Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

INSID3 segments objects across domains using ONLY ONE annotated example it works entirely without a segmentation decoder, task-specific fine-tuning, or external mask generators like SAM CVPR 2026 paper with enormous practical potential

SkalskiP

49,785 subscribers

55,422 views • 3 months ago •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Robotics keeps hitting the same wall. Single task RL works, but... it does not scale to hundreds of tasks or new embodiments. This new paper looks like a real step toward fixing that. The team introduces MMBench, a benchmark with 200 tasks across many domains and robots, and Newt, a language conditioned world model trained online across all 200 tasks at once. The simple idea behind Newt: The model learns from demos to get the right priors It trains across many tasks through online interaction It uses language to ground the goal It adapts fast when a new task shows up What stood out to me: ✅ One model trained on 200 tasks at the same time ✅ Language conditioned control for both states and RGB ✅ Better data efficiency than strong baselines ✅ Strong open loop control ✅ Fast adaptation to new tasks and embodiments ✅ Full release of 200 checkpoints, 4000 demos, code, and benchmark This is a good push toward general control instead of one model per task. If you want the full paper: Project page: —- Weekly robotics and AI insights. Subscribe free:

Robotics keeps hitting the same wall. Single task RL works, but... it does not scale to hundreds of tasks or new embodiments. This new paper looks like a real step toward fixing that. The team introduces MMBench, a benchmark with 200 tasks across many domains and robots, and Newt, a language conditioned world model trained online across all 200 tasks at once. The simple idea behind Newt: The model learns from demos to get the right priors It trains across many tasks through online interaction It uses language to ground the goal It adapts fast when a new task shows up What stood out to me: ✅ One model trained on 200 tasks at the same time ✅ Language conditioned control for both states and RGB ✅ Better data efficiency than strong baselines ✅ Strong open loop control ✅ Fast adaptation to new tasks and embodiments ✅ Full release of 200 checkpoints, 4000 demos, code, and benchmark This is a good push toward general control instead of one model per task. If you want the full paper: Project page: —- Weekly robotics and AI insights. Subscribe free:

Ilir Aliu

70,090 views • 9 months ago

We’re excited to introduce Text-to-LoRA: a Hypernetwork that generates task-specific LLM adapters (LoRAs) based on a text description of the task. Catch our presentation at #ICML2025! Paper: Code: Biological systems are capable of rapid adaptation, given limited sensory cues. For example, our human visual system can quickly adapt and tune its light sensitivity to our surroundings. While modern LLMs exhibit a wide variety of capabilities and knowledge, they remain rigid when adding task-specific capabilities. Traditionally, customizing these models requires gathering large datasets and performing often expensive, time-consuming fine-tuning for specific applications. To bypass these limitations, Text-to-LoRA (T2L) meta-learns a “hypernetwork” that takes in a text description of a desired task, as a prompt, and generates a task-specific LoRA that performs well on the task. In our experiments, we show that T2L can encode hundreds of existing LoRA adapters. While the compression is lossy, T2L maintains the performance of task-specifically tuned LoRA adapters. We also show that T2L can even generalize to unseen tasks given a natural language description of the tasks. Importantly, Text-to-LoRA is parameter-efficient. It generates LoRAs in a single, inexpensive step, based solely on a simple text description of the task. This approach is a step towards dramatically lowering the technical and computational barriers, allowing non-technical users to specialize foundation models using plain language, rather than needing deep technical expertise or large compute resources.

We’re excited to introduce Text-to-LoRA: a Hypernetwork that generates task-specific LLM adapters (LoRAs) based on a text description of the task. Catch our presentation at #ICML2025! Paper: Code: Biological systems are capable of rapid adaptation, given limited sensory cues. For example, our human visual system can quickly adapt and tune its light sensitivity to our surroundings. While modern LLMs exhibit a wide variety of capabilities and knowledge, they remain rigid when adding task-specific capabilities. Traditionally, customizing these models requires gathering large datasets and performing often expensive, time-consuming fine-tuning for specific applications. To bypass these limitations, Text-to-LoRA (T2L) meta-learns a “hypernetwork” that takes in a text description of a desired task, as a prompt, and generates a task-specific LoRA that performs well on the task. In our experiments, we show that T2L can encode hundreds of existing LoRA adapters. While the compression is lossy, T2L maintains the performance of task-specifically tuned LoRA adapters. We also show that T2L can even generalize to unseen tasks given a natural language description of the tasks. Importantly, Text-to-LoRA is parameter-efficient. It generates LoRAs in a single, inexpensive step, based solely on a simple text description of the task. This approach is a step towards dramatically lowering the technical and computational barriers, allowing non-technical users to specialize foundation models using plain language, rather than needing deep technical expertise or large compute resources.

Sakana AI

403,354 views • 1 year ago

AI in robotics gets all the attention right now, but sometimes the most interesting work is very practical. Viet built a small vision system that counts potatoes on a conveyor belt. No giant dataset. No huge model. Just a clear problem and a smart setup. He used Ultralytics’ ObjectCounter, trained a tiny YOLO11 nano model, and because there was no potato dataset, he annotated a single frame with SAM 2 and trained from that. One frame. Still works across the whole video. It is a good reminder that useful AI in industry often looks like this. Focused. Lightweight. Solves a real task. If you work in manufacturing or robotics, these small systems are usually the fastest wins. They save time, reduce errors, and do not need massive infrastructure. Nice work, Viet. His projects: —- Weekly robotics and AI insights. Subscribe free:

AI in robotics gets all the attention right now, but sometimes the most interesting work is very practical. Viet built a small vision system that counts potatoes on a conveyor belt. No giant dataset. No huge model. Just a clear problem and a smart setup. He used Ultralytics’ ObjectCounter, trained a tiny YOLO11 nano model, and because there was no potato dataset, he annotated a single frame with SAM 2 and trained from that. One frame. Still works across the whole video. It is a good reminder that useful AI in industry often looks like this. Focused. Lightweight. Solves a real task. If you work in manufacturing or robotics, these small systems are usually the fastest wins. They save time, reduce errors, and do not need massive infrastructure. Nice work, Viet. His projects: —- Weekly robotics and AI insights. Subscribe free:

Ilir Aliu

1,676,995 views • 9 months ago

🚀 The Segment Anything Model (SAM) has been upgraded to SAM2, featuring an efficient image encoder for segmenting images and videos. But does SAM2 outperform SAM1 in medical image and video segmentation? We're thrilled to present our paper "Segment Anything in Medical Images and Videos: Benchmark and Deployment"! We comprehensively benchmark SAM2 across 11 medical image modalities and videos. 📄 Paper: 💻 Code: **Highlights:** 1. SAM2 doesn’t always outperform SAM1 in 2D medical images, but excels in video segmentation, making it more accurate and efficient for 3D images, such as CT and MR scans. 2. MedSAM still outperforms SAM2 on most 2D modalities, but SAM2 surpasses MedSAM for 3D image segmentation in a slice-by-slice approach. 3. Segmentation performance varies with model size; sometimes the smallest model outperforms larger ones. 4. Fine-tuning SAM2 significantly boosts its performance for medical image segmentation. While SAM2 may struggle with challenging objects that have unclear boundaries or low contrast, it excels in generating good initial segmentation masks for common medical images and videos. However, the official interface doesn’t support medical data formats and has limitations on video length. To address this, we've developed a 3D Slicer Plugin and Gradio API for efficient 3D medical image and video segmentation. We invite you to try them out and provide feedback! 🔧 Deployment: - 3D Slicer Plugin: - Gradio API: (Note: Due to GPU limitations, the online API is available for only 12 hours and may be slow. We highly recommend deploying the Gradio API with your own computing resources: A big shoutout to Jun Ma (JunMa) who recently joined our UHN AI hub (UHN AI Hub) as Machine Learning Lead, and kudos to all co-authors: Sumin Kim, Feifei Li, Mohammed Baharoon (Mohammed Baharoon), Reza Asakereh, and Hongwei Lyu! This is true teamwork! Looking forward to collaborating with the community to advance 3D medical image and video segmentation foundation models! University Health Network U of T Department of Computer Science Department of Laboratory Medicine & Pathobiology Temerty Centre for AI in Medicine (T-CAIREM) Vector Institute #MedTech #AIinHealthcare #DeepLearning #MedicalImaging #SAM2 #MedSAM #AIResearch

