@GoogleDeepMind This is wild One brain for any robot multi-robot teamwork? Gemini Robotics 2 feels like the future arriving
English
Reena Tech & Tool
69.5K posts
Reena Tech & Tool
@ReenaTech0
Testing Al tools, exploring trends & sharing what works Open Collab working with @invideoOfficial DM for paid collaboration 📨[email protected]
Katılım Mayıs 2024
2.4K Takip Edilen17.2K Takipçiler
One brain. For any robot. 🤖
We’re launching Gemini Robotics 2: our next-generation physical AI bringing full body intelligence to humanoids, advanced dexterity, multi-robot teamwork and more.
English
@BytePlusGlobal Seedance 2.5 👀 Those webinars look packed with value. Aug 7 Aug 13 locked in
English
Discover what's next with Dreamina Seedance 2.5.
Join 2 sessions on the API upgrade, real-world applications, and case insights.
Webinar 01 | Aug 7
API Upgrade & Real-World Applications
Webinar 02 |Aug 13
Real-World Case Study
18:00 BJT / 12:00 CEST / 11:00 BST
English
@appGarly This is the kind of progress that matters 38.6% conversion steady engagement trust building right. Onward
English
Steady progress compounds quietly.
Nearly every registered account is still engaging. Short-term demo conversion hit 38.6% in the last day. Interest is converting into real usage.
We’re still early. Very early.
But the foundation is forming the way it should, building something that may one day protect someone is not a race metrics but It’s a process of earning trust, one reliable step at a time.
Onward. 🧄
English
@cfabricacom Love this No credits no uploads runs locally? Studio Desktop is such a win for creators
English
⚡ Studio Desktop: No Credits. No Uploads. No Waiting Rooms.
Run powerful AI tools locally on your computer with Studio Desktop. Easily remove backgrounds, upscale images, and enhance your designs — 100% for free.
🔗 Stop Paying for Credits. Go Desktop: desktop.creativefabrica.com/?utm_source=x&…
English
@MiniMax_AI That’s huge 🔥 Breaking task modality barriers with an open model? MiniMax H3 sounds like a big leap
English
MiniMax H3: An open model breaking the boundaries between tasks and modalities
Leanna Ren@RenLeanna
This exactly echoes our training methodology: language/image/video/audio are all deeply interconnected modalities of human experience. And multimodal intelligence should be grounded in language which is a generalizable and scalable computational system. Really, really appreciate you digging this up!🙌
English
@Replit Big week for Replit 🔥 Design live smarter Agent longer sessions chef's kiss
English
This Week in Replit Our biggest design moment yet. Here's what shipped: 1) Replit Design is live, the next era of design for everyone 2) Model Selector, choose the best intelligence for the task 3) Follow-up Tasks, an Agent that knows what's next 4) A rebuilt usage page + longer sessions Thread 🧵
English
@askOkara Option 2 all the way An X agent that sounds like you? That’s the sustainable move
English
there are only 2 ways to create content in your voice:
1/ write everything yourself (time-consuming and hard to sustain)
2/ let our X agent learn your voice from your past posts and write tweets that sound like you
Okara@askOkara
Today we're introducing Claude for X Enter your X username and it learns your voice, style, and content themes from your past posts, then writes tweets in your voice every day. Try it now at okara.ai/x
English
@CreaoAI Finally someone explains it live 👏 One prompt agent plans, codes, builds. Zero setup is huge
English
Still don't get what an AI agent actually is?
New CREAO Academy episode shows you, live:
→ Sign in
→ Type one prompt
→ Watch your agent plan, write and run code, and build in real time
No workspace setup. No config screen. Zero to your first working session in under 5 minutes.
Plus: why most AI tools forget everything the second they finish a task, and what makes this one different.
English
Train a policy only on successful demos and it recovers from failure 0% of the time.
Recovery data is the fix, but teleoperating it is slow. Egocentric human recovery clips are ~10x faster to collect — and we transfer intent, not motion.
EgoRecovery 👇
arxiv.org/abs/2607.19745
English
Day 3: Turning Ideas into 3D Assets with V2Fun
Real places, reimagined. 🌍
Using V2Fun, I turned real-world references into reusable 3D assets—one more step toward building my dream town.
#V2Fun #AI3D #3DModeling #WorldBuilding #Blender3D #GameArt #CGI #LevelDesign
English
A speculative campaign for @oakley featuring the iconic Radar® EV Path®.
