Paolo Tosolini

Seeing a lot of creators on X especially from the Western side constantly yapping about “Seedance nerf, Seedance nerf.” Bro… what nerf? Everything works perfectly fine on my side. Seedance 2.0 is still one of the best video models right now. At this point it just feels like a black campaign. If you want to compete with China’s AI progress, build better models. Not fake narratives while hyping tools that break the moment you actually use them. #seedance2

7

6 Comments

NVIDIA's Physical AI Data Factory announced today! NVIDIA + Nebius + Voxel51 team up to build a Synthetic Data Generation pipeline for AV and Robotics that actually works! Porsche AG is already using it to massively accelerate autonomous model development. Mario Viviani Evan Helda Daniel Gural Trisha Tripathi Itai H Zadok Mohamed H.

53

3 Comments

This pioneering project, delivered in collaboration with Bahri | البحري Chemicals & PropVR, marked the first-ever use of Unreal Engine in the shipping industry to create a fully immersive digital twin of an operational ocean-going vessel. Designed to transform maritime training and stakeholder engagement, the experience allows users to explore the Fatimah interactively. Bringing the ship to life through animated, functional simulations including lifeboats, cranes, fire safety systems, and gangway operations. The PropVR team spent two intensive days onboard in Khor Fakkan capturing scans, audio, and fine details, with special focus on the three-level engine room, one of the vessel’s most complex spaces. Real-time user testing and rapid iteration ensured a polished, high-impact result. The outcome is an award-referenced, conference-showcased experience that sets a new benchmark for digital twins in the shipping industry. We look forward to extending this innovation across the fleet. Thank you to Bahri | البحري Chemicals team especially Faisal Al-Husseini, Capt. Aleksejs Glitens, Paweł Cegielski, Capt. Pavels Fedosejevs, Khalid Alotaibi, Tariel ZAKARADZE, Hendrik Jan Atsma, Germaine Porsuelo the opportunity to PropVR collaborate on this exceptional project. Sunder Jagannathan Srinath Kandala Vivek Agarwal Karthick Musuvathi Eswarlal #UnrealUnrealUnreal Engine #Digitaltwin #Maritime #Marine #shipping PropVR

132

10 Comments

Workflows are still going to share a lot of common steps for control, no matter how much they change. Human control matters so much for proper storytelling.

9

For the last four years, our innovation team at Moment Factory has been researching generative AI. Not because it’s trendy. Because we’re curious. Curious about synthetic intelligences producing form. About where they fail. About where they expand visual language. We spoke with ComfyUI about one of those explorations: pushing diffusion models into architectural-scale 3D projection mapping. Working in real architectural space is a stress test for generative systems. At small scale, ambiguity passes. On large, irregular architectural surfaces, it doesn’t. Seams don’t disappear. Depth errors don’t hide. UV logic doesn’t negotiate. Super-resolution amplifies everything. 18K in our case. At that scale, spatial coherence isn’t aesthetic preference. It’s structural obligation. In location-based media, spatial conditioning isn’t optional. It’s the starting point. Generation has to be constrained by geometry from the beginning, not corrected afterward. What made it viable wasn’t better prompting. It was system design: • conditioning as structure through ControlNet and related guidance techniques • cascading models instead of chasing a single “magic” one • integrating production constraints early • human-in-the-loop as authorship, not supervision In the end, the question isn’t what the model can generate. It’s whether the system we build around it can sustain intent. So we keep researching. We keep testing. And yes, it’s genuinely fun. If you’re working on generative pipelines for spatial or location-based experiences, let’s talk. Curious to hear how others are approaching it. https://lnkd.in/eE3ySTnv

87

8 Comments

Caught this fun moment back at AWE 2022 with Proto Hologram. They are bringing holographic communication to life in the most iconic way. Imagine being in two places at once, presenting or pitching in full 3D, all while sipping your matcha miles away ☕ Ori Inbar presented his AWE keynote with Proto last year in a super creative way by having dialogue with his digital twin on stage! Anytime you need the exact likeness of someone presented in multiple places at once- holographic technologies, like this, are a perfect solution. Now, pair this with Hololight’s cloud pixel streaming and you’ve got an incredible product that should be a no-brainer for artists or high profile figures everywhere. As someone who lives at the intersection of immersive tech, AI, and innovation, seeing these kinds of real-time applications reminds me how fast the future is arriving. Shout out to the creators redefining presence, connection, and communication. I can’t wait to find project to use this for! #ImmersiveTech #WomenInTech #SpatialComputing #AWE #TechInnovation