🚀 The Segment Anything Model (SAM) has been upgraded to SAM2, featuring an efficient image encoder for segmenting images and videos. But does SAM2 outperform SAM1 in medical image and video segmentation? We're thrilled to present our paper "Segment Anything in Medical Images and Videos: Benchmark and Deployment"! We comprehensively benchmark SAM2 across 11 medical image modalities and videos. 📄 Paper: 💻 Code: Highlights: 1. SAM2 doesn’t always outperform SAM1 in 2D medical images, but excels in video segmentation, making it more accurate and efficient for 3D images, such as CT and MR scans. 2. MedSAM still outperforms SAM2 on most 2D modalities, but SAM2 surpasses MedSAM for 3D image segmentation in a slice-by-slice approach. 3. Segmentation performance varies with model size; sometimes the smallest model outperforms larger ones. 4. Fine-tuning SAM2 significantly boosts its performance for medical image segmentation. While SAM2 may struggle with challenging objects that have unclear boundaries or low contrast, it excels in generating good initial segmentation masks for common medical images and videos. However, the official interface doesn’t support medical data formats and has limitations on video length. To address this, we've developed a 3D Slicer Plugin and Gradio API for efficient 3D medical image and video segmentation. We invite you to try them out and provide feedback! 🔧 Deployment: - 3D Slicer Plugin: - Gradio API: (Note: Due to GPU limitations, the online API is available for only 12 hours and may be slow. We highly recommend deploying the Gradio API with your own computing resources: A big shoutout to Jun Ma (JunMa) who recently joined our UHN AI hub (UHN AI Hub) as Machine Learning Lead, and kudos to all co-authors: Sumin Kim, Feifei Li, Mohammed Baharoon (Mohammed Baharoon), Reza Asakereh, and Hongwei Lyu! This is true teamwork! Looking forward to collaborating with the community to advance 3D medical image and video segmentation foundation models! University Health Network U of T Department of Computer Science Department of Laboratory Medicine & Pathobiology Temerty Centre for AI in Medicine (T-CAIREM) Vector Institute #MedTech #AIinHealthcare #DeepLearning #MedicalImaging #SAM2 #MedSAM #AIResearch

Bo Wang

178,579 views • 2 years ago

Full Fine-tuning vs. Freezing Layers. Interact 👉 and == Full Fine-tuning == A real network has many — three layers in this example, billions of parameters in a production model. What does fine-tuning look like when you update all of them? That’s full fine-tuning: continue training every weight in the pretrained network on your new task. Every layer’s W gets its own ΔW. Nothing is frozen — every parameter is in play. Think of an MLP as a chain of prerequisites leading to an advanced course. Layer 1 might be Linear Algebra, layer 2 Probability, layer 3 Advanced Machine Learning — each one building on what came before. Fine-tuning is what happens during graduate study: the foundations are already there from undergrad, so you’re not re-learning. Full fine-tuning is reviewing every prerequisite to see what new topics have appeared and what discoveries the field has made since the last time you sat through them. Effective — but exhausting. This diagram shows the same three-layer MLP twice, side by side. On the left, the pretrained network runs on input X: three weight matrices W₁, W₂, W₃, each followed by a ReLU activation. Full fine-tuning gives the model the most freedom to specialize. Every parameter can move — and every parameter that can move must be stored. But not every prerequisite needs revisiting. The further you go back in the chain, the less the material has changed since pretraining — the linear-algebra basics under your computer-vision course are largely the same as they ever were. The next page does exactly that: freeze the prerequisites that haven’t moved, and only refresh the advanced one closest to your specialization. == Freezing Layers == Full fine-tuning reviewed every prerequisite — Linear Algebra, Probability, Advanced ML — to refresh each subject with the latest topics. Effective, but exhausting. Then you realize something. The prerequisites haven’t actually changed that much. Linear Algebra is still Linear Algebra; the matrix decompositions you learned still hold. Probability is still Probability; the distributions and Bayes’ rule haven’t moved. Almost all the new material — the new ideas, the recent discoveries — lives in the advanced layer at the top. That’s freezing layers: keep the prerequisite layers fixed at their pretrained state, and only update the advanced one. In the diagram below, W1 and W2 — the foundational prerequisites — stay frozen. Only W3 — the layer closest to your task-specific output — gets a ΔW.

Full Fine-tuning vs. Freezing Layers. Interact 👉 and == Full Fine-tuning == A real network has many — three layers in this example, billions of parameters in a production model. What does fine-tuning look like when you update all of them? That’s full fine-tuning: continue training every weight in the pretrained network on your new task. Every layer’s W gets its own ΔW. Nothing is frozen — every parameter is in play. Think of an MLP as a chain of prerequisites leading to an advanced course. Layer 1 might be Linear Algebra, layer 2 Probability, layer 3 Advanced Machine Learning — each one building on what came before. Fine-tuning is what happens during graduate study: the foundations are already there from undergrad, so you’re not re-learning. Full fine-tuning is reviewing every prerequisite to see what new topics have appeared and what discoveries the field has made since the last time you sat through them. Effective — but exhausting. This diagram shows the same three-layer MLP twice, side by side. On the left, the pretrained network runs on input X: three weight matrices W₁, W₂, W₃, each followed by a ReLU activation. Full fine-tuning gives the model the most freedom to specialize. Every parameter can move — and every parameter that can move must be stored. But not every prerequisite needs revisiting. The further you go back in the chain, the less the material has changed since pretraining — the linear-algebra basics under your computer-vision course are largely the same as they ever were. The next page does exactly that: freeze the prerequisites that haven’t moved, and only refresh the advanced one closest to your specialization. == Freezing Layers == Full fine-tuning reviewed every prerequisite — Linear Algebra, Probability, Advanced ML — to refresh each subject with the latest topics. Effective, but exhausting. Then you realize something. The prerequisites haven’t actually changed that much. Linear Algebra is still Linear Algebra; the matrix decompositions you learned still hold. Probability is still Probability; the distributions and Bayes’ rule haven’t moved. Almost all the new material — the new ideas, the recent discoveries — lives in the advanced layer at the top. That’s freezing layers: keep the prerequisite layers fixed at their pretrained state, and only update the advanced one. In the diagram below, W1 and W2 — the foundational prerequisites — stay frozen. Only W3 — the layer closest to your task-specific output — gets a ΔW.

