Freelance Augmented Reality Developer Lucky Bulldozer LTD, Virtual Cinematographer, Photogrammetry, iOS, ARKit/RealityKit & visionOS! Flotogram & ArC creator.
Auckland, New Zealand
Joined February 2015
- Tweets1K
- Following1.1K
- Followers1.1K
- Likes2.2K
Pinned Tweet
I'm hoping I'm the first in the world to get Apple's Object Capture code working on the iPhone, I pulled an all nighter as soon as it was announced - was I? Official documentation comes out tomorrow -#augmentedreality #ios #iPhone #WWDC23
Dan Monaghan retweeted
Not one, but six new Agent Skills! Early adopters improved Xcode build times by 78%.
github.com/AvdLee/Xcode-Buil…
After weeks of work, I launched my latest Agent Skill at Let’s Vision in Shanghai. Here’s how it works:
It runs 3 clean + 3 incremental builds
It analyzes build logs and project settings
It proposes an optimization plan
It applies improvements after you approve
It re-benchmarks and reports the improvement
All free → try it on your Xcode projects.
Let me know how it goes in the replies, and open a PR if you add your results to the README!
Dan Monaghan retweeted
OpenUSD v26.03 now supports 3D Gaussian Splats! Makes me confident we'll see it supported in the updated RealityKit at WWDC26 🔥
Dan Monaghan retweeted
braintrance.net has one of the only public world generation models available for use --Today!
Our gaussian splat worlds generate in *less* than a minute, and are really good at keeping the visuals of people!
Now that’s how you test a depth model!
Just tested Depth Anything 3 inside ComfyUI + Blender — this thing is insanely good!
Massive respect to @bingyikang & amazing team for the model, to PozzettiAndrea for bringing it into ComfyUI, and to @Wildminder for spreading the word.
#DA3 #DepthAnything3 #AI #ComfyUI #blender
Dan Monaghan retweeted
An archaeologist friend of mine wished he had a certain software for 10 years. I vibecoded it for him that evening over a bottle of wine.🍷
He never even used ChatGPT before, and his mind was blown. 🤯
Link, repo, and story below.👇
This is an incredible new way to render texture data - decoding the JPEG (or any compressed image) after visibility is known, right on the GPU. Deferred texturing meets compression, only visible 16×16 blocks ever get decoded. Game changer, literally, gigs could be saved.
A really fun project we've been working on: Real-Time Rendering with JPEG textures! Ended up working way better than we expected (1500+ fps). Will be interesting extending this to modern formats like AVIF or JPEG XL.
Paper: arxiv.org/abs/2510.08166
Code: github.com/elias1518693/jpeg…
Wrapped up the final day of @SwiftServerConf at the National Science Museum in Kensington. As an AR Swift dev, I was blown away by how far Swift on the server has come. Staying in Swift for AWS/Cloud/local changes everything. Huge thanks to the team - see you in 2026!
An interesting trade... an invite from Natively.dev to the new video generation model from OpenAI, Sora - or is it RockStar GTA6 - all the vids I've seen are GTA. Hola@ sora.natively.dev or the Sora Invite Manager in the App Store #Sora #SoraInvite #AI #Natively
Did LumaLabsAI remove the Gaussian Splat processing from their API - docs now have no mention? Who are the alternatives - is there a professionally aimed service out there that can render out at speed and adjustable quality? #GaussianSplatting
So cool you got mono in there! I tried to run local on my Mac and it still went for the Cuda option and asked for 96 gigs when I threw a folder at it. Super impressed, I processed a single frame of a chandeliers and it got the depth really well, def good enough for some VFX elems
Meet MapAnything – a transformer that directly regresses factored metric 3D scene geometry (from images, calibration, poses, or depth) in an end-to-end way. No pipelines, no extra stages. Just 3D geometry & cameras, straight from any type of input, delivering new state-of-the-art results 🚀
One universal model enables SoTA for:
🔥 Mono Depth Estimation
🔥 Multi-View SfM
🔥 Multi-View Stereo
🔥 Depth Completion
🔥 Registration
… and many more possibilities! – plus everything is metric 🎯
We release code for data processing, training, benchmarking & ablations – everything Apache 2.0!
Details & Links 👇
From ArC, the iPhone cinema camera simulator I’ve led for a few years, to real-time point-cloud scanning, Gaussian splats, here’s my latest AR work and prototypes. Always open to projects pushing AR & filmmaking. #ARKit #ios #VirtualProduction #photogrammetry
Epic vision for spatial positioning, if you’re an Augmented Reality developer, watch this, TWICE. Not too tech, it’s about the vision of how our apps *should work. No more starting the user from scratch, all 3D assets remain where they last left off, fusing them into meat space.
I’m wanting to talk to someone with real experience making iOS app for AR with Unreal Engine. My initial investigation is …. slow af build times, on an M2 Mac. I’m used to Xcode & I’m bored if a build takes more than a minute. Is it desktop on big windows box or not worth it?
Finally.... I can't wait to see what comes of this concept.
📢📢📢 𝐓𝐫𝐢𝐚𝐧𝐠𝐥𝐞 𝐒𝐩𝐥𝐚𝐭𝐭𝐢𝐧𝐠 𝐟𝐨𝐫 𝐑𝐞𝐚𝐥-𝐓𝐢𝐦𝐞 𝐑𝐚𝐝𝐢𝐚𝐧𝐜𝐞 𝐅𝐢𝐞𝐥𝐝 𝐑𝐞𝐧𝐝𝐞𝐫𝐢𝐧𝐠
arxiv.org/abs/2505.19175
trianglesplatting.github.io
TL;DR: use triangles rather than Gaussians, and everything works just as well.
Code (soon) github.com/trianglesplatting…
Dan Monaghan retweeted
This is called SigNeRF and the goal is to make #NeRF inpainting easier and more accessible. It is just a matter of time before #3D #generativeAI platforms implement more powerful ways to take control of the #creative process.
Full article below 👇