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.

Let me introduce the Blender CNC toolpath material plugin I recently made: it quickly generates pocket tool marks, parallel finishing, continuous spiral, and sidewall vertical textures. #blender #CAD #rendering #CNC #hardsurface

Bida Wang

2,068 subscribers

18,821 views • 7 days ago •via X (Twitter)

#blender #CAD #rendering #CNC #hardsurface

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

#XiangliYao Learning Blender recently, so I haven't drawn much. Here is a render I made! I retextured and shaded a lot of the material. It was a lot of fun! (if you exclude the debugging and countless program crashes)

#XiangliYao Learning Blender recently, so I haven't drawn much. Here is a render I made! I retextured and shaded a lot of the material. It was a lot of fun! (if you exclude the debugging and countless program crashes)

奈揆

37,293 views • 28 days ago

I took Karpathy’s experiment and pushed it into Star Wars. I gave Opus 5 Max the opening scene of A New Hope, it formulated and made every single Star Wars SVG, and recreated the entire sequence in Three.js as LEGO. (Including all editing, voices, etc) It worked for 8 hours and actually completed the full opening scene. For me, this represents a far more artful and aesthetically coherent direction for AI filmmaking than end to end diffusion. By 2029, I expect a proliferation of short form fan films created by pairing models like these with Blender and other 3D rendering software.

I took Karpathy’s experiment and pushed it into Star Wars. I gave Opus 5 Max the opening scene of A New Hope, it formulated and made every single Star Wars SVG, and recreated the entire sequence in Three.js as LEGO. (Including all editing, voices, etc) It worked for 8 hours and actually completed the full opening scene. For me, this represents a far more artful and aesthetically coherent direction for AI filmmaking than end to end diffusion. By 2029, I expect a proliferation of short form fan films created by pairing models like these with Blender and other 3D rendering software.

Chris

89,646 views • 20 days ago

On Figma Motion, my demo & review. This is one of those updates that clicked swiftly into my workflow. Like in my 3D pipeline, using this stick shift animation, I emphasize hardware feedback with motion from Figma Motion. Assets are animated in Fig Motion and loaded into Blender as video textures for mesh and material control. It's advantageous since it is built into an ubiquitous ecosystem. A lot of work these days start in Figma even for someone like me that works outside the tool frequently. Can't over-emphasize my appreciation for the schema used to introduce motion to nearly all properties. Reminds me of how motion works in Blender - with better complexity management. Better than introducing more menus for those. Ease curves iD between keyframes is also an improvement over existing systems. Having spent some time on the tool, here are updates I'll like to see to Figma Motion. These are based on my immediate needs as at the time of working on this demo: • Using modifier keys (CMD, Shift, Ctrl) to jump / nudge across the timeline. E.g: holding down Shift key to make the playhead jump 100ms etc. • Jump to previous or next keyframe. • Better anchor point trails • Renaming layers from the timeline as well. • Editable custom bezier styles: after saving a style, I should be able to edit the curve - just as you can edit colour styles. • Enable transform on the timeline. I should be able to flip or scale keyframes. I hope to continue trying more ideas. Craft focused tools have had to take a backseat in recent times. Pleasing to see this come to light.

On Figma Motion, my demo & review. This is one of those updates that clicked swiftly into my workflow. Like in my 3D pipeline, using this stick shift animation, I emphasize hardware feedback with motion from Figma Motion. Assets are animated in Fig Motion and loaded into Blender as video textures for mesh and material control. It's advantageous since it is built into an ubiquitous ecosystem. A lot of work these days start in Figma even for someone like me that works outside the tool frequently. Can't over-emphasize my appreciation for the schema used to introduce motion to nearly all properties. Reminds me of how motion works in Blender - with better complexity management. Better than introducing more menus for those. Ease curves iD between keyframes is also an improvement over existing systems. Having spent some time on the tool, here are updates I'll like to see to Figma Motion. These are based on my immediate needs as at the time of working on this demo: • Using modifier keys (CMD, Shift, Ctrl) to jump / nudge across the timeline. E.g: holding down Shift key to make the playhead jump 100ms etc. • Jump to previous or next keyframe. • Better anchor point trails • Renaming layers from the timeline as well. • Editable custom bezier styles: after saving a style, I should be able to edit the curve - just as you can edit colour styles. • Enable transform on the timeline. I should be able to flip or scale keyframes. I hope to continue trying more ideas. Craft focused tools have had to take a backseat in recent times. Pleasing to see this come to light.