Tom Yeh

27,740 views • 4 months ago

Boom! Grok Tasks Make It One Of The Most POWERFUL Real-Time AI Systems In The World. — My How to Use Grok Tasks With Hidden Tools For Powerful Daily Output. Grok Tasks are customizable AI workflows that integrate a variety of tools to streamline daily activities, from research and analysis to creative planning and problem-solving. I have been using them for quite sometime and because of the vital heartbeat of news and first person data on X, it is the most powerful AI platform available. By combining Tasks with tools like web searches, X platform interactions, code execution, and media viewers, you can build efficient, automated processes. These tasks work by prompting Grok with a clear description of what you want to achieve, and Grok will intelligently call the necessary tools in sequence or parallel to deliver results. Here's a step-by-step guide to creating and using Grok Tasks: Step 1: Define Your Task Start by clearly outlining the daily activity or goal. Consider what inputs you have (e.g., a URL, a query, or an attachment) and what output you need (e.g., a summary, calculation, or visual analysis). Break it down into subtasks to identify tool needs. For example, if your task involves researching current events, note that you'll need search and browsing capabilities. Step 2: Review Available Tools Familiarize yourself with the tools Grok can access. Here's a quick overview: - Code Execution: Run Python code for calculations, data processing, or simulations using libraries like numpy, pandas, or sympy. - Browse Page: Fetch and summarize content from any website URL with custom instructions. - Web Search: Perform general internet searches, returning results with optional operators like site:. - Web Search With Snippets: Get quick, detailed excerpts from search results for fact-checking. - X Keyword Search: Advanced search for X posts using operators like from:, since:, or filter:. - X Semantic Search: Find semantically related X posts based on a query, with filters for dates or users. - X User Search: Locate X users by name or handle. - X Thread Fetch: Retrieve a full X post thread, including context like replies and parents. - View Image: Analyze an image from a URL or conversation ID. - View X Video: Extract frames and subtitles from an X-hosted video. - Search PDF Attachment: Query a PDF file for relevant pages using keyword or regex modes. - Browse PDF Attachment: View specific pages of a PDF with text and screenshots. Select tools that align with your task. Aim for a mix to handle data gathering, processing, and visualization. Step 3: Craft Your Prompt Write a detailed prompt to Grok describing the task. Include: - The overall goal. - Specific steps or subtasks. - References to tools if you want to guide the process (e.g., "Use web_search to find sources, then code_execution to analyze data"). - Any constraints, like dates or limits. Example prompt: "Create a Grok Task for my morning routine: Search recent X posts about tech news using x_keyword_search, fetch a key thread with x_thread_fetch, and summarize with browse_page on linked articles." Step 4: Submit and Interact Send your prompt to Grok. It will process the task by calling tools as needed, often in parallel for efficiency. Review the output and refine with follow-up prompts if required (e.g., "Expand on that using view_image for visuals"). Iterate to fine-tune the workflow for reuse. Step 5: Save and Reuse Once refined, note the prompt as a template for future use. You can adapt it for similar tasks, making Grok Tasks a habitual part of your day. Finding Grok Tasks To discover existing Grok Tasks or inspiration for new ones, use X searches with tools like x_keyword_search or x_semantic_search (e.g., query: "Grok Tasks examples" with mode: Latest). Browse community-shared threads via x_thread_fetch, or web_search for tutorials on xAI features. Prompt Grok directly: "Show me popular Grok Tasks for productivity." 1 of 3

Boom! Grok Tasks Make It One Of The Most POWERFUL Real-Time AI Systems In The World. — My How to Use Grok Tasks With Hidden Tools For Powerful Daily Output. Grok Tasks are customizable AI workflows that integrate a variety of tools to streamline daily activities, from research and analysis to creative planning and problem-solving. I have been using them for quite sometime and because of the vital heartbeat of news and first person data on X, it is the most powerful AI platform available. By combining Tasks with tools like web searches, X platform interactions, code execution, and media viewers, you can build efficient, automated processes. These tasks work by prompting Grok with a clear description of what you want to achieve, and Grok will intelligently call the necessary tools in sequence or parallel to deliver results. Here's a step-by-step guide to creating and using Grok Tasks: Step 1: Define Your Task Start by clearly outlining the daily activity or goal. Consider what inputs you have (e.g., a URL, a query, or an attachment) and what output you need (e.g., a summary, calculation, or visual analysis). Break it down into subtasks to identify tool needs. For example, if your task involves researching current events, note that you'll need search and browsing capabilities. Step 2: Review Available Tools Familiarize yourself with the tools Grok can access. Here's a quick overview: - Code Execution: Run Python code for calculations, data processing, or simulations using libraries like numpy, pandas, or sympy. - Browse Page: Fetch and summarize content from any website URL with custom instructions. - Web Search: Perform general internet searches, returning results with optional operators like site:. - Web Search With Snippets: Get quick, detailed excerpts from search results for fact-checking. - X Keyword Search: Advanced search for X posts using operators like from:, since:, or filter:. - X Semantic Search: Find semantically related X posts based on a query, with filters for dates or users. - X User Search: Locate X users by name or handle. - X Thread Fetch: Retrieve a full X post thread, including context like replies and parents. - View Image: Analyze an image from a URL or conversation ID. - View X Video: Extract frames and subtitles from an X-hosted video. - Search PDF Attachment: Query a PDF file for relevant pages using keyword or regex modes. - Browse PDF Attachment: View specific pages of a PDF with text and screenshots. Select tools that align with your task. Aim for a mix to handle data gathering, processing, and visualization. Step 3: Craft Your Prompt Write a detailed prompt to Grok describing the task. Include: - The overall goal. - Specific steps or subtasks. - References to tools if you want to guide the process (e.g., "Use web_search to find sources, then code_execution to analyze data"). - Any constraints, like dates or limits. Example prompt: "Create a Grok Task for my morning routine: Search recent X posts about tech news using x_keyword_search, fetch a key thread with x_thread_fetch, and summarize with browse_page on linked articles." Step 4: Submit and Interact Send your prompt to Grok. It will process the task by calling tools as needed, often in parallel for efficiency. Review the output and refine with follow-up prompts if required (e.g., "Expand on that using view_image for visuals"). Iterate to fine-tune the workflow for reuse. Step 5: Save and Reuse Once refined, note the prompt as a template for future use. You can adapt it for similar tasks, making Grok Tasks a habitual part of your day. Finding Grok Tasks To discover existing Grok Tasks or inspiration for new ones, use X searches with tools like x_keyword_search or x_semantic_search (e.g., query: "Grok Tasks examples" with mode: Latest). Browse community-shared threads via x_thread_fetch, or web_search for tutorials on xAI features. Prompt Grok directly: "Show me popular Grok Tasks for productivity." 1 of 3

Brian Roemmele

152,242 views • 8 months ago

Right now, you may not have access to models like GPT‑5.6 Sol, GPT‑4.6 Terra, GPT‑5.6 Luna, Claude Mythos 5, or Claude Fable 5. But you can run something surprisingly powerful today, locally, and completely free. in the next 10 mins on your 8 GB VRAM gaming laptop. Gemma 4 26B A4B QAT (MoE) delivers strong performance on a standard 8 GB VRAM GPU using Ollama, with no API, no usage limits, and no external dependencies. Out of the box, it reaches around 20 tokens per second without any optimizations. Only one command in your terminal: Ollama run gemma4:26b This means: Full offline capability (privacy by default) Zero recurring cost Competitive performance for many real world tasks Fast enough for interactive use on cheap consumer hardware If you're waiting for cutting edge cloud models, you're missing what is already practical today: a capable, local LLM that runs entirely on your own machine.

Right now, you may not have access to models like GPT‑5.6 Sol, GPT‑4.6 Terra, GPT‑5.6 Luna, Claude Mythos 5, or Claude Fable 5. But you can run something surprisingly powerful today, locally, and completely free. in the next 10 mins on your 8 GB VRAM gaming laptop. Gemma 4 26B A4B QAT (MoE) delivers strong performance on a standard 8 GB VRAM GPU using Ollama, with no API, no usage limits, and no external dependencies. Out of the box, it reaches around 20 tokens per second without any optimizations. Only one command in your terminal: Ollama run gemma4:26b This means: Full offline capability (privacy by default) Zero recurring cost Competitive performance for many real world tasks Fast enough for interactive use on cheap consumer hardware If you're waiting for cutting edge cloud models, you're missing what is already practical today: a capable, local LLM that runs entirely on your own machine.