51

4 Comments

Hot: Procedural splats are strong primitives for prompt-driven procedural authoring (now available in Arrival.Space) They abstract away many complex patterns needed for fire, volume fog, emissive particles and ground cover. You shape probability, not polygons. Generate. Tweak. Re-prompt - Directly in the browser. Scene link in comments. 🔥

64

16 Comments

Many studios and VFX shops get frustrated because AI generated shots are limited to 8 bit color depth out of the gate. These shots look fine on instagram but break down in the color grading process in Resolve and can derail production timelines for high-value projects. Topaz Labs develops deep learning models that help with this problem, notably SDR - HDR conversion. Working with a customer to deploy this model directly to their pipelines, let me know if you're interested in API access. We will get it up our API docs soon.

67

31 Comments

I am thrilled to share that OurWorlds, Inc. has been awarded U.S. Patent No. 12,394,161: XR360º Extended Reality System and Tool Suite. This describes our method of compositing elements in a 360º environment on mobile and desktop devices. This technology makes it possible to create immersive 360º environments with holograms and AR overlays—directly from your smartphone or in a desktop webapp. These experiences can be anchored by geolocation or unlocked at specific times, making it possible to connect place, time, and story in powerful new ways. We will be releasing it publicly this fall. Congratulations to my co-founders, Kilma S. Lattin and Maximilian Mackh. What started as a conversation over a campfire at Santa Ysabel Reservation in 2017 became the spark for reimagining what the education of the future could be like. How do we preserve the experience of being in community with our elders? How does charisma, and the timbre of a voice affect how we parse knowledge? We are hardwired over millions of years of evolution to read nuance—the expressions, gestures, and cadence matter. In an age when we question the authenticity of every image and video, how can XR and AI together create worlds that are not only transformative, but authentic and profoundly human? Making immersive reality workflows accessible is the key. Our goal has never been to gatekeep. We sought this patent to ensure these technologies remain accessible, stay in the right hands and are used mindfully—to revitalize language, to deepen understanding of our environments, and to support communities who want to share knowledge on their own terms. Please reach out if you’d like to explore how this can fit into your workflows. I came into this project as a developer, a designer, working alongside community to pursue new answers to these questions. The journey hasn’t always been easy. Innovation, learning, and collaboration rarely are. But every challenge has made the work stronger. To everyone who has supported the work—thank you. To those who have challenged it—thank you for teaching us to be better. With deep gratitude to our supporters, partners and funders in education and technology—Agog, NewSchools, Apple, World Design Capital San Diego Tijuana 2024, UC San Diego, University of San Diego, The City of San Diego, San Diego Public Library, SXSW, SXSW EDU, ASU+GSV Summit, Media Arts Center San Diego, Creative Visions, J. Paul Getty Trust, Birch Aquarium at Scripps, e3 Civic High, Codefy... Let’s continue this work together. It’s more important now than ever. #XR360 #AR #EdTech #XR #AI #LanguageRevitalization #Geolocation

75

20 Comments

Vision-3D v1.3 is now live. This update focuses on performance, usability, and better scene presentation. Key highlights: • Interactive Level of Detail for smooth navigation with large point clouds • New Scene States to quickly switch between different scene configurations • Enhanced viewpoints with focal points, animation control, and fixed perspectives • Improved camera controls (field of view + clipping planes) • Simplified interface with grouped tools and tags • New audio and URL tags for added scene context • Lock feature to prevent accidental movement of point clouds Scene States are a major step forward—allowing multiple scenarios (e.g., vehicle positions, crash stages) to be presented within a single scene without duplication. These carry into the viewer, making them ideal for presentations and courtroom use. This release is about working faster, cleaner, and communicating scenes more effectively. #Forensics #3D #PointCloud #AccidentReconstruction #Visualization #Vision3D Watch the video below for updates. https://lnkd.in/eJNT5fQq

37

In this guide, I break down exactly how GarageBand’s built in effects work, how third party AUv3 plugins fit into your workflow, and the difference between AUv3 instruments and AUv3 effects in GarageBand for iPhone. 👉 https://lnkd.in/eEHwStc5

