Sam Sartor
@samsartor.com
I research 3D graphics stuff at William & Mary, when not distracted by rusty sideprojects and fiddle music! Right now I live in Colorado.
Just published docs.rs/weakref/, a Copy+static reference type which checks liveness at runtime!
I saw verdagon.dev/blog/surpris... a few years back and started using a similar weakref implementation in my own weird projects. But now it has its own crate 📦
I saw verdagon.dev/blog/surpris... a few years back and started using a similar weakref implementation in my own weird projects. But now it has its own crate 📦
weakref - Rust
Weakref provides a cheap `Copy + 'static` reference type `Ref<T>`. You can pass it anywhere almost effortlessly, then check if the reference is alive at runtime. The single owner `Own<T>` increments a...
docs.rs
July 5, 2025 at 6:37 PM
Just published docs.rs/weakref/, a Copy+static reference type which checks liveness at runtime!
I saw verdagon.dev/blog/surpris... a few years back and started using a similar weakref implementation in my own weird projects. But now it has its own crate 📦
I saw verdagon.dev/blog/surpris... a few years back and started using a similar weakref implementation in my own weird projects. But now it has its own crate 📦
Just read open.substack.com/pub/garymarc... and there is a take really I want to unpack.
> For what I think are mostly sociological reasons, people who have built neural networks such as LLMs have mostly tried to do without explicit [world] models
> For what I think are mostly sociological reasons, people who have built neural networks such as LLMs have mostly tried to do without explicit [world] models
June 29, 2025 at 8:11 PM
Just read open.substack.com/pub/garymarc... and there is a take really I want to unpack.
> For what I think are mostly sociological reasons, people who have built neural networks such as LLMs have mostly tried to do without explicit [world] models
> For what I think are mostly sociological reasons, people who have built neural networks such as LLMs have mostly tried to do without explicit [world] models
the real paperclip maximizer was the MBAs all along
February 12, 2025 at 10:33 PM
the real paperclip maximizer was the MBAs all along
I really don't like Three Body Problem. I was promised metaphysical dread in the face of a fundamentally impossible unpredictable world, and got <spoiler>it was aliens doing a prank lol</spoiler>. Does anyone know a book that executes on the "science suddenly breaks" premise for real?
February 6, 2025 at 12:36 PM
I really don't like Three Body Problem. I was promised metaphysical dread in the face of a fundamentally impossible unpredictable world, and got <spoiler>it was aliens doing a prank lol</spoiler>. Does anyone know a book that executes on the "science suddenly breaks" premise for real?
Starting at SIGGRAPH and ending at LeamLab Borderless has been really nice. Computer graphics book-ends my Tokyo trip!
December 15, 2024 at 11:41 AM
Starting at SIGGRAPH and ending at LeamLab Borderless has been really nice. Computer graphics book-ends my Tokyo trip!
I setup btrfs on my Linux laptop to get all the cool timeshift stuff. I didn't expect it would also be 100× easier to mount when dual-booting windows!
github.com/maharmstone/... is badass
github.com/maharmstone/... is badass
GitHub - maharmstone/btrfs: WinBtrfs - an open-source btrfs driver for Windows
WinBtrfs - an open-source btrfs driver for Windows - maharmstone/btrfs
github.com
November 26, 2024 at 9:22 PM
I setup btrfs on my Linux laptop to get all the cool timeshift stuff. I didn't expect it would also be 100× easier to mount when dual-booting windows!
github.com/maharmstone/... is badass
github.com/maharmstone/... is badass
Ok hear me out, C is typescript for assembly
November 22, 2024 at 8:14 PM
Ok hear me out, C is typescript for assembly
The "ai-generated art is stolen" argument came up in a conversation today. I think it is bad because regardless of training data I'd still still object to:
- People passing off AI art as their own
- Bots flooding the world with "content"
- Automating away fun parts of life
- People passing off AI art as their own
- Bots flooding the world with "content"
- Automating away fun parts of life
November 22, 2024 at 1:23 AM
The "ai-generated art is stolen" argument came up in a conversation today. I think it is bad because regardless of training data I'd still still object to:
- People passing off AI art as their own
- Bots flooding the world with "content"
- Automating away fun parts of life
- People passing off AI art as their own
- Bots flooding the world with "content"
- Automating away fun parts of life
I went mad today trying to embed a WebGL canvas into PowerPoint! Surprisingly learn.microsoft.com/en-us/office... supports WebGL just fine. But the logistics of actually inserting into a slide and saving plugin state are insane.
Don't follow me into this nightmare
Don't follow me into this nightmare
November 18, 2024 at 10:32 PM
I went mad today trying to embed a WebGL canvas into PowerPoint! Surprisingly learn.microsoft.com/en-us/office... supports WebGL just fine. But the logistics of actually inserting into a slide and saving plugin state are insane.
Don't follow me into this nightmare
Don't follow me into this nightmare
Yah, to me "synthetic data" is still "this big pile of blender renders". But I think we lost the terminology battle on this one
For some reason the term “synthetic data” has become synonymous with LLM/Diffusion model outputs for many people over the past couple of years. Let it be known that there exist ways to generate data other than neural networks. Those are, as of yet, not illegal.
November 17, 2024 at 11:38 PM
Yah, to me "synthetic data" is still "this big pile of blender renders". But I think we lost the terminology battle on this one
I honestly gave up on PIL/skimage/iio/torchvision/cv2 and started doing all my image processing in Rust with the amazing image crate, then yeeting a numpy array over the wall with pyo3. I had to write my own transforms, but that wound up being really useful anyway.
both PIL and scikit-image have had the absolutely bonkers idea that a matrix transform happens in unnormalized pixel space. I have waster days writing conversions every time I see code using a pixel-space transform instead of a texel-space one. why. exblain.
November 17, 2024 at 5:29 PM
I honestly gave up on PIL/skimage/iio/torchvision/cv2 and started doing all my image processing in Rust with the amazing image crate, then yeeting a numpy array over the wall with pyo3. I had to write my own transforms, but that wound up being really useful anyway.
My first grad school paper (samsartor.com/matfusion) was so much more work, but way less fun to look at. Funny how that goes!
Sam Sartor - MatFusion: A Generative Diffusion Model for SVBRDF Capture
SVBRDF estimation from photographs for three different lighting conditions (directional, natural, and flash/no-flash illumination) is shown by refining a novel SVBRDF diffusion backbone model, named M...
samsartor.com
November 17, 2024 at 4:58 AM
My first grad school paper (samsartor.com/matfusion) was so much more work, but way less fun to look at. Funny how that goes!
Since I'm bootstrapping a profile here, I figure I should share the coolest thing I made recently
November 17, 2024 at 4:48 AM
Since I'm bootstrapping a profile here, I figure I should share the coolest thing I made recently
Reposted by Sam Sartor
Sam Sartor, Pieter Peers
Content-aware Tile Generation using Exterior Boundary Inpainting
https://arxiv.org/abs/2409.14184
Content-aware Tile Generation using Exterior Boundary Inpainting
https://arxiv.org/abs/2409.14184
September 24, 2024 at 7:31 AM
Sam Sartor, Pieter Peers
Content-aware Tile Generation using Exterior Boundary Inpainting
https://arxiv.org/abs/2409.14184
Content-aware Tile Generation using Exterior Boundary Inpainting
https://arxiv.org/abs/2409.14184