Alok

65,387 views • 2 months ago

𝗗𝗼𝗻'𝘁 𝗳𝗶𝗻𝗲-𝘁𝘂𝗻𝗲 𝗿𝗼𝗯𝗼𝘁 𝗳𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻 𝗺𝗼𝗱𝗲𝗹𝘀. 𝗦𝘁𝗲𝗲𝗿 𝘁𝗵𝗲𝗺 𝘄𝗶𝘁𝗵 𝗵𝘂𝗺𝗮𝗻 𝗰𝗼𝗿𝗿𝗲𝗰𝘁𝗶𝗼𝗻𝘀 𝗶𝗻𝘀𝘁𝗲𝗮𝗱, 𝘄𝗶𝘁𝗵𝗼𝘂𝘁 𝗰𝗵𝗮𝗻𝗴𝗶𝗻𝗴 𝘁𝗵𝗲 𝗯𝗮𝘀𝗲 𝗽𝗼𝗹𝗶𝗰𝘆 Modern VLAs and world-action models can perform impressive manipulation skills, but adapting them reliably to new robots and tasks remains challenging. A natural solution is DAgger-style online imitation learning: deploy the robot, collect human corrections, and update the policy. Yet foundation models are fragile in the low-data regime, fine-tuning on a handful of interventions can improve one behavior while degrading others. Online post-training or reinforcement learning can require costly data collection and exploration, making real-world learning expensive and potentially unsafe. In our new paper, 𝗙𝗹𝗼𝘄𝗗𝗔𝗴𝗴𝗲𝗿, we take a different approach: 𝗜𝗻𝘀𝘁𝗲𝗮𝗱 𝗼𝗳 𝗰𝗵𝗮𝗻𝗴𝗶𝗻𝗴 𝘁𝗵𝗲 𝗳𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻 𝗺𝗼𝗱𝗲𝗹, 𝘄𝗲 𝗹𝗲𝗮𝗿𝗻 𝗵𝗼𝘄 𝘁𝗼 𝘀𝘁𝗲𝗲𝗿 𝗶𝘁 𝗳𝗿𝗼𝗺 𝗵𝘂𝗺𝗮𝗻 𝗰𝗼𝗿𝗿𝗲𝗰𝘁𝗶𝗼𝗻𝘀. The key idea is 𝗮𝗰𝘁𝗶𝗼𝗻 𝗶𝗻𝘃𝗲𝗿𝘀𝗶𝗼𝗻: we map human corrective actions back into the latent noise space of the frozen generative policy. These latent targets train a lightweight controller that adapts the robot while preserving the original model's capabilities. Across simulation and real robots, FlowDAgger: 📈 Learns from only 5–20 human intervention episodes 🏆 Outperforms supervised fine-tuning and latent-space reinforcement learning 🤖 Works across VLAs, diffusion policies, and world-action models ✔️ Provides reliable improvements without modifying the pretrained policy We believe this offers a practical path toward making robot foundation models improve during deployment, learning from the way humans naturally teach: through corrections. 📄 Paper: 🌐 Project: 💻 Code: This project was led by my amazing colleague Michael Murray with help from Daphne Chen, Simran Bagaria, Dean Fortier, Tess Hellebrekers, Harshavardhan Reddy Gajarla, Galen Mullins and Andrey Kolobov at Microsoft Research and Maya Cakmak at University of Washington

𝗗𝗼𝗻'𝘁 𝗳𝗶𝗻𝗲-𝘁𝘂𝗻𝗲 𝗿𝗼𝗯𝗼𝘁 𝗳𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻 𝗺𝗼𝗱𝗲𝗹𝘀. 𝗦𝘁𝗲𝗲𝗿 𝘁𝗵𝗲𝗺 𝘄𝗶𝘁𝗵 𝗵𝘂𝗺𝗮𝗻 𝗰𝗼𝗿𝗿𝗲𝗰𝘁𝗶𝗼𝗻𝘀 𝗶𝗻𝘀𝘁𝗲𝗮𝗱, 𝘄𝗶𝘁𝗵𝗼𝘂𝘁 𝗰𝗵𝗮𝗻𝗴𝗶𝗻𝗴 𝘁𝗵𝗲 𝗯𝗮𝘀𝗲 𝗽𝗼𝗹𝗶𝗰𝘆 Modern VLAs and world-action models can perform impressive manipulation skills, but adapting them reliably to new robots and tasks remains challenging. A natural solution is DAgger-style online imitation learning: deploy the robot, collect human corrections, and update the policy. Yet foundation models are fragile in the low-data regime, fine-tuning on a handful of interventions can improve one behavior while degrading others. Online post-training or reinforcement learning can require costly data collection and exploration, making real-world learning expensive and potentially unsafe. In our new paper, 𝗙𝗹𝗼𝘄𝗗𝗔𝗴𝗴𝗲𝗿, we take a different approach: 𝗜𝗻𝘀𝘁𝗲𝗮𝗱 𝗼𝗳 𝗰𝗵𝗮𝗻𝗴𝗶𝗻𝗴 𝘁𝗵𝗲 𝗳𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻 𝗺𝗼𝗱𝗲𝗹, 𝘄𝗲 𝗹𝗲𝗮𝗿𝗻 𝗵𝗼𝘄 𝘁𝗼 𝘀𝘁𝗲𝗲𝗿 𝗶𝘁 𝗳𝗿𝗼𝗺 𝗵𝘂𝗺𝗮𝗻 𝗰𝗼𝗿𝗿𝗲𝗰𝘁𝗶𝗼𝗻𝘀. The key idea is 𝗮𝗰𝘁𝗶𝗼𝗻 𝗶𝗻𝘃𝗲𝗿𝘀𝗶𝗼𝗻: we map human corrective actions back into the latent noise space of the frozen generative policy. These latent targets train a lightweight controller that adapts the robot while preserving the original model's capabilities. Across simulation and real robots, FlowDAgger: 📈 Learns from only 5–20 human intervention episodes 🏆 Outperforms supervised fine-tuning and latent-space reinforcement learning 🤖 Works across VLAs, diffusion policies, and world-action models ✔️ Provides reliable improvements without modifying the pretrained policy We believe this offers a practical path toward making robot foundation models improve during deployment, learning from the way humans naturally teach: through corrections. 📄 Paper: 🌐 Project: 💻 Code: This project was led by my amazing colleague Michael Murray with help from Daphne Chen, Simran Bagaria, Dean Fortier, Tess Hellebrekers, Harshavardhan Reddy Gajarla, Galen Mullins and Andrey Kolobov at Microsoft Research and Maya Cakmak at University of Washington

Oier Mees

13,276 views • 2 months ago

🚨 JUST IN: CHINA just released an AI EMPLOYEE that works 24X7 on its own. 100% OPEN SOURCE. It researches, codes, builds websites, creates slide decks, and generates videos. All by itself. All on your computer. It's called DeerFlow. You give it a task. It makes a plan, spins up its own team of sub-agents, and gets to work. You come back and there's a finished deliverable waiting. Not a draft. Not a summary. The actual thing. Not a chatbot. Not a research assistant. An AI with its own computer that works while you sleep. Here's what it does on its own: → Spawns multiple sub-agents in parallel, each tackling a different piece of your task, then combines everything into one finished output → Writes real code, runs it, reads the results, and fixes its own mistakes without asking you once → Builds slide decks, websites, full research reports, and data dashboards from scratch → Remembers you across sessions. Your writing style. Your tech stack. Your preferences. Gets better every time. → Reads files you upload, works with them inside its own filesystem, hands you clean finished outputs → Searches the web, runs commands, calls any tool you plug in Here's how it thinks: You give one instruction. The lead agent makes a plan. Sub-agents fan out and work in parallel. Results come back. Everything gets synthesized. You get a deliverable. A single research task might split into a dozen sub-agents, each exploring a different angle, then converge into one finished website with generated visuals. Here's the wildest part: DeerFlow 2.0 launched on February 28th 2026 and hit number 1 on all of GitHub Trending the same day. Version 2.0 was a complete rewrite. Zero shared code with version 1. Because users kept using it for things the team never intended. Data pipelines. Dashboards. Entire content workflows. The community told them what it needed to become. So they burned it down and rebuilt it. 22.7K GitHub stars. 2.7K forks. Built by ByteDance 100% Open Source. MIT License.