4

1 Comment

Luma just updated their “Modify Video” feature that allows you to act out any scene! Head to Luma AI: https://lnkd.in/e7Z8xBbP Modify Video allows you to record a live-action performance and then apply AI style transfer on top, replacing backgrounds, characters, and entire visual worlds while keeping your original motion. This is a huge leap for AI filmmaking, visual effects, and creative storytelling without complex VFX pipelines. #LumaLabs #ModifyVideo #AIVideoEditing #AIStyleTransfer #creativeai

261

13 Comments

“Every Person Matters” is our first principle for a very important reason, and it may not be what you think… Ethics may be based in religion or morality or even social responsibility. But not everyone has those naturally, or even shares the same understandings if they do. Focusing only on one religion, family, team, company, class, or even country can cause related harm to those outside these circles of consideration. If someone instead focuses on effecting the greatest good and doing the least harm with their work, they can also be quite ethical. Even a high-functioning sociopath can act ethically on that basis, if they simply understand how it benefits them as well as everyone else to do good and avoid harm. However, such “utilitarianism” (as philosophers will explain) has also been used to justify atrocities in the name of some “Greater Good.” Eugenics, Nazism and Effective Altruism will cite some imagined “Greater Good” benefitting the group in power, while actively discounting the harms (like exclusion, sterilization, poverty, or even death) to those who have less power. So it’s imperative that any consideration of ethics takes everyone into account. Every person matters. No exclusions of affected parties. Are there any ethical schools of thought that are free of the trap of externalities or otherwise excluding some people to benefit others?

21

3 Comments

I recently got early access to Luma and spent some time experimenting with it. Naturally, I decided to make a small micro-drama pilot about Luma itself, just to see what the workflow actually felt like. One thing I liked straight away is that you can approach it in two different ways. You can be very hands-on, working directly in the canvas. Prompting, adjusting shots, switching between different video models, and shaping the scene as you go. Or you can work with the agent. You give it your brief, references, storyboard, and then have a back-and-forth while it helps move the project forward. I like that you can decide how involved you want to be at any moment. Sometimes you want to craft things shot by shot, other times you want help pushing the idea along. Having everything in one shared canvas also works really well. The context of the project stays in one place instead of being scattered across different tools, which makes the process feel much closer to collaborating than just prompting. It’s definitely not perfect yet, but AI video models have come such a long way. It’s definitely not perfect yet, but lip sync is allllllmost there, and it’s amazing how quickly the quality of keeps improving. Also worth mentioning: Luma has now opened access to the public, so anyone can jump in and try it. Always exciting to see tools that expand what a solo creative can make. Curious to see what people build with it.

164

22 Comments

NAB Show 2026. Here’s what we’re hearing this year. The model has shifted. Clients don’t want hardware or file movement. Studio in a Box is resonating. Everything in one place. People, workstations, storage, tools, content. Edit while capture is real and live workflows are the unlock for everyone from YouTube channels to brand clients. In-house teams are feeling it. The idea is simple. The team moves to the media. The media never moves. #cree8 #NAB26 #studioinabox

16

4 Comments

i just saw the first tiny moves toward making AI music work in a live setting. train a model like this with your own output and what you get is your own custom remixer! or don't and maybe get hit with a bunch of copyright violations if your performance goes viral. he's figured out a way to run ACE Step live and using MIDI controllers to morph the noise seed to get variations on output. https://lnkd.in/gCf7dWXi

Big update to the 3D AI Canvas on aianimation.com The latest release makes it easier than ever to build, restyle, and iterate on your scenes using powerful AI tools and 2D/3D assets: 🔹 New Features Projects now saved securely in the cloud Integrated Poly Pizza 3D asset library (+ upload your own models) Major improvements to AI-powered 3D scene restyling Support for Kling 2.1 video generation (7 video models now available) Generate images (with optional transparency) and drop them straight into your scene Main Camera field of view settings. (*plus scene editing view and orthographic support) 🎬 New Canvas View Build multi-scene 3D projects with flexible scene linking Link assets or entire scenes to speed up world-building Restyle prompts, models, cameras, and environment settings can be linked or independent per scene Drag & drop to reuse or duplicate scene elements effortlessly Still in early access — but evolving fast. Would love your thoughts or feedback! #AIAnimation #3DTools #GenerativeAI #CreativeTech #WebGL #AIforCreators

24

3 Comments