4 ms·

Mark Zuckerberg Quietly Buries the Metaverse

16 points by galaxyLogic 4y ago

marstall 4y ago
zero support for this thesis in the article
- kwertyoowiyop 4y ago
  The article ended right where the first paragraph of justification for the thesis should have been.
brucethemoose2 4y ago
I wouldn't write it off so quickly. Content creation tools are a bottleneck of historical and existing "metaverses," which generative AI kinda revolutionizes. Now, I don't think Facebook could even make a 2009 Spore quality metaverse, but I think they are still going to try.
- pipeline_peak 4y ago
  2009 Spore didn’t even end up to be Spore lol. Make a 3 eyed monster with claws! Sorry Maxis, it isn’t 1999 anymore…wrapping it up in a Civ clone doesn’t cut it.
  - brucethemoose2 4y ago
    I was thinking of Spore Galactic Adventures specifically, which is a very early and oddly unknown crack at a "game maker game" kinda thing.
- wkat4242 4y ago
  Yeah exactly, I was just basically saying the same. Developing content is the biggest barrier right now. AI will revolutionize this.
  - galaxyLogic 4y ago
    Good point. In the early days of web there was some VRML or something like that. I tried it out but in the end it fizzled fast, maybe because it just takes too much code to represent virtual reality. But so how in particular do you think AI could create the metaverse?
    - brucethemoose2 4y ago
      Its going to (further) democratize content creation. Think of a novelist who could produce a whole interactive movie by themself, or an engineer who could roughly prototype a vehicle in minutes. But don't get me wrong... its going to be awful for anything real-world. And companies like Meta wants us mindlessly scrolling through feeds in VR, not making cool stuff ourselves.
      - galaxyLogic 4y ago
        I assume AI generated metaverse would have to consist of some kind of "markup language" describing the data-structures that make up the metaverse. But markup language is still about computation, describing the computation needed to render the metaverse. Then while the AI can easily generate syntactically correct markup, we can't easily specify what we want that output of that computation to be. Perhaps some entertaining random game-worlds? But is that really useful for anything? Some entertainment yes, but how will it actually accomplish something useful, is my question.
hindsightbias 4y ago
Whatever Meta makes, it will be better than reality for many millions. > Candy Crush had 255 million users in 2021 The bar is super low.
- 1letterunixname 4y ago
  Microsoft owns 7%. There will be more enterprise integrations. Already, virtual desktops and virtual monitors are killer features. The amount of cash already invested in gear to run AI, it's only a matter of building apps on top of it. Meta should be running an AWS for AI with graph integrations and serverless before AWS or OpenAI does.
1letterunixname 4y ago
This article is fact-free bullshit. Last Employee FYI, I can't quote or divulge anything, but the term was used.
- carabiner 4y ago
  > Last Employee FYI, I can't quote or divulge anything, but the term was used. What does this sentence mean?
hulitu 4y ago
> Mark Zuckerberg Quietly Buries the Metaverse in a mount of sugar. /s
wkat4242 4y ago
Why does it have to be either AI or VR? Both can be very complementary. Part of the problem of virtual reality is the high time sink of developing content. Autogenerating content can really help there. Zuckerberg himself had a demo about this at meta connect, where you'd say "show me a tree, no a bit bigger, and more like a palm tree" for example and the system would comply. This could be a real boon for content generators.

16 points by galaxyLogic 4y ago

marstall 4y ago

zero support for this thesis in the article

kwertyoowiyop 4y ago
The article ended right where the first paragraph of justification for the thesis should have been.

brucethemoose2 4y ago

I wouldn't write it off so quickly. Content creation tools are a bottleneck of historical and existing "metaverses," which generative AI kinda revolutionizes. Now, I don't think Facebook could even make a 2009 Spore quality metaverse, but I think they are still going to try.

pipeline_peak 4y ago
2009 Spore didn’t even end up to be Spore lol. Make a 3 eyed monster with claws! Sorry Maxis, it isn’t 1999 anymore…wrapping it up in a Civ clone doesn’t cut it.
- brucethemoose2 4y ago
  I was thinking of Spore Galactic Adventures specifically, which is a very early and oddly unknown crack at a "game maker game" kinda thing.
wkat4242 4y ago
Yeah exactly, I was just basically saying the same. Developing content is the biggest barrier right now. AI will revolutionize this.
- galaxyLogic 4y ago
  Good point. In the early days of web there was some VRML or something like that. I tried it out but in the end it fizzled fast, maybe because it just takes too much code to represent virtual reality. But so how in particular do you think AI could create the metaverse?
  - brucethemoose2 4y ago
    Its going to (further) democratize content creation. Think of a novelist who could produce a whole interactive movie by themself, or an engineer who could roughly prototype a vehicle in minutes. But don't get me wrong... its going to be awful for anything real-world. And companies like Meta wants us mindlessly scrolling through feeds in VR, not making cool stuff ourselves.
    - galaxyLogic 4y ago
      I assume AI generated metaverse would have to consist of some kind of "markup language" describing the data-structures that make up the metaverse. But markup language is still about computation, describing the computation needed to render the metaverse. Then while the AI can easily generate syntactically correct markup, we can't easily specify what we want that output of that computation to be. Perhaps some entertaining random game-worlds? But is that really useful for anything? Some entertainment yes, but how will it actually accomplish something useful, is my question.

hindsightbias 4y ago

Whatever Meta makes, it will be better than reality for many millions. > Candy Crush had 255 million users in 2021 The bar is super low.

1letterunixname 4y ago
Microsoft owns 7%. There will be more enterprise integrations. Already, virtual desktops and virtual monitors are killer features. The amount of cash already invested in gear to run AI, it's only a matter of building apps on top of it. Meta should be running an AWS for AI with graph integrations and serverless before AWS or OpenAI does.

1letterunixname 4y ago

This article is fact-free bullshit. Last Employee FYI, I can't quote or divulge anything, but the term was used.

carabiner 4y ago
> Last Employee FYI, I can't quote or divulge anything, but the term was used. What does this sentence mean?

hulitu 4y ago

> Mark Zuckerberg Quietly Buries the Metaverse in a mount of sugar. /s

wkat4242 4y ago

Why does it have to be either AI or VR? Both can be very complementary. Part of the problem of virtual reality is the high time sink of developing content. Autogenerating content can really help there. Zuckerberg himself had a demo about this at meta connect, where you'd say "show me a tree, no a bit bigger, and more like a palm tree" for example and the system would comply. This could be a real boon for content generators.