seyi

29,044 views • 1 month ago

I started using Blender through MCP about two weeks ago, and I quickly realized that you can build almost anything with AI. This model was created using Blender, Hunyuan3D, Gemini, and ChatGPT. Here’s how I did it: I opened Gemini, uploaded an image of the Gundam model, and asked it to generate a clean front-view image. I uploaded that front view to ChatGPT and asked it to generate two additional angles: a back view and a 45-degree front-left view. I went to: I signed up with my email and translated the page into English. Then I opened Image to 3D and selected the multi-image option. You’ll see a diagram of a whale from several angles. Upload each reference image in its corresponding position, such as front, 45-degree front-left, and back. I selected the 1.5M-face option. This produces a very high-poly model, but don’t worry, we’ll fix that next. Once the generation is complete, download the model as a GLB file. From the Hunyuan homepage, open 3D Studio using one of the dropdown menus. Select the topology or retopology tool and upload your GLB. I chose the High setting to preserve as much detail as possible. After a few seconds, the model was retopologized. It kept most of its visual detail while using far fewer polygons. The original head and rifle didn’t look very good, so I generated them separately. I returned to ChatGPT and created dedicated reference images for the Gundam’s head and rifle. I generated each part individually in Hunyuan at the 1.5M setting, then ran both through the same retopology process. Next came the textures. Open the texture section, select the multi-image option, and upload the same reference images according to the whale orientation indicators. However, instead of using the original clean textures, I asked ChatGPT to recreate them with wear, rust stains, scratches, and other surface damage. This gave the Gundam a much older and more authentic appearance. Once every part was textured, I imported everything into Blender. You can ask Codex through MCP to remove the original head and rifle, or you can do it manually. Select the main model and press Tab to enter Edit Mode. Press 3 to enable face selection, then press C to activate Circle Select. Paint over the faces belonging to the original helmet or rifle. You can press X to delete those faces or P to separate them into another object. Then position the newly generated, more detailed head and rifle in their place. I demonstrate this process in one of my older videos. And that’s it. You now have a very cool 3D Gundam model! Afterward, I created the cockpit, separated the model into movable sections, and rigged everything for use in my Three.js game. That process deserves its own tutorial, though. If anyone wants to see it, let me know. Or just ask your AI, I guess. They seem to know everything these days. xD

I started using Blender through MCP about two weeks ago, and I quickly realized that you can build almost anything with AI. This model was created using Blender, Hunyuan3D, Gemini, and ChatGPT. Here’s how I did it: I opened Gemini, uploaded an image of the Gundam model, and asked it to generate a clean front-view image. I uploaded that front view to ChatGPT and asked it to generate two additional angles: a back view and a 45-degree front-left view. I went to: I signed up with my email and translated the page into English. Then I opened Image to 3D and selected the multi-image option. You’ll see a diagram of a whale from several angles. Upload each reference image in its corresponding position, such as front, 45-degree front-left, and back. I selected the 1.5M-face option. This produces a very high-poly model, but don’t worry, we’ll fix that next. Once the generation is complete, download the model as a GLB file. From the Hunyuan homepage, open 3D Studio using one of the dropdown menus. Select the topology or retopology tool and upload your GLB. I chose the High setting to preserve as much detail as possible. After a few seconds, the model was retopologized. It kept most of its visual detail while using far fewer polygons. The original head and rifle didn’t look very good, so I generated them separately. I returned to ChatGPT and created dedicated reference images for the Gundam’s head and rifle. I generated each part individually in Hunyuan at the 1.5M setting, then ran both through the same retopology process. Next came the textures. Open the texture section, select the multi-image option, and upload the same reference images according to the whale orientation indicators. However, instead of using the original clean textures, I asked ChatGPT to recreate them with wear, rust stains, scratches, and other surface damage. This gave the Gundam a much older and more authentic appearance. Once every part was textured, I imported everything into Blender. You can ask Codex through MCP to remove the original head and rifle, or you can do it manually. Select the main model and press Tab to enter Edit Mode. Press 3 to enable face selection, then press C to activate Circle Select. Paint over the faces belonging to the original helmet or rifle. You can press X to delete those faces or P to separate them into another object. Then position the newly generated, more detailed head and rifle in their place. I demonstrate this process in one of my older videos. And that’s it. You now have a very cool 3D Gundam model! Afterward, I created the cockpit, separated the model into movable sections, and rigged everything for use in my Three.js game. That process deserves its own tutorial, though. If anyone wants to see it, let me know. Or just ask your AI, I guess. They seem to know everything these days. xD