🚨 JUST IN: CHINA just released an AI EMPLOYEE that works 24X7 on its own. 100% OPEN SOURCE. It researches, codes, builds websites, creates slide decks, and generates videos. All by itself. All on your computer. It's called DeerFlow. You give it a task. It makes a plan, spins up its own team of sub-agents, and gets to work. You come back and there's a finished deliverable waiting. Not a draft. Not a summary. The actual thing. Not a chatbot. Not a research assistant. An AI with its own computer that works while you sleep. Here's what it does on its own: → Spawns multiple sub-agents in parallel, each tackling a different piece of your task, then combines everything into one finished output → Writes real code, runs it, reads the results, and fixes its own mistakes without asking you once → Builds slide decks, websites, full research reports, and data dashboards from scratch → Remembers you across sessions. Your writing style. Your tech stack. Your preferences. Gets better every time. → Reads files you upload, works with them inside its own filesystem, hands you clean finished outputs → Searches the web, runs commands, calls any tool you plug in Here's how it thinks: You give one instruction. The lead agent makes a plan. Sub-agents fan out and work in parallel. Results come back. Everything gets synthesized. You get a deliverable. A single research task might split into a dozen sub-agents, each exploring a different angle, then converge into one finished website with generated visuals. Here's the wildest part: DeerFlow 2.0 launched on February 28th 2026 and hit number 1 on all of GitHub Trending the same day. Version 2.0 was a complete rewrite. Zero shared code with version 1. Because users kept using it for things the team never intended. Data pipelines. Dashboards. Entire content workflows. The community told them what it needed to become. So they burned it down and rebuilt it. 22.7K GitHub stars. 2.7K forks. Built by ByteDance 100% Open Source. MIT License.

Kanika

738,832 views • 5 months ago

Sitting here watching my Cybertruck charge with the mountains of Utah in the background, it hit me again… I really believe this is the most beautiful Tesla ever made. Every angle is different. It looks futuristic without trying too hard, and it fits just as well in the middle of a city as it does surrounded by nature like this. What makes this truck special is so much more than just the stainless steel or design. For me it’s how capable and practical it is. Long road trips, hauling gear, camping, tough weather, and FSD… it feels like THE one vehicle built to do almost everything. I’ve owned and driven many Teslas over the years, and every one has been amazing in their own unique way, but the Cybertruck is the first one that genuinely makes me stop and look back every time I park it. Driving across the U.S. in this beast has only reinforced that feeling. And to me, the Cybertruck is the best vehicle Tesla has ever built.

Sitting here watching my Cybertruck charge with the mountains of Utah in the background, it hit me again… I really believe this is the most beautiful Tesla ever made. Every angle is different. It looks futuristic without trying too hard, and it fits just as well in the middle of a city as it does surrounded by nature like this. What makes this truck special is so much more than just the stainless steel or design. For me it’s how capable and practical it is. Long road trips, hauling gear, camping, tough weather, and FSD… it feels like THE one vehicle built to do almost everything. I’ve owned and driven many Teslas over the years, and every one has been amazing in their own unique way, but the Cybertruck is the first one that genuinely makes me stop and look back every time I park it. Driving across the U.S. in this beast has only reinforced that feeling. And to me, the Cybertruck is the best vehicle Tesla has ever built.

Teslaconomics

27,057 views • 2 months ago

You might have seen a lot of thai fans saying that #PerthSanta spend a lot of time online because they always know the latest memes and it’s sooo true. Examples of what they’ve been using lately: - “You scared me, give me a hug” wasn’t actually created by Ta, this phrase has been going around on TikTok since early July but he couldn’t have timed it more perfectly 😂 You can find a lot of videos if you search ตกใจหมดมากอดหน่อย or similar - “WHATS” easy to miss this one but it went viral after a thai driver got into a car accident and tried to pretend to speak english only just to avoid getting into trouble. She spoke with such a thick accent that instead of saying “what” (the easiest word to pronounce) she said something that sounded like “whats” and people have been using it as a joke ever since. 2nd clip is PS saying it, 3rd clip is the original of the girl - “dtdraai” or “dress” people have been using these words instead of saying “oh I’m d€ad” and you’ll see it often in thai posts too. I’ve found a good video explaining it in more detail (4th clip) If this ever comes across our thai little devils I’d love to know if there are more! It’s pretty hard to keep up without actually living in thailand but it’s so much fun and we’d love to laugh about it with you 🙇🏻‍♀️

You might have seen a lot of thai fans saying that #PerthSanta spend a lot of time online because they always know the latest memes and it’s sooo true. Examples of what they’ve been using lately: - “You scared me, give me a hug” wasn’t actually created by Ta, this phrase has been going around on TikTok since early July but he couldn’t have timed it more perfectly 😂 You can find a lot of videos if you search ตกใจหมดมากอดหน่อย or similar - “WHATS” easy to miss this one but it went viral after a thai driver got into a car accident and tried to pretend to speak english only just to avoid getting into trouble. She spoke with such a thick accent that instead of saying “what” (the easiest word to pronounce) she said something that sounded like “whats” and people have been using it as a joke ever since. 2nd clip is PS saying it, 3rd clip is the original of the girl - “dtdraai” or “dress” people have been using these words instead of saying “oh I’m d€ad” and you’ll see it often in thai posts too. I’ve found a good video explaining it in more detail (4th clip) If this ever comes across our thai little devils I’d love to know if there are more! It’s pretty hard to keep up without actually living in thailand but it’s so much fun and we’d love to laugh about it with you 🙇🏻‍♀️

Mia 🦋

20,001 views • 17 days ago

🚨 COULD A PASSING DWARF PLANET CAUSE A MASS EXTINCTION WITHOUT EVER HITTING EARTH? Physicist Daniele Fargion proposes that close passages of planetary-mass objects from the outer Solar System could have generated powerful, long-lasting tidal forces on Earth. These flybys may have caused: • Massive, persistent tsunamis lasting years • Sudden sea level changes and coastal flooding • Crustal deformation triggering large volcanic episodes • Climate disruption • And even the diversion of asteroids or comets toward Earth The idea is that these events could help explain several major extinctions over the past 600 million years including some where conventional causes (like the Chicxulub impact for the dinosaurs) don’t fully account for the geological record. Why this matters: Most mass extinctions show complex, multi-factor signatures. A single flyby could simultaneously produce several of the observed effects (tidal waves + volcanism + climate shift), offering a potential unifying mechanism that current explanations sometimes lack. The deeper implication: If correct, it would mean Earth’s biosphere has been periodically “reset” or heavily stressed by gravitational interactions with objects we’re only now beginning to catalog in the outer Solar System. It also offers one possible answer to the Fermi Paradox: advanced civilizations may be rare because life on habitable worlds is repeatedly interrupted by such cosmic events. This remains a hypothesis presented in conference proceedings. The correlations are intriguing, but direct geological proof of specific flybys is still lacking. Could repeated gravitational “visits” from the outer Solar System be one of the hidden drivers of Earth’s evolutionary resets? Follow for more frontier planetary science and cosmic threat hypotheses.

