Pinned Tweet
I left Stability AI in late July 2024 to join a new stealth HW startup. I want to thank my former Stability AI colleagues for a year and a half filled with exciting challenges, discoveries and technical achievements. I wish them all the best.
5
2
95
16,195
Christian Laforte retweeted
Exciting to see the community building on FastGen and pushing distillation techniques forward! 🚀
This cool new work builds on ideas from our PDD and PiFlow, exploring novel parameterizations of student heads. Great to see these ideas evolving!
This cool new work builds on ideas from our PDD and PiFlow, exploring novel parameterizations of student heads. Great to see these ideas evolving!
Excited to introduce Simplified Matching Distillation (SMD) for few-step diffusion🧵
No auxiliary score model. No adversarial training. No generator modification.
You will get:
• Photorealistic appearance
• High-quality dynamics
• Memory-efficient training
• Mode covering MD works well across images, video, joint audio–video, and 3D gen, validated across model scales up to 33B parameters.
Besides, SMD supports:
- Multi-teacher capability merging (like OPD in LLM)
- Cross-model capability transfer
- Bidirectional teacher to causal distillation
Full release. Paper, code, model and data.
Project Page: zjwfufu.github.io/SMD-Page/
Paper: arxiv.org/abs/2610.18472
Code: github.com/zjwfufu/SMD
Model Card & Data: huggingface.co/zjwfufu/SMD
Demo Video 👇
0
8
62
4,122
Christian Laforte retweeted
Speaker spotlight for Frontier Research in World Model (#SFTechWeek, Fri Oct 9): Dalton Omens from Moonlake @moonlake on "Building Worlds with Code". 1:05 PM PT.
luma.com/video-tnp8
luma.com/video-tnp8
1
5
14
613
Christian Laforte retweeted
Been seeing a lot of Astra for real2sim, but what you need for robotics: physics and affordance.
Our SOTA recipe specifically solves for:
• Affordance: model object details that convey functionality.
• Physics: calibrate simulation dynamics against real-world observations. ion space of a “full-stack simulation engineer” and jointly considers geometry, materials, colliders, articulation, and physics. It automatically calibrates physics against real trajectories and matches task-specific dynamics.
Simulation environments is useful only if it can provide meaningful policy level signals. Methods, comparisons, and results below.
Our SOTA recipe specifically solves for:
• Affordance: model object details that convey functionality.
• Physics: calibrate simulation dynamics against real-world observations. ion space of a “full-stack simulation engineer” and jointly considers geometry, materials, colliders, articulation, and physics. It automatically calibrates physics against real trajectories and matches task-specific dynamics.
Simulation environments is useful only if it can provide meaningful policy level signals. Methods, comparisons, and results below.
31
66
581
122,340
Christian Laforte retweeted
Big milestone for autonomous drug discovery! 🧪
Congrats to @muni_bio on closing the design-to-test loop with @adaptyvbio! Using NVIDIA Proteina-Complexa, their autoresearch agent generated 3 sub-nanomolar TREM2 binders, outperforming previous leaderboards.
Congrats to @muni_bio on closing the design-to-test loop with @adaptyvbio! Using NVIDIA Proteina-Complexa, their autoresearch agent generated 3 sub-nanomolar TREM2 binders, outperforming previous leaderboards.
@muni_bio now gives agents access to the wet-lab. Using @NVIDIAHealth's Proteina-Complexa + @adaptyvbio validation, muni found 3 sub-nanomolar binders, all of which were better than our hackathon winners.
Even if assays take days or weeks, the results automatically feed back g you to orchestrate autonomous scientific research in parallel.
0
8
34
3,212
Christian Laforte retweeted
Over the past year, I've been volunteering with California Search and Rescue occasionally.
If you're in the Bay Area, love hiking or backpacking, want to learn skills like wilderness first aid & navigation, meet like-minded people, and give back to the community, Search and Rescortunity:
cal-esar.org/join-us
If you're in the Bay Area, love hiking or backpacking, want to learn skills like wilderness first aid & navigation, meet like-minded people, and give back to the community, Search and Rescortunity:
cal-esar.org/join-us
3
1
34
2,448
Christian Laforte retweeted
Today, we're announcing not one, but two of the @moonlake team on stage at AUTONOMOUS.
@sharonal_lee and @sunfanyun, Co-Founders of Moonlake AI, will be on our Frontier Stage to discuss Where Robots Learn Before the Real World.
🗓️San Francisco, July 16th
📍 The Midway
🎟️ autonomousfuture.co
Moonlake builds generative world models that turn a plain-language prompt into an interactive 3D environment in minutes. From text to the system
compiling a training experience with physics, agents, and logic.
Out of @Stanford's AI lab, and then again out of stealth in late 2025 with a flying start, with a round led by NVIDIA's NVentures & others.
@sharonal_lee and @sunfanyun, Co-Founders of Moonlake AI, will be on our Frontier Stage to discuss Where Robots Learn Before the Real World.
🗓️San Francisco, July 16th
📍 The Midway
🎟️ autonomousfuture.co
Moonlake builds generative world models that turn a plain-language prompt into an interactive 3D environment in minutes. From text to the system
compiling a training experience with physics, agents, and logic.
Out of @Stanford's AI lab, and then again out of stealth in late 2025 with a flying start, with a round led by NVIDIA's NVentures & others.
1
3
16
5,805
Christian Laforte retweeted
Today, we're announcing not one, but two of the @moonlake team on stage at AUTONOMOUS.
@sharonal_lee and @sunfanyun, Co-Founders of Moonlake AI, will be on our Frontier Stage to discuss Where Robots Learn Before the Real World.