Spectro

79,320 views • 29 days ago

🙌Meet Artifig: A Figma Plugin to Generate Figma Plugins Do you use Figma and ever feel like this: - Your mind is bursting with plugin ideas, but you can't bring them to life because you don't know how to code? - You want to focus on design, but repetitive tasks keep slowing you down? - You dream of creating custom tools for your team, but lack the time or resources? I’ve been there too. That’s why I created Artifig. ✨ What is Artifig? Artifig is an AI-powered Figma plugin that empowers anyone to build their own Figma plugins using just natural language. No coding needed—simply describe what you want, and watch as your idea transforms into a fully functional, real-time plugin. 🚀 Redefining Figma Plugin Development The core philosophy of Artifig is simple: Designers often have countless ideas and creative visions, but many of them remain unrealized due to a lack of technical skills. We believe designers shouldn’t be limited by their inability to code. You should focus on creating, not be held back by technical barriers or repetitive tasks. Artifig takes you directly from "description" to "implementation." 🛠️ How Does It Work? 1. Describe Your Needs: Tell Artifig what you want, like “Create a skew transformation tool for objects, supporting horizontal and vertical skew with real-time preview functionality.” 2. Generate and Run the Plugin: Artifig instantly generates the plugin and runs it right within Figma. For example, the generated plugin can apply skew transformations to objects, precisely controlled via matrix transformations, with an intuitive user experience. 3. Optimize and Iteration: Need adjustments? Simply describe them, and Artifig will Iterating the plugin step by step. 4. Share Your Creations: Publish your plugins to the Artifig community, or remix plugins shared by others to build on their ideas. No learning curve. No complex steps. It’s as simple as that. 🌟 Key Features - Zero Barrier to Entry: No coding experience needed—any Figma user can create plugins effortlessly. - Multilingual Support: Works in multiple languages, including English, Chinese, French, Japanese, and German. - What-You-See-Is-What-You-Get: Generated plugins run in real-time, so you can quickly validate and refine your ideas. - Open and Flexible: The generated plugin code is 100% yours—modify it, distribute it, even use it commercially. - Global Community: Share your plugins, explore others’ creations, and publish your plugins to the Figma community. 🎯 Why is Artifig a Game-Changer? 1. No More Repetitive Work Let AI handle the tedious, time-consuming tasks: batch renaming layers, auto-aligning elements, or applying styles in bulk. All you need to do is say, “Import a PDF and arrange each image on the canvas with 20px spacing.” 2. Quickly Bring Ideas to Life From color contrast checks to data imports and custom components, all your “what if we could” ideas can now become plugins. Just one natural language description, and Artifig makes it happen. 3. Custom Tools for Your Team Build tailored tools for your team, creating unique solutions to streamline your workflow. 4. Not Just a Tool, But a Learning Experience Artifig explains the logic behind the code it generates, helping you understand Figma APIs and JavaScript. Today, you’re a designer; tomorrow, you could also be a design engineer. 🧑‍🚀👩🏻‍💻🥷🏻 Who is Artifig For? - Beginners: No development experience needed—just describe your ideas and let Artifig do the rest. - Experts: Save time and focus on high-value tasks while Artifig handles the repetitive work. - Learners: Use Artifig as a bridge to deepen your understanding of development. - Teams: Build custom tools to enhance collaboration and efficiency. 🎉 Ready to Get Started? I believe designers’ time and focus should be spent on creating, not on wrestling with complex tools. Artifig is the first step toward realizing this vision. Try Artifig now and experience an unprecedented flow of creativity!

