Lyell

Interactive real-time, streaming music generation is something I’ve been looking forward to for quite a while. Lyria RealTime (Experimental) by @GoogleDeepMind controlled via #TouchDesigner @1null1 #LOPs

Lil Test Output from the new StreamDiffusion V3! So FuN! #metagnome @dotsimulate @DaydreamLiveAI

Just posted a StreamDiffusionTD update on patreon for first time in many months... On april fools day… I promise it’s not a joke.

Never been able to explore latent space like this before.. 👀

Llm is a powerful tool

Ai is dumb

@TvanHelsdingen By deploy I mean like convert to idea, not like deploy some ai agent slop online

@TvanHelsdingen Yeah it’s super fun to build/watch but hard to deploy. these deterministic babies need more powerful procedural pipelines and lots of design/architecture monitoring

Pretty much my experience. Love agentic coding, and i’m building stuff that I couldn’t have done without them… but you really have to babysit them throughout the process. They don’t make smart architectural decisions themselves.

We are on the cusp of realtime getting interesting again thanks to the fine work of @dotsimulate and @cerspense If you loved deforum, you're gonna love this!

Opus kinda lazy ngl

@JoelSchooling Yes for sure. Grab the discord link from about page on the op

@dotsimulate I'm upgrading StreamDiffusion to the new one and having difficulties. is there a forum for common errors? i don't want to bother you directly,....

From SFMOMA website: “Interact with the work of Alejandro Cartagena through The Latent Space, an AI-driven large-scale projection. Designed and developed by Hugues Bruyère in collaboration with Cartagena, The Latent Space is an AI model trained on the artist’s photographs of homes and urban environments to explore representational bias in training datasets. By manipulating physical objects, viewers can shape and transform the projected image, revealing new outputs generated by the model.” @halecar2, this is so cool 😎🙌❤️🔥

GROUND RULES open at @SFMOMA. Unbelievable to think that so many people will be seeing the work.

@evrythngissoft @SFMOMA it was an amazing experience!

Exciting to see real-time interactive AI projects find their home on our community hub. Today, we're featuring Real-Time Canvas, a live drawing app built for rapid prototyping that utilizes our hosted version of StreamDiffusionTD from @dotsimulate. Grab the project file on our site: app.daydream.live/creators/Formi…

Full house at our @MUTEK_JP Beginners workshop at @TunnelTokyo!

Daydream Release: Open-source real-time AI video just leveled up. SDXL + IPAdapters are now live in the Daydream API and in both local and hosted versions of @dotsimulate’s StreamDiffusionTD component for TouchDesigner x.com/i/broadcasts/1…

@DaydreamLiveAI @as_ws__ @1null1 Looking forward to it!

Tomorrow’s the day! @dotsimulate & @as_ws__ will be walking through the latest update of the TouchDesigner (@1null1) StreamDiffusion component now featuring: ✨ Native SDXL support 🖼️ Image prompting ☁️ Hosted inference via Daydream 🎥 Real-time video generation RSVP: luma.com/k7x8enky

let’s go nyc