Moshi is one of the best open source full-duplex voice models out there. The architecture is dense, so we spent a few days studying it and wrote up what we learned, with diagrams to make it click faster.
Let us know if it was helpful 🤠
trained a MIRA style multiplayer pong world model on a single RTX 5080. I used their training code + added audio, so every pixel and sound is denoised from the same latents (no game engine anywhere).
this clip is the model playing itself in realtime
We tried building Google DeepMind’s SIMA2… on a Temu budget, in Roblox. Result? It played worse than my 4-year-old nephew and kept 360’ing into trees like a drunk Roomba. Full post: https://t.co/TE6LTP6HXb
Every founder I meet worries about missing this moment.
The best worry about wasting it. There’s a difference.
It’s a waste to ignore how much the world has changed. It’s a waste to think you can capture value with pure software the same as 5 years ago.
It’s a waste to build something small.
We have seen the shift slowly and then very quickly at @southpkcommons.
Here’s who we want in the Founder Fellowship now: hardware tinkerers, mad scientists, obsessives, biohackers, people who build nuclear reactors in their basements.
People who want to get their hands dirty and touch grass and atoms.
If you are only building software, then please (for your own sake!) have a thesis that all your friends laugh at you about. Heresy is the price of ambition.
Then put yourself in the right environment to maximize your ambition.
(Apply by August 2nd)
Introducing MIRA.
A playable, multiplayer world model.
A dream of Rocket League.
Trained on 10k hours of data collected with publicly available bots, MIRA learns the dynamics of a four-player game. The model runs in real time at 20 fps, based on the keys you and the other players press.
Built by General Intuition and @kyutai_labs, in collaboration with Epic Games. Not used to develop Rocket League.
▶️ Play the demo, read the technical report, and explore the open-source code at https://t.co/JjlsamGE1D
At ICML? Find us at Booth 111 to try it yourself and dig into the results with the team.