Creative Direction & Visual Production: @zarixstudio
Made using Claude and Nano Banana Pro within Flora
Zazzy@zazzygfx
Campaign Concept- Golden Kiss by @beguilebyomaa Every frame was intentionally art directed to reflect the emotion behind the scentnot just the bottle. Creative Direction, Art Direction & Visual Production: @zarixstudio
English
@ArtificialAnlys @deepseek_ai Huge drop 🔥 284B with 13B active and MIT license? DeepSeek V4 Flash 0731 is a beast for open weights
English
DeepSeek V4 Flash 0731 is now open weights!
@deepseek_ai has just released the weights for its new flash tier model, DeepSeek V4 Flash 0731. With a score of 50 on the Artificial Analysis Intelligence Index, it lands among the top 3 open weights models on the leaderboard. The weights are released under the MIT license, allowing unrestricted commercial use and modification.
DeepSeek V4 Flash 0731 shares identical architecture and pricing with the earlier DeepSeek V4 Flash. At a size of 284B total parameters (13B active), released in mixed FP4/FP8 precision at ~167GB total file size, it lands on our Pareto frontier for Intelligence Index vs. Total Parameters. Among open weights models, DeepSeek V4 Flash 0731 delivers a significant leap in intelligence for its size class. DeepSeek V4 Flash 0731 is also available now through DeepSeek's first-party API.
Check out Artificial Analysis to compare DeepSeek V4 Flash 0731 with other leading open weights and proprietary models: artificialanalysis.ai/models
English
@fuserstudio Love it A chess clock in just a few prompts? Fuser is magic
English
@HappyHorseATH This is huge Storyboard to high-end ad in minutes? Happyhorse 1.1 is a game changer for saving time budget
English
With just a storyboard and a few minutes, you can generate a high-end ad. Happyhorse 1.1 can save your costs and free up your time! 🐎
Synthia@AIwithSynthia
What if a premium matcha commercial could be created from a single storyboard? 🍵 @HappyHorseATH This is a Matcha Latte concept ad created entirely with HappyHorse 1.1. Instead of planning an expensive commercial shoot, I started with a storyboard reference that defined the entire visual narrative from the opening hero product shot to the final golden-hour café celebration. HappyHorse accurately interpreted each storyboard frame, adhered closely to both the visual references and prompt, and transformed them into a polished commercial with 12 seamless cinematic scenes. The model maintained consistent character appearance, outfit, product design, glass shape, matcha color, and branding details throughout the video, making every shot feel like part of the same professional campaign. I also used Camera Control to create smooth push-ins, cinematic tracking shots, macro product reveals, handheld lifestyle moments, and elegant transitions, giving the final video the look of a high-end beverage advertisement without the complexity of traditional production. This workflow shows how creators and brands can rapidly produce commercial-quality advertising with dramatically lower production costs while keeping product consistency and storytelling intact. #happyhorse
English
@riverflow_ai Left side energy More ideas, still on brand. Riverflow is
English
@BRContentNews @bhphoto That anamorphic look is unreal Horizontal flares oval bokeh just hit different
English
Anamorphic lenses compress a wide field of view onto a standard sensor, then get "de-squeezed" in post. The signature look: horizontal flares, oval bokeh. Genuinely distinctive, genuinely expensive.
Featured video by @bhphoto
bottlerocketcontent.com/what-is-anamor…
English
@Anima_Labs This looks amazing 🍿 Love the witch chase vibe + that potion escape. Cinematic style is a huge step up!
Midjourney Nano Banana Seedance 2 came together perfectly. Can’t wait to see what world you build next
English
New Short Film 🍿
I wanted to experiment with a new animation style and push the cinematic side a little further.
I also had a lot of fun creating this character and imagining her world.
A young witch is spotted in a crowded market, chased through the streets, and uses a mysterious potion to escape her pursuers.
Midjourney + Nano Banana @magnific
Animation : Seedance 2 sur @dreamina_ai
#DreaminaCPP
English
@BlockadeLabs Love this 🌎 True HDRI 3.0 faster better lighting. Huge for games automotive
English
True HDRI 3.0 environments are coming in Skybox AI 1.14.11 🌎
Rebuilt engine. Faster generation. More accurate light values. Decisions you shaped in previous feedback.
Built for what's next in immersive tech 🚗 Automotive, 🎶 Entertainment, 🕹️ Games.
English
@Telegraphprotoc Makes so much sense 👌 One maintained network beats a pipeline that rots. Smart shift
English
Most internal verification pipelines start strong and gradually decay, because maintaining them competes for engineering time against everything else on the roadmap. Telegraph externalizes that layer: one network, maintained continuously, serving many integrations - not a one-time build that quietly becomes unreliable once the team that built it moves on. See the docs: docs.telegraphprotocol.com/docs
English
@JoinPond Last call 🔥 $150 to hunt bugs in 2 hours. Go get it!
English
