The glass materials are insanely good — the light refraction and detail are next level
Lux3D’s material generation is perfect for game assets, especially the metal and ceramics
Already eyeing that hackathon.
THE NEXT CHAPTER STARTS WITH YOU. 🧪
50 new researchers wanted.
Try NOAH free. Investigate a token. Help build what comes next.
Airdrop: 25 planned places. Rewards require launch, funding, final rules and admission.
✅ https://t.co/VKe2yVGZK1
✅For more updates Visit X:
https://t.co/szGoYkdS9N
Dreamina is ByteDance’s AI creation platform; CapCut and Seedance come from the same company.
Use your CapCut Pro account. Open the link below and click “Sign in to reveal.”
All three reward types are included; amounts vary by draw. Basic discounts go up to 98% off. Free generations are 720p at any supported length.
Dreamina and CapCut have something for CapCut Pro members to use in their next edit.
Up to 10 free Seedance 2.0 generations, bonus credits, and a month of Basic for as little as $0.29 if you’re new to Dreamina. Every eligible user gets rewards.
https://t.co/5KWNp1EzxG
#DreaminaPartner
Building a $100M revenue run rate with just $5M raised.
That's not just impressive growth. It's a different way of building an AI company.
Retell is proving that enterprise AI voice agents can deliver real business value at scale.
The future of AI voice is getting interesting.
Static tools = Digital storage.
AI Systems = Digital workforce.
Most people are stuck with static sheets.
Lark Base helps you build an intelligent AI system.
Here is how to build intelligent workflows without code 🧵👇
Anyone else find themselves testing every new AI coding tool that comes out? 😂
Here's another one to explore.
Qoder is Alibaba's coding agent (the company behind Qwen), and they've just launched a new
stealth model called Qoduck.
Try it free for a limited time:
https://t.co/eKoKGB2FOh
this is actually a crazy offer for creators
@frevana_ai is matching your 2026 spend on Higgsfield, Runway and other AI creation tools with 1:1 credits
and yes, your old Higgsfield and Runway receipts count
check it out, link below 👇
Introducing the YouArt Desktop App: your AI creative teammate on your computer.
It uses the best LLMs like Fable 5.2 to make creatives. It can operate creative software on your computer, including CapCut, DaVinci Resolve, Adobe Premiere and many more.
It brings client feedback onto a native canvas, and you can ask agents to address all the comments while you have a coffee.
It’s the most cost-efficient video agent. It supports all third-party API keys and uses smart routing to send every clip to the cheapest provider.
Have fun!
Just one quick AI image.
Also me 20 minutes later: What if I changed the background? What if I tried retro film? What if I made it cinematic? 👀
This week removes that creative friction—purchase or upgrade to CapCut Web Ultra Nov 13–19 to enjoy 7 days of unlimited image generation on Seedream (5.5 / 5.0 Pro / 5.0 / 4.5 / 4.3) and GPT Image 2.5 at 1K & 2K resolution (available in the UK, France, Germany, Italy, Spain, Mexico, Brazil, Canada, Japan, South Korea, and Australia).
Consider this your official invitation to fall down the creative rabbit hole.
Head to Design Studio or use Generate Image in the Video Editor on CapCut Web to test every ""what if""—unlock Ultra before Nov 19!
https://t.co/8u0Snae7mH
#CapCutWeb #CapCutWebDesignStudio #CapCutWebVideoEditor #AIImage
One thing PAN 2 highlights is that visual prediction alone isn't enough.
Without proper reasoning, even a powerful diffusion model struggles to complete multi-step tasks.
The LLM handles the sequence and logic, while the diffusion model handles the visual side. Both are needed to make the plan work.
Physical intelligence needs more than just recognizing what's in front of it.
An agent also needs to understand what will happen after each action.
PAN 2 takes an interesting approach by combining conceptual planning with visual simulation, allowing models to think through multiple steps before acting.
The gap between predicting a single action and completing an entire plan is bigger than it looks.
PAN 2 addresses this by combining language-based reasoning with visual world simulation.
Getting each step right matters, because one wrong move can break the whole sequence.
Instead of committing to the first action, imagine an agent testing a few possible outcomes before making its move.
That's what makes PAN 2 interesting.
It can simulate different actions, compare what happens next, and choose the option most likely to work. A useful direction for physical AI.
What stands out about PAN 2 is that it doesn't just predict what comes next. It lets an agent explore different possibilities before choosing an action.
Combining LLM-based reasoning with video generation gives the system a way to evaluate potential outcomes rather than simply guessing.
An agent that can visualize the consequences of its actions before taking them has a very different way of approaching decisions.
PAN 2 is exploring exactly that.
Test multiple possibilities, understand the outcomes, and make a more informed choice instead of acting blindly.
World models become much more useful when they can help agents evaluate decisions, not just generate realistic scenes.
PAN 2 combines conceptual reasoning and visual prediction to simulate possible futures.
And making the entire model available under Apache 2.0 makes this even more interesting for the research community.