S. Eric Chen

https://lnkd.in/g9FrD3NM Dr. Deep Sen, lead immersive audio architect with Apple describe Apple Spatial Audio Format (ASAF) and a new codec called APAC that seem to be pointed at integration to the Vision Pro and Video that uses audio in new ways. This video describes the technology and Apple's strategy, what is interesting and different is that this isn't the Dolby "Cinema Room" model. The idea is the "Audio" must serve the "In the Game" user scenario instead of the "Theatre/Room" environment. If you wish to understand Apple's path forward with this technology this is an excellent talk.

5

Thank Dani Cherkassky Great piece on the architecture shift voice AI needs. The core thesis: cloud-centric voice systems fail in real-world conditions because they are too slow, too expensive to keep always-on, and deaf to spatial context. The solution is a hybrid architecture—fast, always-on edge processing (Spatial Hearing AI and Cognition AI) handling 80% of interactions locally, with cloud LLMs reserved for complex reasoning. But the Kahneman thesis runs deeper than the article suggests. System 1 and System 2 are not just useful metaphors—they reflect an actual biological sequence. Fast, intuitive processing (System 1) evolved first; slow, deliberate reasoning (System 2) was layered on top. The brain's architecture reflects this: spatial-acoustic processing happens in milliseconds at the brainstem and primary auditory cortex, long before language centers engage. Kardome is not just borrowing Kahneman as marketing. Their architecture is neurologically correct—it replicates the hierarchical, parallel, spatially-aware processing that makes human auditory cognition so robust. Here is the deeper point: hearing preceded language in human evolution, and spatial hearing preceded both. The mammalian auditory system evolved to detect predators and locate prey—survival functions that required sub-second latency and continuous environmental awareness. Language processing was layered on top of this infrastructure. Current voice AI inverts this evolutionary logic: it starts with language models and treats spatial audio as an afterthought. Kardome restores the natural hierarchy. 🙌 Danny Shapiro, Rory Sutherland

35

2 Comments

Introducing 🔥GenS🔥 (Generative Frame Sampler) — a plug-and-play module that greatly enhances long video understanding in existing LMMs (10+ gain for GPT-4o on LongVideoBench), by selecting fewer yet more informative frames. GenS builds on advanced long-context LMMs, reframing key frame sampling as a generative task. Unlike CLIP-based methods, it supports sophisticated multi-hop temporal reasoning and nuanced query interpretation, leading to more concise and informative frame selection. We introduce GenS-Video-150K, a large-scale synthetic dataset tailored for training frame sampling models, featuring fine-grained dense annotation of frame relevance scores. Project page: https://lnkd.in/gVq9Sw8c Paper: https://lnkd.in/gjkBEdh3

87

1 Comment

I’ve been thinking a lot about screens lately. Not resolution or specs, and not the usual portrait versus landscape debates, but something more fundamental. We’ve spent decades designing displays as rectangles because that’s what paper, cameras, and early computing handed us. Over time, that shape stopped feeling like a constraint and started feeling neutral. But when you spend your days around kiosks, POS systems, and purpose-built devices in the real world, you start to notice where that assumption breaks down. A few things pushed this into focus for me. Netflix experimenting with vertical content. Curved casino displays that feel oddly optimized for long engagement. Immersive venues like Cosm that rethink how attention works in a space. And the quiet reality that most of our software stacks and SDKs still assume every pixel deserves equal importance. That curiosity turned into a longer Substack post I just published: “We Don’t See in Portrait or Landscape.” It’s an exploration of human vision, attention, display technology, and what it might mean to design future kiosks and customer-facing devices around how people actually perceive and interact, not just around inherited screen shapes. It also digs into what that shift implies for software development, SDKs, and device management once displays stop behaving like flat documents. If you work on digital experiences that live outside a laptop or phone, especially in retail, hospitality, or dedicated device environments, this might resonate. https://lnkd.in/gPwqwXxt

22

2 Comments

After years of leading engineering teams and developing new-to-the-world products in optics, display, and instrumentation, I’ve learned that innovation thrives where collaboration lives. Magic Leap has shown how bringing together diverse expertise can turn complex challenges into real progress. This new direction of becoming an AR accelerator for technology partners is a powerful way to build on that experience. By combining our deep technical knowledge with the strengths of our partners, we can move the AR ecosystem forward faster and with greater impact. The future of augmented reality isn’t built in isolation. It’s built together. This article reflects that evolution and the shared vision driving it forward.

132

1 Comment

Matterport / CoStar Group continue R&D investment to help everyone re-imagine properties that matter with this strong follow-up to the original paper: "An Empty Room is All We Want: Automatic Defurnishing of Indoor Panoramas" https://lnkd.in/gW5jWBJU

35

1 Comment

Spatial training and remote collaboration in BILT is seamless on an Apple Vision Pro. Manipulating the same 3D model and interacting in real time is a game changer for increased productivity, comprehension and knowledge retention. It’s so fun to build incredible and impactful experiences!

5

One of the most underrated benefits of 3D printing is optionality. When design iterations are quick and low-risk, teams explore better solutions instead of settling early. This changes how problems are approached — from “What can we afford to try?” to “What’s the best possible outcome?” In many organisations, the value of 3D printing shows up not just in parts produced, but in decisions improved. #3DPrinting #Innovation #Engineering

2

This is a fantastic blog post by Massimiliano Pippi showing how we built a custom solution for managing a massive LlamaIndex monorepo with 650+ (!?) Python packages. 1. We replaced poetry with uv (thanks Astral + Charlie Marsh) 2. We built our own build system LlamaDev designed specifically to work more efficiently, have more reproducibility, and scale better vs. pants. Check out our blog: https://lnkd.in/gT8HAYRU

202

14 Comments

Companies have seen remarkable efficiency gains by leveraging Document3D for their work instructions. Reports indicate a substantial 90% reduction in time spent on updating instructions when transitioning to interactive 3D documents. Additionally, a notable 30% enhancement in training efficiency has been observed with the adoption of Document3D. Technical communication is a key component of digital transformation.

4

The world moved from filament to LED because the newer substrate was better. That wasn’t disruption for its own sake. It was progress. AI runtime will be no different.

4

4 Comments

Niantic Spatial SDK v3.15 has been released, bringing significant mixed reality capabilities to Quest 3 and Quest 3S headsets. This update provides developers with advanced features, including a centimeter-accurate outdoor Visual Positioning System

1

Innovation often requires operating at the absolute edge of technical possibility. Vicki Dobbs Beck explains how Vader Immortal utilized every available resource on the Quest hardware, to the point where a standard OS update rendered the app unlaunchable. This "Rebel Alliance" strategy of pushing boundaries is what defines the high quality immersive storytelling at Lucasfilm. Watch the full conversation for more of this analysis. Ted Schilowitz Rony Abovitz The AI/XR Podcast #XR #AI #Metaverse #GameDev #HardwareLimits #Storytelling #Lucasfilm

8

3 Comments