🗓️San Francisco, July 16th
📍 The Midway
🎟️ autonomousfuture.co
Moonlake builds generative world models that turn a plain-language prompt into an interactive 3D environment in minutes. From text to the system
compiling a training experience with physics, agents, and logic.
Out of @Stanford's AI lab, and then again out of stealth in late 2025 with a flying start, with a round led by NVIDIA's NVentures & others.
1
3
21
2,826
Christian Laforte retweeted
Makes me happy when we put all our gaming GPUs + Workstations and Macs to work.
OxCapsule QWEN 3.5-397B demo youtu.be/gDGGy5r29eQ?is… via @YouTube
OxCapsule QWEN 3.5-397B demo youtu.be/gDGGy5r29eQ?is… via @YouTube
4
10
37
7,554
Christian Laforte retweeted
Introducing Moonlake's 3D Agent.
Our agent acts like a technical artist that can build and reconstruct articulated assets and large-scale editable scenes with hundreds of objects from a single image and can improve its generations continuously.
Learn more in the thread below.
Our agent acts like a technical artist that can build and reconstruct articulated assets and large-scale editable scenes with hundreds of objects from a single image and can improve its generations continuously.
Learn more in the thread below.
42
175
1,379
1,169,796
Christian Laforte retweeted
We are happy to announce that Moonlake AI has been named on the 2026 @CBinsights AI 100 list.
One of the 4 on infrastructure and world models. Read more here: cbinsights.com/research/repor…
One of the 4 on infrastructure and world models. Read more here: cbinsights.com/research/repor…
3
9
39
5,658
Christian Laforte retweeted
Taking main stage at AWE 2026 to demystify world models alongside @bilawalsidhu.
9:55–10:45 AM, Grand Ballroom. Details in comments
#awe2026
9:55–10:45 AM, Grand Ballroom. Details in comments
#awe2026
4
5
53
10,444
Christian Laforte retweeted
🚀
Taking main stage at AWE 2026 to demystify world models alongside @bilawalsidhu.
9:55–10:45 AM, Grand Ballroom. Details in comments
#awe2026
0
3
12
2,385
Christian Laforte retweeted
Excited to share that we are working with Nvidia to bring @moonlake's 3D Agent capabilities into NVIDIA's simulation ecosystem (@nvidiaomniverse) to help teams build assets, digital twins, and scenarios at scale.
Here are 8 pairs of input images and the generated .
Read more here: moonlakeai.com/blog/simulatio…
Here are 8 pairs of input images and the generated .
Read more here: moonlakeai.com/blog/simulatio…
27
37
424
919,394
Christian Laforte retweeted
Real-time neural rendering model @moonlake
Works on from any physics/game engine to close the sim-to-real gap.
If this is relevant to your team, I'd love to chat (my DM is open)!
Works on from any physics/game engine to close the sim-to-real gap.
If this is relevant to your team, I'd love to chat (my DM is open)!
13
35
291
50,885
Christian Laforte retweeted
One of the coolest project I've been involved in NVIDIA.
The quality is insane. 👀
The quality is insane. 👀
Today, we released Lyra 2.0, a framework for generating persistent, explorable 3D worlds at scale, from NVIDIA Research.
Generating large-scale, complex environments is difficult for AI models. Current models often “forget” what spaces look like and lose track of movement over o shift, blur, or appear inconsistent. This prevents them from creating the reliable 3D environments required for downstream simulations. Lyra 2.0 solves these issues by:
✅ Maintaining per-frame 3D geometry to retrieve past frames and establish spatial correspondences
✅ Using self-augmented training to correct its own temporal drifting.
Lyra 2.0 turns an image into a 3D world you can walk through, look back, and drop a robot into for real-time rendering, simulation, and immersive applications.
➡️ Learn more: research.nvidia.com/labs/sil/proje…
📄 Read the paper: arxiv.org/abs/2604.13036
5
6
167
13,499
Christian Laforte retweeted
Quote retweet the post and we'll fast track your beta access application 💫
Sign up: app.moonlakeai.com
Read more here: moonlakeai.com/blog/building-…
Sign up: app.moonlakeai.com
Read more here: moonlakeai.com/blog/building-…
4
17
90
25,460
Christian Laforte retweeted
🚀 Diffusion too slow? Fix it in a few steps.
📢 Introducing NVIDIA FastGen — a plug-and-play research library for turning slow diffusion models into high-quality few-step generators.
⚡ What’s inside:
• Consistency & MeanFlow (CM, sCM, TCM, MeanFlow)
• Distribution Matching (DMDng-video generation (CausVid, Self-Forcing)
• Fine-tuning & KD (SFT, CausalSFT, KD, Causal KD)
🧠 Includes:
📷 EDM, DiT, SD 1.5, SDXL, Flux
🎬 WAN (T2V / I2V / VACE), CogVideoX, Cosmos Predict2
✨ One unified interface. Research-ready. Apache-2.0.
🔗 Blog: nvda.ws/3LARhFy
💻 Code: github.com/NVlabs/FastGen
📢 Introducing NVIDIA FastGen — a plug-and-play research library for turning slow diffusion models into high-quality few-step generators.
⚡ What’s inside:
• Consistency & MeanFlow (CM, sCM, TCM, MeanFlow)
• Distribution Matching (DMDng-video generation (CausVid, Self-Forcing)
• Fine-tuning & KD (SFT, CausalSFT, KD, Causal KD)
🧠 Includes:
📷 EDM, DiT, SD 1.5, SDXL, Flux
🎬 WAN (T2V / I2V / VACE), CogVideoX, Cosmos Predict2
✨ One unified interface. Research-ready. Apache-2.0.
🔗 Blog: nvda.ws/3LARhFy
💻 Code: github.com/NVlabs/FastGen
8
54
363
49,488