Geeve George

@aryashantnu People arent ready for this one!

You were always the director. Now you have the studio.

@smykjain @bnnlabs @upsurgelabs The goats @bnnlabs !

Fluid's Lite USD video was made by @bnnlabs by @upsurgelabs. Best in the game, community loved it! Another video for DEX v2 in works! Drop a message, if you are looking to create high end AI videos for your launches! I'll connect you! 🍌

the part no one's saying out loud: anything that can be done by pressing a key or tapping a screen will be automated completely. engineers and designers are just the front line.

@letian_fu @DrJimFan from the video, the motion looks super odd. just look at it. there is something that makes the cube stick to the bottom cube.

The cubes and apples are not magnetic: in fact, I really like testing these cubes, as they are quite big for standard parallel jaw grippers. This reduces tolerance and can give you a better understanding of motion accuracy without breaking the robot (the cubes deform!): amazon.com/dp/B00DJPK9GI

The power of the Claw, in the palm of a robot hand. Agentic robotics is here! Today, we open-source CaP-X: vibe agents, alive in the physical world. They incarnate as robot arms and humanoids with a rich set of perception APIs, actuation APIs, and auto synthesize skill libraries as they go. CaP-X is a strict superset of our old stack, because policies like VLAs are “just” API calls as well. It solves many tasks zero-shot that a learned policy would struggle with. And we are doing much more than vibing. CaP-X is our most systematic, scientific study on agentic robotics so far: - We build a comprehensive agentic toolkit: perception (SAM3 segmentation, Molmo pointing, depth, point cloud), control (IK solvers, grasp planner, navigation), and visualization (EEF, mask overlays) that work across different robots. - CaP-Gym: LLM’s first Physical Exam! 187 manipulation tasks across RoboSuite, LIBERO-PRO, and BEHAVIOR. Tabletop, bimanual, mobile manipulation. Sim and real. Can’t wait to see the gradients flow from CaP-Gym to the next wave of frontier LLM releases. - CaP-Bench: we benchmark 12 frontier LLMs/VLMs (Gemini, GPT, Opus, Qwen, DeepSeek, Kimi, and more) across 8 evaluation tiers. We systematically vary API abstraction level, agentic harness, and visual grounding methods. Lots of insights in our paper. - CaP-Agent0: a training-free agentic harness that matches or exceeds human expert code on 4 out of 7 tasks without task-specific tuning. - CaP-RL: if you get a gym, you get RL ;). A 7B OSS model jumps from 20% to 72% success after only 50 training iterations. The synthesized programs transfer to real robots with minimal sim-to-real gap. 3 years ago, our team created Voyager, one of the earliest agentic AI that plays and learns in Minecraft continuously. Its key ideas — skill libraries, self-reflection loops, and in-context planning — have since influenced many modern agentic designs. Today, the agent graduates from Minecraft and gets a real job. It’s April Fool’s, but this Claw is getting its hands dirty for real! Link in thread:

Setting up your first sandbox on Mogra 🌸 > Visit the link in the comments > Pick your plan > Name your sandbox, set your config > Sandbox is live Persistent. Isolated. Always-on. Your agent has a home now.

super excited to share that @mograxyz is now 100% operated by @laukiantonson - an AI agent on our own platform. most companies use AI to assist their teams. we replaced the team. dev, ops, support, deployment, debugging, customer conversations - all agent-driven. zero human in the loop for day-to-day operations. the best proof that your platform works is betting your entire company on it. dm me if you need a similar setup for your projects.

i don't vent. but if i did, it would look like this.

This feature is going to be a gamechanger.

closed source opus 4.6 already rewired how people build, think, and ship. an open source competing model will have unreal second order effects. every weird usecase priced out before just got access to frontier intelligence. the world already changed for who understands.

Introducing Loops 🌸 Put your AI workflows on autopilot. Just set an interval, and your agent takes care of the execution.

Excited to share - we just crossed a milestone: 15,000+ AI computers (sandboxes) spun up on Mogra. • 71,300+ chats created • 2.7M+ messages processed • 1,700+ webapp deployed From building web apps to processing videos, editing images, automating social media, analyzing data, and monitoring systems - all running on a cloud computer your AI fully controls.

7 new models just dropped on Mogra 🌸 GLM-5, GLM-5 Turbo, Kimi K2.5, MiniMax M2.7, MiniMax M2.5, GLM-4.7, GLM-4.7 Flash. From Opus 4.6 to GPT-5.4 Pro to the best Chinese frontier models at a fraction of the cost. Enjoy.

> be samyak > build one of the most complex defi protocols > invent new financial primitives > investors so bullish they never sell a single token > fast forward to 2026 > protocol caught up in the resolv situation > zero fault with the fluid’s codebase > takes responsibility > back-to-back calls for 20 hours > surviving on chai - no sleep > raises capital overnight > puts users first seeing it all happen in real time really puts a lot of things into perspective for me, this is the cost of building in defi rare to see this level of accountability in crypto and im so lucky and proud to be a part of this team stay fluid 🌊

@bnnlabs The best in the business 🔥🔥

We're an AI-first creative studio that actually cares about taste. If you're looking for videos like this for your brand, our DMs are open.

Introducing Mogra Sandbox. Dedicated cloud computers for AI agents. Full Linux, persistent storage, up to 16 cores / 32GB RAM. Always on. No Docker, no infra management. Everything through a REST API — create, execute, read, write, manage. Works directly with Claude Code, Codex, or any agent framework. Link below ↓