Welcome back to Spatial Beats, where we round up all the top news and happenings from around the spatial computing spectrum, including its escalating infusions with AI and other emerging tech. Let’s dive in…
The Lede
RunwayML Tees Up a $450 Million Investment. According to The Information, General Atlantic will lead the new round, valuing the company at $4 billion. The New York company has raised $237 million to date. Investors include Google, Nvidia, Salesforce Ventures, Felicis Ventures, Amplify Partners, Lux Capital Management, Coatue Management and Madrona Venture Group.
One of the most impressive things about Runway Gen 3 is how well it understands and interprets prompts. It almost always comes back exactly the way I imagined it.
"A miniature baby zebra walking on a fingertip' pic.twitter.com/DRqBv98g97
— Jonas Peterson (@jonaspeterson) July 2, 2024
Gen-3 Alpha can create videos up to ten seconds. And they’re good. Really good. But. Generating AI images can be expensive. Runway Gen-3 is $1/per each 10-second video generated. If you take image generation seriously, and you’re going to make multiple projects, this will get expensive fast, because for each usable shot there can be dozens of “takes.” Director Paul Trillo told me it took over 700 generations to make the Sora music video, Washed Out. 700 x $6 = $4,200. Even though it would have cost millions to bring Trillo’s vision to life without AI, I would not call $4,200 “democratizing the medium.” Indeed, Luma AI charges $400 for just 30 Dream Machine AI generations.
mini puppies with gen-3 pic.twitter.com/HO1WIx2YtB
— Proper (@ProperPrompter) July 2, 2024
“Gen-3 is cool. I’ve been testing it quite a bit, it seems to really excel at real-life simulations and establishing shots,” says Caleb Ward, CEO and co-founder of Curious Refuge, best known for “Wes Anderson Star Wars,” and other YouTube AI sensations. “The true utility power of Gen-3 will be here once you can upload reference images. Without that feature it is very challenging to get your desired end result and, as a result, you end up using a lot of generative credits that can get costly.” Eric Schwartz of TechRadar says “This move suggests Runway is ready to professionalize its products after having the chance to refine them, thanks to all of the people playing with the free models.”
Every time I touch a new #GenAI tool I use the same initial test prompt: "A spaceship landing in the jungle." This is what @runwayml Gen-3 did on the first try. Mind. Officially. Blown. @c_valenzuelab @iamneubert pic.twitter.com/H1WIsmGQFZ
— Charlie Fink (@CharlieFink) July 1, 2024
Early access user The Dor Brothers demonstrated Gen-3’s capabilities with the AI fashion runway show above. They launched a Discord Server to teach others how to create their own imaginative fashion show editions.
"Rogue Runway: Intergalactic Edition"
Made with @runwayml Gen-3 🤯 🔥
Feel free to join our 'Rogue Runway' Discord to get the prompts, make suggestions and create your own editions: https://t.co/jN45L4Wp4S#runway #gen3 #aiart #aigenerated #aifilm #aimovie #aifashion pic.twitter.com/kBDZn6cJXq
— The Dor Brothers (@thedorbrothers) July 1, 2024
When it launched in February, Sora appeared to be ay ahead of the rest of the AI video industry. With the release of Runway Gen-3 Alpha, Dream Machine from Luma AI, Vudu from Google, Kling from China, and more to come from Pika and others, the text-to-video category is crowded with companies and venture dollars. Popcorn, please!
🎥🎞️🎧 Short Movie: Skydiving
My first short movie using @runwayml Gen-3 pic.twitter.com/L2ZVGHTwit
— Abel Art (@A_B_E_L_A_R_T) June 29, 2024
Feeling Spatial
Meta’s New AI System Can Prompt 3D Assets. Three years ago, Meta teased “Builder Bot,” a voice-to-3D world builder. “Builder Bot, make me a beach scene,” asked a cartoony Zuck avatar in a demo video. Boom. Builder Bot immediately produced a whole scene, replete with ocean, sand, beach chairs, towels, and a palm tree. Now, with Meta’s new AI system, 3D Gen, that will be a reality. 3D Gen combines AssetGen and TextureGen to rapidly create high-quality 3D assets with realistic textures from text descriptions. The implications of 3D Gen are far-reaching, both inside and outside of the Metaverse.