🚨 COULD A PASSING DWARF PLANET CAUSE A MASS EXTINCTION WITHOUT EVER HITTING EARTH? Physicist Daniele Fargion proposes that close passages of planetary-mass objects from the outer Solar System could have generated powerful, long-lasting tidal forces on Earth. These flybys may have caused: • Massive, persistent tsunamis lasting years • Sudden sea level changes and coastal flooding • Crustal deformation triggering large volcanic episodes • Climate disruption • And even the diversion of asteroids or comets toward Earth The idea is that these events could help explain several major extinctions over the past 600 million years including some where conventional causes (like the Chicxulub impact for the dinosaurs) don’t fully account for the geological record. Why this matters: Most mass extinctions show complex, multi-factor signatures. A single flyby could simultaneously produce several of the observed effects (tidal waves + volcanism + climate shift), offering a potential unifying mechanism that current explanations sometimes lack. The deeper implication: If correct, it would mean Earth’s biosphere has been periodically “reset” or heavily stressed by gravitational interactions with objects we’re only now beginning to catalog in the outer Solar System. It also offers one possible answer to the Fermi Paradox: advanced civilizations may be rare because life on habitable worlds is repeatedly interrupted by such cosmic events. This remains a hypothesis presented in conference proceedings. The correlations are intriguing, but direct geological proof of specific flybys is still lacking. Could repeated gravitational “visits” from the outer Solar System be one of the hidden drivers of Earth’s evolutionary resets? Follow for more frontier planetary science and cosmic threat hypotheses.

TheNewPhysics

42,024 views • 2 months ago

i think it’s been enough time & we can finally address this particular discourse, and i think there’s a healthy discussion we can have about it. i genuinely think a fully traditional rendition in 2026, by artists operating at a global pop level, would have changed how the project was received internationally. there’s a very real possibility it would’ve been boxed into “world music,” treated more like cultural exhibition than a modern musical statement, keeping in mind that Arirang Album is not for 'aesthetics' nor is it a history lesson. using 808s, trap production, pop structures, and english hooks while still centering Arirang was actually a significant artistic choice. historically, Arirang was the song of ordinary people. in today’s world, pop culture/pop music is the common language. using contemporary pop to carry that sentiment to millions of listeners across countries feels very aligned with the spirit of what Arirang represented in its own time. and ‘unapologetically korean’ does not have to mean “only in korean” or “frozen in tradition.” it can also mean Korean artists having full control over how their culture is presented, modernized, exported, and understood. i also think if they had leaned entirely into historical styling and traditional presentation, western media may have reduced it to a “cultural curiosity” instead of engaging with it as a current, competitive pop release. critics likely would’ve framed it as heritage performance before artistry. i am glad they chose this route for the album.

i think it’s been enough time & we can finally address this particular discourse, and i think there’s a healthy discussion we can have about it. i genuinely think a fully traditional rendition in 2026, by artists operating at a global pop level, would have changed how the project was received internationally. there’s a very real possibility it would’ve been boxed into “world music,” treated more like cultural exhibition than a modern musical statement, keeping in mind that Arirang Album is not for 'aesthetics' nor is it a history lesson. using 808s, trap production, pop structures, and english hooks while still centering Arirang was actually a significant artistic choice. historically, Arirang was the song of ordinary people. in today’s world, pop culture/pop music is the common language. using contemporary pop to carry that sentiment to millions of listeners across countries feels very aligned with the spirit of what Arirang represented in its own time. and ‘unapologetically korean’ does not have to mean “only in korean” or “frozen in tradition.” it can also mean Korean artists having full control over how their culture is presented, modernized, exported, and understood. i also think if they had leaned entirely into historical styling and traditional presentation, western media may have reduced it to a “cultural curiosity” instead of engaging with it as a current, competitive pop release. critics likely would’ve framed it as heritage performance before artistry. i am glad they chose this route for the album.

ruz 𓍯

26,131 views • 4 months ago

China's central bank has now bought gold for 19 months straight, the largest official buyer on earth. And this week, as gold broke 4,000 dollars, China's biggest banks moved to push ordinary Chinese out of leveraged gold trading, with at least one warning it will liquidate any position not closed by month-end. Both are true at once, and together they explain what this crash really is. Start with what is being banned, because the words matter. ICBC and a string of other banks are shutting down retail trading in what the Chinese themselves call paper gold, the margined, leveraged contracts where you bet on the price without ever owning a bar. Some banks lifted the margin requirement to 140 percent to choke the leverage off before closing the products outright. Physical gold, meanwhile, stays wide open. Coins, bars, savings plans, ETFs, all fine. It is only the paper, the leverage, the casino, that is being shut, the last step in a five-year retreat that the crash just finished. Officially this is about protecting small investors, and that part is real. The same kind of leverage wiped out a wave of Chinese retail in a 2020 commodity blowup. But set the ban beside what the state is doing and something larger comes into view. While its citizens are pushed out of the paper, the People's Bank of China has spent those same 19 months buying the physical metal, more than two thousand three hundred tonnes of it now, accumulating straight through a 28 percent crash that scared everyone else out. Beijing is not trading gold. It is hoarding it. That is the strategy in one frame. China looked at the two things both called gold, the paper bet and the physical bar, and made a choice no Western government would make. It is taking the metal for the state and closing the casino for everyone else. The reason sits in a single date. 2022, when Russia's reserves were frozen with a keystroke. That taught every country outside the Western system one lesson: dollars in an account can be switched off, gold in your own vault cannot. So China is building its monetary independence out of the one asset nobody can freeze, and it does not want that foundation in the hands of leveraged traders who panic-sell in a crash, or priced by a paper market it does not control. Watch this month and the two worlds split in real time. Western investors were forced out of their gold by margin calls and a rate scare. China's central bank bought that exact dip with both hands. One side treats gold as a trade. The other treats it as the floor under a currency. The West is selling paper gold and calling it a crash. China is buying physical gold and calling it a foundation. In ten years, only one of them will look like it understood what gold was for. The metal is already moving to that side.

China's central bank has now bought gold for 19 months straight, the largest official buyer on earth. And this week, as gold broke 4,000 dollars, China's biggest banks moved to push ordinary Chinese out of leveraged gold trading, with at least one warning it will liquidate any position not closed by month-end. Both are true at once, and together they explain what this crash really is. Start with what is being banned, because the words matter. ICBC and a string of other banks are shutting down retail trading in what the Chinese themselves call paper gold, the margined, leveraged contracts where you bet on the price without ever owning a bar. Some banks lifted the margin requirement to 140 percent to choke the leverage off before closing the products outright. Physical gold, meanwhile, stays wide open. Coins, bars, savings plans, ETFs, all fine. It is only the paper, the leverage, the casino, that is being shut, the last step in a five-year retreat that the crash just finished. Officially this is about protecting small investors, and that part is real. The same kind of leverage wiped out a wave of Chinese retail in a 2020 commodity blowup. But set the ban beside what the state is doing and something larger comes into view. While its citizens are pushed out of the paper, the People's Bank of China has spent those same 19 months buying the physical metal, more than two thousand three hundred tonnes of it now, accumulating straight through a 28 percent crash that scared everyone else out. Beijing is not trading gold. It is hoarding it. That is the strategy in one frame. China looked at the two things both called gold, the paper bet and the physical bar, and made a choice no Western government would make. It is taking the metal for the state and closing the casino for everyone else. The reason sits in a single date. 2022, when Russia's reserves were frozen with a keystroke. That taught every country outside the Western system one lesson: dollars in an account can be switched off, gold in your own vault cannot. So China is building its monetary independence out of the one asset nobody can freeze, and it does not want that foundation in the hands of leveraged traders who panic-sell in a crash, or priced by a paper market it does not control. Watch this month and the two worlds split in real time. Western investors were forced out of their gold by margin calls and a rate scare. China's central bank bought that exact dip with both hands. One side treats gold as a trade. The other treats it as the floor under a currency. The West is selling paper gold and calling it a crash. China is buying physical gold and calling it a foundation. In ten years, only one of them will look like it understood what gold was for. The metal is already moving to that side.

