Fan Jiang

Robotics PhD @GeorgiaTech
Cosplayer & Photographer & Part-time CTF player :)

Comments are my own and is in no way endorsed by my employer.

[andrew blinn in NYC @disconcision]

Drag-and-drop makes the digital graspable, but building it well is fiddly enough that many don’t bother. Declarative Dragging (with Josh Horowitz & Jeff Heer, #uist2026 best paper award), lets you make an element draggable just by listing the states dragging it can reach.

[Fan Jiang @FanOnRobotics]

@DominoTree OLD MCDONALD WARNING

[Gracia @gracia_hq]

Real-time relighting on 4D Gaussian Splatting characters in Unreal Engine with the Gracia Plugin. As the key light moves, the characters take on the scene’s lighting, with cast shadows and self-shadowing across clothes, hair and skin.

[Ken Goldberg @Ken_Goldberg]

My brilliant postdoc @Shuangyu0_0Xie found that Gemini 4 Argon did very well on spatial robotics queries so she used Claude to reverse-engineer their skill library and found it also works well for Agentic Robotics: @carolina_parada

[Fan Jiang @FanOnRobotics]

Always the same top quality research from @keenanisalive

[Keenan Crane @keenanisalive]

If anyone's interested in what I've been up to lately, I've been leading development of some 3D foundation models at @Roblox—like a next-gen texture generator, released today: https://t.co/YVpMG1Y0Jp

4k resolution, PBR materials… and it's free. :-)

Also gets some things right that are important in practice (and annoying with other tools), like generating good texture even in spots where different parts of the mesh intersect.

Was happy this required not just ML, but some classic signal processing to really get right. Great work, by a great team.

Has definitely been a journey of understanding how generative AI tools can be designed as just another level of abstraction for creation—rather than a “one shot” replacement for artists. Still a lot of work to do, and more to come…

[Giseop Kim @GiseopK]

Our LT-Mem paper won the Best Paper Award at IROS 2026 — 1 out of 1,933 papers!

Huge congratulations to Yumin Lee and Hyoseok Ju. Proud of our APRL team at DGIST!

https://t.co/VSp54fd3nE

#IROS2026 #Robotics #SpatialIntelligence #DGIST #APRL

[Fan Jiang @FanOnRobotics]

Pair this with the amazing Aerospace i3-like tiling window manager (https://t.co/4vw4NGSyMn) by @nikitabobko you can have global window switching without any hotkey conflicts.

[Fan Jiang @FanOnRobotics]

I made https://t.co/QT8vATkprR, a web visual tool that allows you to program your Macbook keyboard just as if it is a fancy keyboard powered by QMK/ZMK Studio. Vim-mode for all apps now :)

It is a frontend programming tool for the amazing Karabiner Elements by @tekezo .

[Fan Jiang @FanOnRobotics]

It achieves this by compiling the layer transitions into a finite state machine encoded in Karabiner Elements' Complex Modifications language. This eliminates the need for a person to manually write configs.

[MMitchell @mmitchell_ai]

Fun fact: The term “AI” was chosen to avoid the “biological baggage” of cybernetics, and because John McCarthy hated Norbert Wiener.

[Lu Gan @ganlumomo]

🎉 Excited to share our recent work on deformable object manipulation, 𝗣𝗵𝘆𝘀𝗖𝗼𝗥𝗲, accepted to 𝗖𝗼𝗥𝗟 𝟮𝟬𝟮𝟲. Congrats to @hcy1n @ShuohanT!

𝗖𝗼𝗱𝗲 & 𝗗𝗮𝘁𝗮: https://t.co/OE1o5P06PP

#CoRL2026 #RobotLearning #DeformableObjects

[Richard Wei @rxwei]

It supports both the on-device model and Private Cloud Compute! We built it to enable effortless automations for Mac users. It's free, zero setup, no privacy compromises.

[Richard Wei @rxwei]

Jev is a fantastic example of co-designing SDKs and models to unlock a new paradigm!

When we were designing the Swift API for Apple’s Foundation Models framework two years ago, we wanted to make type-safe structured responses the standard for apps on our platform. That is why we created the `@​Generable` API, making constrained decoding approachable to app developers. This effectively eliminated structural hallucinations in both tool call arguments and responses, which were especially important for a small on-device model. https://t.co/TnDSoCImNw

It would be fascinating to see decision models conform to the `LanguageModel` protocol and work with `@​Generable` natively! The best part is that all the APIs are already there: Bools are "noul", enums represent choices, and `@​Guide` descriptions are instructions. A Swift model implementation simply needs to handle a schema conversation, and output probabilities can be stored directly in the response metadata. https://t.co/q5PATIZ0bs

[Fan Jiang @FanOnRobotics]

@keenanisalive How about multi-part geometry (assemblies)? The current models generate and replicate single-piece models really well but I found them still struggle to generate working assemblies with complex geometrical interactions.

[Mingrui Zhao @mingrui_zhao1]

3D models can look right and still fall apart.

Introducing Functionalization: graph completion adds missing structure and connectors to rectify motion and make models work.

SIGGRAPH Asia 2026
Paper, code, data available at: https://t.co/P5PYq6HFzU

[Fan Jiang @FanOnRobotics]

Bosstown Dynamics

[cocktail peanut @cocktailpeanut]

We Must Pace the Frontier

[Nick Price @DominoTree]

Fixed a bug in ZFS that has existed since when Sun Microsystems existed (RIP)

https://t.co/vg38ETYi4m

[Simone Foti @simo_foti]

🔥🏆Heat Kernel Textures (HKTex) received the Best Paper Award at #ECCV2026. 🏆🔥

UV-mapping anchors 3D texturing to the past. Inspired by 3DGS, we rethink textures using heat kernels and NO splatting!

https://t.co/02WZhcxT9B

[Fan Jiang @FanOnRobotics]

@ducha_aiki Ah it's Jakob on the stage!

[Fan Jiang @FanOnRobotics]

@LvZhaoyang Recently there has been an influx of PRs that has been generated purely based on an open issue report on GitHub. They look legit at first glance that "fixes issue x" but in fact the original issue was bogus or invalid. I imagine the same would happen with mathematics :D