NASCAR launched a new Chicago Street Course map. It’s styled on the real NASCAR Chicago Street Race taking place this Sunday, July 7th. On the Chicago Street Map in Fortnite, players navigate multiple routes, master the art of speed, drift and avoid obstacles in multiplayer races or solo time trials. NASCAR’s official launch on Fortnite brings its second annual event in the center of one of the world’s greatest cities, bringing a digital twin of one of the world’s greatest, most iconic citites into the Fortnite Metaverse.
The 9th Annual VRTO Spatial Media World Conference & Expo Opens Weds, July 10th. OCAD University in the heart of Toronto’s media arts district is the venue this year. Notable speakers is cofounder & Chief Creative Officer of The VOID, Author & Magician, Curtis Hickman, Dave Cardwell, Head of 3D & Immersive Labs at Adobe, author Caitlin Krause, and Ryan Andal, CEO and co-founder of Secret Location. Attendees include C-suite executives, investors, hackers, gamers, developers, artists, and leading education, healthcare, and enterprise experts. I’ve always found the show to be a super friendly, laid-back, welcome summer interlude in one of the World’s Greatest Cities.
- Quest 3S: Here’s what we know about Meta’s next VR headset
- Meta plans to bring generative AI to metaverse games
- The Vision Pro will get Apple Intelligence and “Go Deeper” in-store demos
- Hands-on: Sony’s New MR Headset Impresses, But Still Needs Tuning
- ‘Half-Life: Alyx’ on PSVR 2 Would be a Win-win-win for Valve, Sony, & Players
- Vision Pro Could Be Getting Controllers Soon from This Third-party Creator
- Tim Sweeney Explains Why Fortnite Has “No Plans” For VR Support
- Deleted Meta Comment Speaks Volumes About the Fate of ‘GTA: San Andreas’ for Quest
- Augmented Reality Laptop Innovators Terminate Staff and Hardware Division – XR Today
- Zuckerberg: Meta “almost ready” to Show Off Prototype AR Glasses
- Apple Reportedly Asks LG & Samsung About Supplying Micro-OLED For Cheaper Vision Headset
- Meta is Pulling the Plug on Quest 1 Security Patches Next Month
The AI Desk
Open AI Licenses Time Library. While it fights lawsuits claiming fair use of every website it scraped, OpenAI is trying to go legit. This week they announced a deal for Time’s hundred years of content. They already have licensing deals with News Corp, Axel Springer, The Financial Times, Reddit, and the Associated Press.
Eleven Labs to Bring Dead Celeb Voices to Its Reader App. AI audio firm ElevenLabs has licensed the voices of Judy Garland, James Dean, and other bygone stars to use their voices to read books, articles, PDFs, and other text material to mobile users of its new Reader App.
Weekend Reading
Every Major Enterprise XR Announcement At AWE (Rory Geener/XR Today)
Fortnite’s Metallica concert showed how sprawling the game has become (Andrew Webster/The Verge)
Tim Sweeney Explains Why Fortnite Has “No Plans” For VR Support (Henry Stockdale/UploadVR)
Listen In
For more spatial commentary & insights, check out This Week in XR, hosted by author Charlie Fink, Ted Schilowitz, former studio executive and co-founder of Red Camera, and Rony Abovitz, founder of Magic Leap. No show on July 5th but on July 12th we’ll be back to our regular weekly schedule with Chandra Devam, co-founder and CEO, Aris MD. Meanwhile, see our latest episode below, recorded from the show floor at AWE USA 2024. You can find subscribe on podcasting platforms Spotify, iTunes, and YouTube.
Charlie Fink is an author and futurist focused on spatial computing. See his books here. Spatial Beats contains insights and inputs from Fink’s collaborators including Paramount Pictures futurist Ted Shilowitz.