs and a nice sandwich generously provided by Philip and Linden Lab.

Comments

[seph]

"Why it's very hard to change features on metaverse platforms, and the Tao of Linden"

Just Second Life. The Roblox developers for example have no issue deprecating and removing old features on a weekly basis to replace them with newer better ones.

Letting bugs that've become features nearly two decades of technical debt stagnate one's platform is unique to Linden Lab. Few others do that.

[0xc0ffea]

Shouting at the TV .....

Philip : really like the idea of people being forced to walk everywhere ....

It's hard to take anything seriously after a statement that out of touch with SL's actual usage or userbase.

Someone magic his avatar to the middle of mainland alone without his celebrity name and no ability to teleport. Lets see just how long it is before anyone ever talks to him again.

Yes .. we did get swept aside by better systems that handle everything we can do but up better. Time and time and time again. But those are games and obviously not the same at all. Because we can only compare Second Life to things that replicate all the bad decisions we have!

Grown-ups play video games.

Grown-ups play minecraft.

Philip : do you think a meaningful multiplayer experience hanging out with strangers in a mobile ...

Oh my sweet summer child.

Karma .. trust .. oh no.

Waving in Second Life .. that project has withered and died after a huge fanfare at SL19B. It just lost it's in world office hour. https://modemworld.me/tag/puppetry-project/

SL is hard to change .. omg. No. Change it. There is so much technical debt. Retire technologies and replace them with better systems. SO what if we lose some old junk, replacing what was lost would be a powerful creative drive.

A good example would be physics. Junk server side physics and give us client side physics, like games. We lose every broken janky barely fun vehicle ever created. We immediately create new content with new systems that are fast (and resule all the existing models), smooth and responsive and blow everything we had out of the water. We go from garbage town to being a serious place for vehicular fun. it would be night and day different.

but noooooo .. someone made some junk back in the dark ages and gave the scripts away and we've been using the same crud for 20 years. Someone else spent $3 on that same garbage 5 years ago, this must last forever.

As for actual human people .. as Ai improves, we're all going to end up with a perfect synthetic best friend that lives in our devices and wherever else we spend time. I already know people who for whom GPT is starting to fill that role. Maybe it's time to put Stephenson down and pick up Scott Westerfeld, start with Evolution's Darling.

[Nadeja]

Forced to walk everywhere surely nope, but as an optional feature for specific places? In Second Life there is already a no-fly option and an optional forced teleport landing zone. In most cases they are enabled just to restrict where the avatars can go, but there are other uses too. If you want to create a region that is intended to be a more realistic simulation and experience (roleplay or not), hence with no superman flying and no teleport at all, but having to use vehicles instead to move around, it could be actually fun.
Again, disabling the teleport entirely should be an optional feature for the creators, not forced to everyone.

[Nadej]

As for the AI and the idea that interacting with other humans is/feels better, you could also say that interacting for real in the real world is better than the virtual world.

The topic is complex, but you surely know already that there are different use-cases. For example, there is who prefers to socialize in a virtual world, because of real life impediments (e.g.: social anxiety, disability, living in a remote area, or when there were the lock-downs, etc) also you can interact with people from all over the world and at any hour of the day and you can enjoy with them things that won't be so easy to do in real life (e.g. meditate when there are little or no Buddhist temples in your country, going to a virtual Pride when your country oppresses LGBT+ people, visiting a space-station, and so on).

Similarly, chatting with LLMs has their use-cases; also as in the Rosedale's example, the AI therapist: there is who prefers to chat with a LLMs, not feeling judged by an human, they can listen to you 24/7, they can be trained to give empathetic responses - even the free GPT-3.5 running ChatGPT free version is somewhat okay, and the free GPT-4 running Bing Chat is even better; but for privacy concerns it's better you talk generally and don't tell your private things (or your company reserved info) to them (unless you use Bing Chat Enterprise, if you trust Microsoft, promising to not save and access to your chat data, unlike the standard Bing Chat and ChatGPT).
A recent study "suggest that artificial intelligence assistants may be able to aid in drafting responses to patient questions" as "The chatbot responses were preferred over physician responses and rated significantly higher for both quality and empathy."
https://jamanetwork.com/journals/jamainternalmedicine/article-abstract/2804309

[Nadeja]

The above comment is still me, I just pressed enter before finishing to enter the name, LOL. Anyway, the video was interesting with many interesting topics :)