Geoffrey Stahl

Generative AI and VR/AR tech is evolving faster than we think. Watch this video and ask yourself: Are we ready for what the next 5 years will bring? Or are we still clinging to old creative methods while the world rushes past us? Here’s Figmin XR on Quest 3, building in passthrough mode outside our house. Combined with Gen 3 from Runway. Will you dive into the future, or do you still have one foot in the past? Or maybe a combination of the two, combining the old and the new.. will be key? Let me know what you think! #CreativeTech #VR #ai

124

11 Comments

In my generative AI work, I rely on prompts as little as possible, preferring the vastly more precise control that visual guidance materials offer. However, this new model by Black Forest Labs, called Flux, is something else. It has remarkable prompt adherence and interpretive ability, even with abstract, hard-to-describe concepts that other models tend to turn into a jumbled mess. These alchemical elemental glyphs I tinkered with are entirely prompt-based, and the clarity and cleanliness of their execution are pretty impressive. #ai #artificialintelligence #aiart #aicc #generativeai #genai #digitalart

15

1 Comment

https://lnkd.in/dTMYhRdj My quick take: 1. Always great to see a-list brands and story tellers taking AR seriously 2. Love Apple's attention to detail in creating for the medium in their early content and experiences on Vision Pro 3. A little scared that they'll get turned off if the numbers aren't what they want to see on this early content (both Apple and Marvel) - so many companies have walked this road before, only to get stuck in consumer AR chicken and egg fog 4. Anticipate a "pivot" to enterprise/industrial in the next 12 months and a period of disenchantment to follow... (unless they actually throw some mega bucks at interactive series, etc.)

3

1 Comment

Yesterday NVIDIA launched its first 'digital human' on-device Small Language Model (SLM), Nemotron-4 4B Instruct. The SLM debut is on Amazing Seasun Games' Mecha BREAK, a new multiplayer mech game. So there's a lot to unpack here. First and foremost, the announcement comes a month after NVIDIA launched James — its digital human brand ambassador (as seen in the accompanying video). James uses RAG to tell users (his voice provided by ElevenLabs) about the latest NVIDIA technologies (you can interact with James directly here: https://lnkd.in/eUBqcpDw). The hope for digital humans is that they can be applied across many roles such as customer service, entertainment, retail and hospitality. However, the deployment of digital humans into gaming marks a seismic shift in storytelling. By 'baking' language models into games, developers can create more immersive worlds for players to explore and enjoy. In this instance, Nemotron-4 4B Instruct improves the conversation abilities of in-game characters, allowing them to comprehend players and respond naturally more intuitively. Finally, the announcement is further evidence of GenAI moving towards the edge and becoming more task-specific — a big step in enhancing user experience, especially in scenarios like gaming, where latency and responsiveness are key. Follow ACQUAINTED for more.

10

Imagine having a robot that can interact with the real world and virtual worlds in the same manner. Game engines and AI perceptions are converging and becoming more and more compatible by the day. Many robots are already trained in games engines. If AI Agents kick in to the level that all the big tech players are aiming at, robotic embodiments is bound to happen for some of them. And since they will be virtual agents already, no reason why they could not have virtual avatars in games and virtual worlds as well (Metaverses and etc.) This is what Mixed Reality could look like. #MR

1

3 pillars of immersive VR experiences: 1) Intuitive interactions 2) Realistic physics 3) Responsive environment Nail these, and users forget they're in VR.

16

8 Comments

Here’s a quick test sketching a scene and doing a Magnific AI realistic style transfer so I can use Luma AI to add some motion to it.

119

19 Comments