Shanaka Anslem Perera ⚡

327,303 views • 2 months ago

Claude Code + Higgsfield MCP is f*cking cracked 🤯 I built an entire DTC ad campaign inside Claude Code using the new Higgsfield MCP. One product URL → hero static, animated hero shot, 2 UGC clips with a creator wearing the product. 5 assets. One Claude conversation. 3 Higgsfield models. All inside Claude Code. Perfect for DTC brands and agencies who need full campaign packages without booking a shoot or briefing a designer. If you're spending hours every week generating statics in one tool, briefing a motion designer for the hero clip, then chasing a UGC creator for the talking-head shots — this MCP eliminates the entire pipeline: → Drop a product URL into Claude Code → Claude pulls the brand brief — voice, hero SKUs, visual style, target customer → Generates the hero static with ChatGPT Images 2.0 → Animates it into a 5-second cinematic opener with Seedance 2.0 → Generates a UGC creator with GPT Image 2 → Drops her in the product and generates 2 native UGC video clips with Seedance 2.0 No tab-switching between tools. No copy-pasting prompts between platforms. No briefing 3 different vendors for one campaign. What you get: → A complete campaign package — static, animation, UGC — from one product URL → Brand-specific outputs that pull from a real brief, not generic AI slop → Claude making creative decisions between every step (which variation wins, which creator fits the persona, which clip needs a re-spin) → A repeatable pipeline you can run for any product in your catalog Built 100% in Claude Code with the Higgsfield MCP. I recorded a full walkthrough showing exactly how this works: the MCP setup, every prompt, every model, the full campaign output. Want the full video walkthrough? > Like this post > Comment "MCP" And I'll send it over (must be following so I can DM)

Mike Futia

29,800 views • 4 months ago

This fact absolutely shatters evolution. ...And makes the intelligent design in Life certain. All the systems in Life that make things function - like everything that enables birds to fly - require a whole set of complex interconnected parts. Nothing works in isolation. Something evolutionists overlook is how the nature of all these systems are intricately interconnected. Every part relies on something else to make it work. You need the whole system, down to the DNA itself, or the function fails. This coordinattion between so many systems makes step-by-step evolution all but impossible. For instance... A wing isn’t just bones and feathers (or membrane). A wing requires a certain type of hollow bones, which the alleged ancestors of birds didn't have. And wings by themselves are useless - what reproductive benefit would wings have without the accompanying systems that lead to actual flight (the simultaneous requirements for bone pneumatization, muscle attachment sites, feather or membrane development, and the regulatory changes that control all of them)? Wings without these features would be harmful. Hollow bones without flight would be detrimental. The step-by-step path leading from non-bird to flying birds is filled with easily killed, unselectable intermediates. This is where evolutionists constantly fail - they look at modification in isolation, without considering that these systems are large, interconnected networks of functional parts all working together. And it all starts at the genetic level, from the DNA itself. DNA genes code for proteins. Regulatory information organizes when these proteins are built and how much should be made. Further regulatory systems coordinate where these proteins go after they’re built. And yet more informational networks place the proteins into the final system in coordination with other proteins. Then these protein systems are fit together to make entire organs, tissues, bones, etc. All of these systems and processes are coordinated from the DNA - which means major modification like turning a fish fin into a bird wing requires many coordinated DNA changes across the entire network. One change at a time simply cannot lead to the coordination of an entire multi-part coordinated system like this. A mutation that changes one protein would need another coordinated mutation to fit that mutated protein into the right place, which would require other proteins to be mutated to fit together into that system, which would also require yet more regulatory mutations to keep everything coordinated... Nothing works in isolation. It's all connected. Which is why it can't evolve one step at a time. One step at a time does nothing. Life requires all or nothing systems. Only intelligence can engineer all or nothing systems.

This fact absolutely shatters evolution. ...And makes the intelligent design in Life certain. All the systems in Life that make things function - like everything that enables birds to fly - require a whole set of complex interconnected parts. Nothing works in isolation. Something evolutionists overlook is how the nature of all these systems are intricately interconnected. Every part relies on something else to make it work. You need the whole system, down to the DNA itself, or the function fails. This coordinattion between so many systems makes step-by-step evolution all but impossible. For instance... A wing isn’t just bones and feathers (or membrane). A wing requires a certain type of hollow bones, which the alleged ancestors of birds didn't have. And wings by themselves are useless - what reproductive benefit would wings have without the accompanying systems that lead to actual flight (the simultaneous requirements for bone pneumatization, muscle attachment sites, feather or membrane development, and the regulatory changes that control all of them)? Wings without these features would be harmful. Hollow bones without flight would be detrimental. The step-by-step path leading from non-bird to flying birds is filled with easily killed, unselectable intermediates. This is where evolutionists constantly fail - they look at modification in isolation, without considering that these systems are large, interconnected networks of functional parts all working together. And it all starts at the genetic level, from the DNA itself. DNA genes code for proteins. Regulatory information organizes when these proteins are built and how much should be made. Further regulatory systems coordinate where these proteins go after they’re built. And yet more informational networks place the proteins into the final system in coordination with other proteins. Then these protein systems are fit together to make entire organs, tissues, bones, etc. All of these systems and processes are coordinated from the DNA - which means major modification like turning a fish fin into a bird wing requires many coordinated DNA changes across the entire network. One change at a time simply cannot lead to the coordination of an entire multi-part coordinated system like this. A mutation that changes one protein would need another coordinated mutation to fit that mutated protein into the right place, which would require other proteins to be mutated to fit together into that system, which would also require yet more regulatory mutations to keep everything coordinated... Nothing works in isolation. It's all connected. Which is why it can't evolve one step at a time. One step at a time does nothing. Life requires all or nothing systems. Only intelligence can engineer all or nothing systems.

Divinely Designed

47,292 views • 26 days ago

OpenAI's AgentKit will be so insane, build every step of agents on one platform. These visual agent builders make the whole process of iterating and launching agents far more efficient. It sits on top of the Responses API and unifies the tools that were previously scattered across SDKs and custom orchestration. It lets developers create agent workflows visually, connect data sources securely, and measure performance automatically without coding every layer by hand. The core of AgentKit is the Agent Builder, a drag-and-drop canvas where each node represents an action, guardrail, or decision branch. Developers can link these nodes into multi-agent workflows, preview results instantly, and version each setup. It supports inline evaluation so that developers can see how changes affect output before deploying. The Connector Registry is a single admin panel that manages how data and tools connect across the OpenAI ecosystem. It centralizes integrations like Google Drive, SharePoint, Dropbox, and Microsoft Teams. Large organizations can govern access and flow of data between agents securely under one global console. ChatKit provides a ready-to-use chat interface for embedding agents inside apps or websites. It manages streaming, message threads, and model reasoning displays automatically. Developers can skin the interface to match their product without writing custom front-end code. Under the hood, all these blocks use the same execution core that runs agent reasoning through OpenAI’s APIs. Workflows in Agent Builder compile down to structured instructions for the Responses API, which handles model calls, tool use, and context passing. Connector Registry handles authentication and routing for external tools, while Evals and RFT provide feedback loops that improve agents over time. This integration means developers no longer need to handle orchestration logic, model evaluation pipelines, or safety layers separately. Everything runs natively within OpenAI’s control plane with managed security, automatic versioning, and built-in testing. In short, AgentKit standardizes the entire life cycle of an AI agent—from visual design to deployment and performance tuning—inside a single unified system.

OpenAI's AgentKit will be so insane, build every step of agents on one platform. These visual agent builders make the whole process of iterating and launching agents far more efficient. It sits on top of the Responses API and unifies the tools that were previously scattered across SDKs and custom orchestration. It lets developers create agent workflows visually, connect data sources securely, and measure performance automatically without coding every layer by hand. The core of AgentKit is the Agent Builder, a drag-and-drop canvas where each node represents an action, guardrail, or decision branch. Developers can link these nodes into multi-agent workflows, preview results instantly, and version each setup. It supports inline evaluation so that developers can see how changes affect output before deploying. The Connector Registry is a single admin panel that manages how data and tools connect across the OpenAI ecosystem. It centralizes integrations like Google Drive, SharePoint, Dropbox, and Microsoft Teams. Large organizations can govern access and flow of data between agents securely under one global console. ChatKit provides a ready-to-use chat interface for embedding agents inside apps or websites. It manages streaming, message threads, and model reasoning displays automatically. Developers can skin the interface to match their product without writing custom front-end code. Under the hood, all these blocks use the same execution core that runs agent reasoning through OpenAI’s APIs. Workflows in Agent Builder compile down to structured instructions for the Responses API, which handles model calls, tool use, and context passing. Connector Registry handles authentication and routing for external tools, while Evals and RFT provide feedback loops that improve agents over time. This integration means developers no longer need to handle orchestration logic, model evaluation pipelines, or safety layers separately. Everything runs natively within OpenAI’s control plane with managed security, automatic versioning, and built-in testing. In short, AgentKit standardizes the entire life cycle of an AI agent—from visual design to deployment and performance tuning—inside a single unified system.

