Nam Hee Gordon Kim

Can we build AI players that are not just great at the game, but play *like people*? New #Eurographics 2026 paper: Robo-Saber: Generating and Simulating Virtual Reality Players (Links in the reply!)

🔗 Project page robo-saber.github.io 📄 Paper arxiv.org/abs/2602.18319

Can we build AI players that are not just great at the game, but play *like people*? New #Eurographics 2026 paper: Robo-Saber: Generating and Simulating Virtual Reality Players (Links in the reply!)

6/ Special thanks to the Beat Saber modding and custom mapping community. This work would not exist without the ecosystem they built. And many thanks to my wonderful colleagues across the world who made this work possible! @perttu_h @jaakkolehtinen @xjamesobrienx @xbpeng4

5/ Robo-Saber points toward a future where AI player models help us test content across a much broader space of human player styles. Human playtesters might still assess subjective qualities like funness, but the AI players afford well-defined difficulty/playability metrics.

Code release for our @SIGGRAPHAsia paper on simulating fatigue! Check it out on github.com/noshaba/SIGGRA…

Project page and YouTube video for our @SIGGRAPHAsia paper on discovering fatigued movements for virtual character animation are up: vcai.mpi-inf.mpg.de/projects/Fatig… youtube.com/watch?v=FVOWVz… @DFKI @VcaiMpi @aaltomediaai @suppose628 @NamHeeGordonKim @perttu_h @VGolyanik @marc_habermann

3. "G stands for graphical" is a poor argument. City University of New York is abbreviated as CUNY but it's not pronounced "soo-knee".

2. The closest word to GIF is not GIFT but GIN. Both have edit distance of 1 but GIFT is 4 characters. I say we shall break ties by the complexity of the end product and GIN is definitely simpler.

I found a hill to die on: GIF is most definitely pronounced "jif".

Unpopular opinion: "write so your grandma can understand" is a well-intentioned but bad writing advice. Leave my grandma out of it.

Unpopular opinion: the word "directly" is confusing and doesn't add anything to explaining your method

The schools don't ask me how well you did in my course. They ask me about your potential for graduate school that go way beyond grades. If I don't know you, my answer to these questions is just "I don't know". I get a very strong sense that this will not help your chances.

As a first time reference letter writer, I made the mistake of not turning down a request I wasn't confident about, and I'm really regretting it. This person took my class and did okay (top 30%), but I actually don't remember ever interacting with them.

Dear prospective graduate students, PLEASE choose your referees wisely.

I am looking for more students! If you are interested in the practical and theoretical challenges of creating generalist problem-solving robots, apply to @Mila_Quebec mila.quebec/en/cours/super…

@CalNewport2 Are they so good that they can't ignore you?