🙌Meet Artifig: A Figma Plugin to Generate Figma Plugins Do you use Figma and ever feel like this: - Your mind is bursting with plugin ideas, but you can't bring them to life because you don't know how to code? - You want to focus on design, but repetitive tasks keep slowing you down? - You dream of creating custom tools for your team, but lack the time or resources? I’ve been there too. That’s why I created Artifig. ✨ What is Artifig? Artifig is an AI-powered Figma plugin that empowers anyone to build their own Figma plugins using just natural language. No coding needed—simply describe what you want, and watch as your idea transforms into a fully functional, real-time plugin. 🚀 Redefining Figma Plugin Development The core philosophy of Artifig is simple: Designers often have countless ideas and creative visions, but many of them remain unrealized due to a lack of technical skills. We believe designers shouldn’t be limited by their inability to code. You should focus on creating, not be held back by technical barriers or repetitive tasks. Artifig takes you directly from "description" to "implementation." 🛠️ How Does It Work? 1. Describe Your Needs: Tell Artifig what you want, like “Create a skew transformation tool for objects, supporting horizontal and vertical skew with real-time preview functionality.” 2. Generate and Run the Plugin: Artifig instantly generates the plugin and runs it right within Figma. For example, the generated plugin can apply skew transformations to objects, precisely controlled via matrix transformations, with an intuitive user experience. 3. Optimize and Iteration: Need adjustments? Simply describe them, and Artifig will Iterating the plugin step by step. 4. Share Your Creations: Publish your plugins to the Artifig community, or remix plugins shared by others to build on their ideas. No learning curve. No complex steps. It’s as simple as that. 🌟 Key Features - Zero Barrier to Entry: No coding experience needed—any Figma user can create plugins effortlessly. - Multilingual Support: Works in multiple languages, including English, Chinese, French, Japanese, and German. - What-You-See-Is-What-You-Get: Generated plugins run in real-time, so you can quickly validate and refine your ideas. - Open and Flexible: The generated plugin code is 100% yours—modify it, distribute it, even use it commercially. - Global Community: Share your plugins, explore others’ creations, and publish your plugins to the Figma community. 🎯 Why is Artifig a Game-Changer? 1. No More Repetitive Work Let AI handle the tedious, time-consuming tasks: batch renaming layers, auto-aligning elements, or applying styles in bulk. All you need to do is say, “Import a PDF and arrange each image on the canvas with 20px spacing.” 2. Quickly Bring Ideas to Life From color contrast checks to data imports and custom components, all your “what if we could” ideas can now become plugins. Just one natural language description, and Artifig makes it happen. 3. Custom Tools for Your Team Build tailored tools for your team, creating unique solutions to streamline your workflow. 4. Not Just a Tool, But a Learning Experience Artifig explains the logic behind the code it generates, helping you understand Figma APIs and JavaScript. Today, you’re a designer; tomorrow, you could also be a design engineer. 🧑‍🚀👩🏻‍💻🥷🏻 Who is Artifig For? - Beginners: No development experience needed—just describe your ideas and let Artifig do the rest. - Experts: Save time and focus on high-value tasks while Artifig handles the repetitive work. - Learners: Use Artifig as a bridge to deepen your understanding of development. - Teams: Build custom tools to enhance collaboration and efficiency. 🎉 Ready to Get Started? I believe designers’ time and focus should be spent on creating, not on wrestling with complex tools. Artifig is the first step toward realizing this vision. Try Artifig now and experience an unprecedented flow of creativity!

yancymin

21,222 views • 1 year ago

2025 was a bit of a special year for me. lots of change, lots of personality-defining moment, and lots of growth as well. i want to thank all my friends: work friends, party friends, friend friends, and especially friends in the intersection of these friend categories (you know who you are) <3 i would like to thank all partners, clients, and everyone else i might have come across in 2025. going into 2026, i'm hopeful that we will be even better and continue to improve ourselves. big things coming! <- i am serious anyway, i made this scene quickly to take a little break from work :) i designed this character with Dimas at the end of 2025, and it looks like i might see her more often in my personal works throughout 2026 <3 created in Blender, Houdini, and DaVinci Resolve. The usual suspects. cheers! -i #b3d #blender3d #houdini #sidefxhoudini

