2025: The Year Imagination Becomes Reality 🌍✨
Android XR arrives, China flexes, Sora stumbles, Veo 2 steals the show, Gemini 2.0 pushes boundaries, and Physical AI rises — plus 3D world building, lost Mayan cities, and more.
Happy new year! I refuse to let someone else write this newsletter for me. But the fact is, I'm doing too darn much. In the interest of sharing stuff regularly with y'all through this forum, I'm trying a new format — a distilled list of my posts, and other cool things I'll curate for you semi-frequently. Let's give it a try and you tell me if you like it via the poll at the end of this edition.
2024 Wrapped 🎉
Here’s the obligatory 2024 by the numbers:
Speaking & Industry Deep Dives 🎤
I gave a keynote at Vimeo REFRAME in NYC and used Gemini to help me prep for it — best speaking coach ever (grab my prompt).
Soon after, I hopped over to Adobe MAX in Miami, interviewed their CTO & VP of Generative AI about Firefly Video, and checked out their latest research like Project Perfect Blend. More on that in a future edition.
Want me at your event? → team@metaversity.us
The Winners in Generative Video 🎬
After all that Sora hype (and hey, I bought into it too), the launch has been quite the whimper. Google’s Veo 2 has emerged as the new state-of-the-art.
Meanwhile, the Chinese duo of Kling and Hailuo AI are quietly becoming the powerhouses in the space, with Luma and Runway nipping at their heels with new offerings like Gen-3 Turbo and Ray 2. Meanwhile, the likes of Pika Labs are differentiating by going after the consumer market — templatizing complex visual effects.
Google Veo 2 thread w/ my community (thanks Sundar for the RT!)
Kling 1.6, Luma’s Ray 2, and Pika effects & ingredients
Beyond the Hype 🎯
Ben Affleck is incredibly well read up on AI
Here’s my take (and Newsweek featuring it)
As the US ban looms, TikTok rolls out GenAI tools for advertisers
The Battle for Mediating Your Reality 🗺️
Spatial Computing is effectively turning into a three way battle between Meta, Apple and Google. Meta of course has been carrying the space, but now Google has partnered with OEMs like Samsung to bring a new XR OS to the party.
Android XR is official – the “Gemini era” for VR headsets & AR glasses
Live Demo of Samsung’s Project Moohan (ft. Gemini, Maps, YTVR)
Press Mentions: Android Authority, Android Police
I managed to try some early AR glasses (no filming allowed, sadly). They felt shockingly intuitive — multimodal AI really is the missing link. And they legit looked like normal glasses! Of course, there’s a huge chasm between “looks cool” in a lab and a product you can ship to millions of people, but I can’t stop thinking about them. For now, my Snap Spectacles are keeping me happy.
The Road to AR Glasses Goes Through Passthrough VR Headsets 👀
For now, it seems we’ll get higher end “pass-through” AR experiences with the likes of a Vision Pro or Project Moohan before we get legit glasses. It was also cool to see my prior work with Google Maps Immersive View + ARCore Geospatial API playing a central role in Google’s XR strategy and Samsung’s answer to the Vision Pro.
But the real star of the show is undoubtedly Gemini 2.0 and it’s new class of fast multimodal models that gave me the closest thing to a JARVIS like experience. Here’s the full live demo video, along with simulated videos of the AR glasses experience that will give you a sense for what select press and creators got to try out in person:
The Signal Tower - Quick Hits ✨
🌆 3D World Models
From single images to fully interactive, dynamic 3D worlds
20 min deep dive video ft. World Labs, Genie 2, CAT4D (Twitter/X, YouTube video)
Generative AI is cool, but procedural 3D remains undefeated
AI Meet Reality - Research & Society 🌐
China’s SOTA AI model uses a fraction of the compute of US Labs
Google Gemini 2.0 Deep dive (Sonnet 3.5 Moment)
PhD student finds lost Mayan city in 2013 LiDAR dataset (?!)
Before/after imagery of LA fires (with analysis ready data)
The TED AI Show - Podcast 🎙️
We’re closing out the season with a great line up — season finale coming soon. Catch up on conversations with the CEO of Perplexity, Synthesia, and Civitai. CSO of Hugging Face, VP of NVIDIA Omniverse and Meta’s Llama. Director of AI at DARPA, and brilliant researchers like Hilke Schellmann and Anil Seth. Stay tuned for an exciting update on what’s next for me and TED!
That's it for this edition. Hit me up with your thoughts on this new format, and I'll see y'all in the next one.
Cheers,
Bilawal Sidhu
https://bilawal.ai