
YC W27 prev. @runwayml @lumalabsai @stabilityai @wand_app @laion_ai & Outlier on YouTube
San Francisco, CA
Joined July 2021
- Tweets717
- Following439
- Followers3K
- Likes2.5K
Pinned Tweet
It finally happened. We are releasing Stable Cascade (Würstchen v3) together with @StabilityAI !
And guess what? It‘s the best open-source text-to-image model now!
You can find the blog post explaining everything here:
stability.ai/news/introducin…
🧵 1/5
Introducing, Runway Characters. Real-time intelligent avatars that turn the internet into a conversation. Deployable anywhere via the Runway API, Runway Characters can be customized in any way across every style. All with the ability to embed bespoke knowledge banks, custom voices and instructions.
Start integrating Runway Characters directly into your apps, websites, products and services today. Available now at the link below.
dome | Outlier retweeted
Beyond excited to share Uni-1 with the rest of the world!! 😍
This is by FAR the most ambitious, impressive and exciting model we have worked on @LumaLabsAI
Multimodal intelligence will shape the future of generative media and I’m thrilled to share our flagship SOTA model 🤩
Introducing Uni-1, Luma’s first unified understanding and generation model, our next step on the path towards unified general intelligence.
lumalabs.ai/uni-1

ALT Combine the black and white curly-haired dog with pink bandana, the Boston Terrier in plaid harness, and the black-and-white cat into a single scene where they are dressed in academic regalia, standing before a whiteboard filled with scientific diagrams and text, with the Luma AI logo placed in the top-left corner.
dome | Outlier retweeted
Really excited to share our newest release on a unified model that does understanding and generation all in one. The team really did a great job here!
Introducing Uni-1, Luma’s first unified understanding and generation model, our next step on the path towards unified general intelligence.
lumalabs.ai/uni-1

ALT Combine the black and white curly-haired dog with pink bandana, the Boston Terrier in plaid harness, and the black-and-white cat into a single scene where they are dressed in academic regalia, standing before a whiteboard filled with scientific diagrams and text, with the Luma AI logo placed in the top-left corner.
dome | Outlier retweeted
A good quote that ChatGPT chose for me today:
"Creativity is the residue of time wasted." (Einstein)
In the upcoming AI era, our worth will come from creativity instead of productivity. We'll have to fight the instincts we cultivated in the past grind era: the 996, the "need to listen to a podcast while cleaning the house to minimize time wasted".
We'll have to embrace boredom, lack of structure and loose identities. Such a generational challenge.
Have you ever wondered how the DPO objective is derived and why it looks the way it does? And especially, where do we even start from to derive it?
In my latest YouTube video I go into the details of the DPO objective. Let me know what you think!
The link is in the next thread ⬇️
Also, are there any suggestions for video topics you have?
Just uploaded a new video on YouTube about Proximal Policy Optimization (PPO) and Group Relative Policy Optimization (GRPO). The video shows the full math derivation of the objectives starting with a super intuitive objective!
The link is in the thread below ⬇️
The video also shows the derivation of the REINFORCE objective and Importance Sampling and touches upon TRPO and much more!
I hope you enjoy the video!
After 1.5 years at Luma AI, I have decided to leave and that the AI field is not my thing anymore and that I will pursue a career as a professional surfer.
Just kidding, I‘m joining @runwayml
Any RLHF experts here? Currently working on a video about PPO / GRPO etc., and I want to introduce importance sampling (see 1st screenshot). Is this a valid step to take, to go from the expected value over trajectories to expected value over actions? And is my explanation for this sound or am I missing something?
Because doing so, makes it very straightforward to introduce importance sampling (see 2nd screenshot).
Lol YouTube suggesting creators new video ideas -> title + thumbnail + description
Soon maybe also generate the video for you haha
5
1,091
dome | Outlier retweeted
. @dome_271 has a new video tutorial on Rotary Positional Embeddings (RoPE, link below).
I am a fan of his excellent tutorials.
After a million years I finally uploaded a new video. This time about RoPE! Let me know what you think!
youtu.be/V8r__fXx7tU?si=sAyH…
dome | Outlier retweeted
LAION proudly presents 2 state-of-the-art emotion detection models for voice and face, surpassing Gemini 2.5 Pro and Hume API. They are completely open under a CC BY 4.0 license, alongside a ~5,000-hour voice-acting dataset & 2 expert-annotated benchmarks. laion.ai/blog/do-they-see-wh…
dome | Outlier retweeted
🚀Excited to announce Dream 7B (Diffusion reasoning model): the most powerful open diffusion large language model to date.
dome | Outlier retweeted
Today, we release Inductive Moment Matching (IMM): a new pre-training paradigm breaking the algorithmic ceiling of diffusion models. Higher sample quality. 10x more efficient. Single-stage, single network, stable training. Read more: lumalabs.ai/news/imm



