Rohan Paul

178,460 views • 11 months ago

THE DEPTH MAP TRICK THAT FIXED DANCE ACCURACY IN SEEDANCE 2.0 Feed the model a video of someone dancing and it tries to interpret everything- the person, the clothes, the lighting, the room, and somewhere in there, the movement. Feed it a depth map and there's nothing left to interpret but the motion. Most creators trying to transfer a dance to a character reference the source footage directly, then wonder why the choreography drifts. The problem isn't the model - it's that you handed it ten variables when you only wanted one. Here's the workflow 1. Lock the character reference in GPT Image 2 first -face, build, costume, so identity holds independently of whatever motion gets applied to it 2. Convert the source dance footage into a depth map instead of using the raw video -this strips out the original performer's appearance, clothing, and environment entirely 3. Feed the depth map as the motion reference and the character sheet as the identity reference- two separate inputs doing two separate jobs, not one input trying to do both 5. Let the depth map carry only spatial movement -the model receives body position and momentum with no competing information about who's moving or what they look like 6. Keep the character and motion inputs isolated throughout - the moment you mix appearance data into the motion reference, the model starts negotiating between two identities Why this works • Raw footage passes the model everything at once- performer, wardrobe, room, lighting -and the choreography competes with all of it for attention • A depth map is pure spatial information, so the only thing left to transfer is movement • Separating identity from motion means the character can stay locked while the dance stays accurate - normally you're trading one for the other • The accuracy gain isn't the model getting better, it's the model getting fewer decisions to make Use cases: ⁃ Dance and choreography transfer onto original characters ⁃ Motion capture-style workflows without motion capture ⁃ Any sequence where a specific movement needs to survive intact ⁃ Character showcase content built on existing performance footage The character sheet answers who's dancing. The depth map answers how - and keeping those two questions separate is the whole trick.

THE DEPTH MAP TRICK THAT FIXED DANCE ACCURACY IN SEEDANCE 2.0 Feed the model a video of someone dancing and it tries to interpret everything- the person, the clothes, the lighting, the room, and somewhere in there, the movement. Feed it a depth map and there's nothing left to interpret but the motion. Most creators trying to transfer a dance to a character reference the source footage directly, then wonder why the choreography drifts. The problem isn't the model - it's that you handed it ten variables when you only wanted one. Here's the workflow 1. Lock the character reference in GPT Image 2 first -face, build, costume, so identity holds independently of whatever motion gets applied to it 2. Convert the source dance footage into a depth map instead of using the raw video -this strips out the original performer's appearance, clothing, and environment entirely 3. Feed the depth map as the motion reference and the character sheet as the identity reference- two separate inputs doing two separate jobs, not one input trying to do both 5. Let the depth map carry only spatial movement -the model receives body position and momentum with no competing information about who's moving or what they look like 6. Keep the character and motion inputs isolated throughout - the moment you mix appearance data into the motion reference, the model starts negotiating between two identities Why this works • Raw footage passes the model everything at once- performer, wardrobe, room, lighting -and the choreography competes with all of it for attention • A depth map is pure spatial information, so the only thing left to transfer is movement • Separating identity from motion means the character can stay locked while the dance stays accurate - normally you're trading one for the other • The accuracy gain isn't the model getting better, it's the model getting fewer decisions to make Use cases: ⁃ Dance and choreography transfer onto original characters ⁃ Motion capture-style workflows without motion capture ⁃ Any sequence where a specific movement needs to survive intact ⁃ Character showcase content built on existing performance footage The character sheet answers who's dancing. The depth map answers how - and keeping those two questions separate is the whole trick.

Nexlow

85,895 views • 2 months ago

THE DEPTH MAP TRICK THAT FIXED DANCE ACCURACY IN SEEDANCE 2.0 Feed the model a video of someone dancing and it tries to interpret everything- the person, the clothes, the lighting, the room, and somewhere in there, the movement. Feed it a depth map and there's nothing left to interpret but the motion. Most creators trying to transfer a dance to a character reference the source footage directly, then wonder why the choreography drifts. The problem isn't the model - it's that you handed it ten variables when you only wanted one. Here's the workflow 1. Lock the character reference in GPT Image 2 first -face, build, costume, so identity holds independently of whatever motion gets applied to it 2. Convert the source dance footage into a depth map instead of using the raw video -this strips out the original performer's appearance, clothing, and environment entirely 3. Feed the depth map as the motion reference and the character sheet as the identity reference- two separate inputs doing two separate jobs, not one input trying to do both 5. Let the depth map carry only spatial movement -the model receives body position and momentum with no competing information about who's moving or what they look like 6. Keep the character and motion inputs isolated throughout - the moment you mix appearance data into the motion reference, the model starts negotiating between two identities Why this works • Raw footage passes the model everything at once- performer, wardrobe, room, lighting -and the choreography competes with all of it for attention • A depth map is pure spatial information, so the only thing left to transfer is movement • Separating identity from motion means the character can stay locked while the dance stays accurate - normally you're trading one for the other • The accuracy gain isn't the model getting better, it's the model getting fewer decisions to make Use cases: ⁃ Dance and choreography transfer onto original characters ⁃ Motion capture-style workflows without motion capture ⁃ Any sequence where a specific movement needs to survive intact ⁃ Character showcase content built on existing performance footage The character sheet answers who's dancing. The depth map answers how - and keeping those two questions separate is the whole trick.

THE DEPTH MAP TRICK THAT FIXED DANCE ACCURACY IN SEEDANCE 2.0 Feed the model a video of someone dancing and it tries to interpret everything- the person, the clothes, the lighting, the room, and somewhere in there, the movement. Feed it a depth map and there's nothing left to interpret but the motion. Most creators trying to transfer a dance to a character reference the source footage directly, then wonder why the choreography drifts. The problem isn't the model - it's that you handed it ten variables when you only wanted one. Here's the workflow 1. Lock the character reference in GPT Image 2 first -face, build, costume, so identity holds independently of whatever motion gets applied to it 2. Convert the source dance footage into a depth map instead of using the raw video -this strips out the original performer's appearance, clothing, and environment entirely 3. Feed the depth map as the motion reference and the character sheet as the identity reference- two separate inputs doing two separate jobs, not one input trying to do both 5. Let the depth map carry only spatial movement -the model receives body position and momentum with no competing information about who's moving or what they look like 6. Keep the character and motion inputs isolated throughout - the moment you mix appearance data into the motion reference, the model starts negotiating between two identities Why this works • Raw footage passes the model everything at once- performer, wardrobe, room, lighting -and the choreography competes with all of it for attention • A depth map is pure spatial information, so the only thing left to transfer is movement • Separating identity from motion means the character can stay locked while the dance stays accurate - normally you're trading one for the other • The accuracy gain isn't the model getting better, it's the model getting fewer decisions to make Use cases: ⁃ Dance and choreography transfer onto original characters ⁃ Motion capture-style workflows without motion capture ⁃ Any sequence where a specific movement needs to survive intact ⁃ Character showcase content built on existing performance footage The character sheet answers who's dancing. The depth map answers how - and keeping those two questions separate is the whole trick.

Nexlow

116,680 views • 15 days ago