Julian Togelius(@togelius) 's Twitter Profileg
Julian Togelius

@togelius

AI and games researcher.
Associate professor at NYU; director of @NYUGameLab; co-founder of https://t.co/FnakJLkAXW.

ID:19510090

linkhttp://julian.togelius.com calendar_today25-01-2009 23:13:42

21,9K Tweets

18,0K Followers

1,0K Following

Eugene Vinitsky(@EugeneVinitsky) 's Twitter Profile Photo

Are you an MS student at NYU interested in doing some RL / multiagent / sim work? My wonderful MS students are graduating so we're looking for new folks.

account_circle
Julian Togelius(@togelius) 's Twitter Profile Photo

Can an LLM beat you at your favorite word game, and what does that say about their ability to connect concepts?
New York University NYU Tandon writes about our work solving Connections puzzles with LLMs, with Graham Todd, The Timerino, smearle:
engineering.nyu.edu/news/nyu-tando…

account_circle
Julian Togelius(@togelius) 's Twitter Profile Photo

Generating game content with Neural Radiance Fields? Yes, but you need to take the structure of the game world into account. DreamCraft, a new project from my students and collaborators (I'm on there as well) demonstrates a new method for NeRF-based PCG in Minecraft.

account_circle
modl.ai(@modl_ai) 's Twitter Profile Photo

Thanks, CNET for shedding light on our role in automating development processes in games! At modl, we're passionate about leveraging AI to streamline QA processes to reduce repetitive labor and enable testers to focus on more impactful tasks.
hubs.la/Q02w9md10

account_circle
Julian Togelius(@togelius) 's Twitter Profile Photo

Several of our speakrs at the AI and Games School 2024 bring practical experience of building games on and around modern generative AI - Anna Winterstein ☕ is the latest addition to our roster. See you at the Summer School on AI and Games!

account_circle
Julian Togelius(@togelius) 's Twitter Profile Photo

I agree, with the proviso that the main reason we are nowhere close to AGI and won't get there with deep learning (alone) is that AGI is a meaningless term. (There are many definitions, all bad, and not in agreement with each other.) But yes, scaling works, in a boring way.

account_circle
Delip Rao e/σ(@deliprao) 's Twitter Profile Photo

“Frontier Model Division”.
“Devs must *assert* models safe under *penalty* of perjury.”

WTF. Sacramento has no idea what they are trying to legislate.

Also, what's up with the fast-tracking? I can smell one/two companies pushing this. The California bill is designed to set a…

account_circle