Conner Ruhl

open-source AI · protocols · programming · paragliding · building · previously · 🇺🇸🏳️‍🌈

[Conner Ruhl @connerruhl]

Hello SF friends 👋

I'll be in town with my co-founders May 18th-July 1st.

DM me if you wanna make some plans!

[Conner Ruhl @connerruhl]

Soon models will simulate massive sets of human minds and what’s occurring within them to achieve any objective.

For most economically relevant decisions we’re pretty far along, i.e. will this person/group want to buy burritos.

You’ll be able to say “sell more burritos” and slight nudges across the economy in the form of bought and repurposed attention will make it so.

[Conner Ruhl @connerruhl]

More deets: https://t.co/V2pzlz0tss

[Conner Ruhl @connerruhl]

Opus 4.8 is very honest 😇

[Conner Ruhl @connerruhl]

Good office warming @capydotai

[Conner Ruhl @connerruhl]

We’re up to 6 dogs at Bar Darling

[Conner Ruhl @connerruhl]

@Hoodline Oh no where else will I buy double cheeseburgers I always forget are too big at 2:30AM

[Conner Ruhl @connerruhl]

@xlr8harder “honest”

[Conner Ruhl @connerruhl]

Lyon street steps are crazy because what do you mean there is final boss of steps directly after?

[Ruilong Li @ruilong_li]

World models are moving beyond offline generation towards interactive, real-time experiences.

Introducing ⚡FlashDreams⚡: an open-source high-performance inference and serving library built for autoregressive world models:

🔥 Up to 3.10× faster LingBot-World inference
🔥 Up to 2.12× faster Self-Forcing inference
🔥 Up to 1.40× faster Wan2.1 inference
🔥 8 integrated models
🔥 Multi-GPU, streaming, low-latency serving
🔥 Agentic skills that teach you how to use it

FlashDreams is designed for a new generation of AI systems that continuously evolve over time while responding to user interactions. It powers applications across robotics, autonomous vehicle simulation, gaming, and virtual worlds.

Github: https://t.co/xM8LuPaRTS
Docs: https://t.co/IInORNIzy3
Research page: https://t.co/mZ6TLQSpIO
Join the #flashdreams Discord channel at https://t.co/GGOQ0k7liY

FlashDreams is also the runtime backbone behind NVIDIA OmniDreams (https://t.co/PLUt55gxxh)

1/n

#AI #WorldModels #FastInference #PhysicalAI #OpenSource #NVIDIA

[Ruohan Gao @RuohanGao1]

Excited to share our new work, SonoWorld, led by @DerongJin and Xiyi Chen!

Silent worlds are incomplete. We turn a single image into an explorable 3D audio-visual scene, enriching visual worlds with immersive spatial sound.

The framework is modular and training-free and can be built entirely with open-source models while also supporting proprietary models such as Marble for enhanced visual fidelity.

Kudos to the team for making this happen. Thanks as well to @theworldlabs and to @drfeifei for the early feedback and encouragement.

Explore the 3D audio-visual world below (🎧 headphones recommended) 👇

[Conner Ruhl @connerruhl]

@war24182236 I just got food from here like four days ago, had no idea it was like this at night.

[Christian Cantrell @cantrell]

This is what it looks like to generate images from code. And this is how a small team beat Nano Banana. Try @reve yourself at https://t.co/l1UinIJxcB.

[Reve @reve]

Today, we’re launching Reve 2.0, the best 4K image model in the world.

We invented a new way to generate and edit any image using precise layouts. For the first time, it’s possible to create images you can touch.

[Bryce Schmidtchen @_bschmidtchen]

real time and multi shot, try it out, it’s crazy!

[Conner Ruhl @connerruhl]

@xxunhuang Congrats!

[Conner Ruhl @connerruhl]

Brain damage: For the first time I found myself starting a message like this to an actual human.

Diagnosis: I am spending too much time talking with Claude.

[Yohei @yoheinakajima]

slowly we're all realizing that tools should be called from code, not from within the llm api

[Tripo @tripoai]

Introducing Project Eden, a world model research preview from @VASTAIResearch

Project Eden is a persistent, multiplayer world model that fundamentally breaks from existing paradigms by decoupling the underlying world state from visual rendering.

Instead of treating the world as a sequence of transient frames, Eden treats it as a structured, evolving environment that runs continuously, can be modified by user actions, and can be consistently observed from any viewpoint.

[Conner Ruhl @connerruhl]

North/West scouting complete 🫡

[Conner Ruhl @connerruhl]

Call me Dora cause I explora

[Conner Ruhl @connerruhl]

Added location-aware timeline tracking for my personal repository.

Lots more to explore.

[Conner Ruhl @connerruhl]

The location tracking stuff I added to my repo is very fun

[Conner Ruhl @connerruhl]

Call me Dora cause I explora