2025 was a bit of a special year for me. lots of change, lots of personality-defining moment, and lots of growth as well. i want to thank all my friends: work friends, party friends, friend friends, and especially friends in the intersection of these friend categories (you know who you are) <3 i would like to thank all partners, clients, and everyone else i might have come across in 2025. going into 2026, i'm hopeful that we will be even better and continue to improve ourselves. big things coming! <- i am serious anyway, i made this scene quickly to take a little break from work :) i designed this character with Dimas at the end of 2025, and it looks like i might see her more often in my personal works throughout 2026 <3 created in Blender, Houdini, and DaVinci Resolve. The usual suspects. cheers! -i #b3d #blender3d #houdini #sidefxhoudini

izzy

13,817 views • 7 months ago

Open source/weight models are often used in regulated industries like Health Care or Financial Services, where they handle personally identifiable data, and can't send it to proprietary LLM providers. We recently chatted to Vaibhav (VB) Srivastav about the partnership VS Visual Studio Code and Hugging Face inference providers have to let you use open weight models directly in your IDE! Something that surprised me was just how fast inference providers like Cerebras make it to generate code! In this episode we made a journaling CLI tool using Qwen! I personally love open source and hope more of these models will become small enough to run locally!

Open source/weight models are often used in regulated industries like Health Care or Financial Services, where they handle personally identifiable data, and can't send it to proprietary LLM providers. We recently chatted to Vaibhav (VB) Srivastav about the partnership VS Visual Studio Code and Hugging Face inference providers have to let you use open weight models directly in your IDE! Something that surprised me was just how fast inference providers like Cerebras make it to generate code! In this episode we made a journaling CLI tool using Qwen! I personally love open source and hope more of these models will become small enough to run locally!

Marlene Mhangami

18,215 views • 9 months ago

New clean 3D AI workflow Control Rig physics in UEFN! Been exploring some more hybrid AI workflows lately, and for 3D assets splitting the model into parts is kinda essential. This video shows my first test using AssetHub, one of the cleanest ways i've used to keep the whole generation process organized. It auto-detected the pose in my main image and broke it down into parts, then generated images for each. From there I can just iterate on each piece separately and generate the 3D models in one workflow. AssetHub is still in closed beta atm, so let me know if you want to try it out. Main reference image came from Grok Imagine. All the part images and high poly meshes were done in AssetHub, selected Tripo 3.1 as the main model. After that I ran it through Hunyuan 3D for retopo + unwrap. Final assembly in Blender. Baked normals from high to low poly. End result: under 60K tris, 3 materials, 2K textures. Also Control Rig physics in Unreal Engine are insane. It finally validates in UEFN too, so of course I had to test it out.

New clean 3D AI workflow Control Rig physics in UEFN! Been exploring some more hybrid AI workflows lately, and for 3D assets splitting the model into parts is kinda essential. This video shows my first test using AssetHub, one of the cleanest ways i've used to keep the whole generation process organized. It auto-detected the pose in my main image and broke it down into parts, then generated images for each. From there I can just iterate on each piece separately and generate the 3D models in one workflow. AssetHub is still in closed beta atm, so let me know if you want to try it out. Main reference image came from Grok Imagine. All the part images and high poly meshes were done in AssetHub, selected Tripo 3.1 as the main model. After that I ran it through Hunyuan 3D for retopo + unwrap. Final assembly in Blender. Baked normals from high to low poly. End result: under 60K tris, 3 materials, 2K textures. Also Control Rig physics in Unreal Engine are insane. It finally validates in UEFN too, so of course I had to test it out.

Jerome | InsaneUnreal

82,769 views • 4 months ago

Good morning! I'm excited to introduce my Partner and Coder of the Exposure Tool, Jody. Here are a few words from him and a video about the tool! Jody haven coding for over 25 years! , This is not Just a simple Tool who deploy Gex.. this its a Unique and Complete Profiling Tool for TOS , This will change the way we trade and how we read the market, On the Discord we have EDU and more info and as well live stream , If you want to Subscribe to the tool , Join discord and open a ticket or just DM me. Im going to Give a bundle of this tool + Option Volume Profile . ConvexSwan Doc aka Trader McGraw ∴ 🕉️ShivaAnalytics🕉️ ∴ Hello Everyone, I wanted to take a moment to introduce myself. My name is Jody and I have been a TOS user for over 25 years and have created hundreds of custom studies and strategies under the name CustomThinkscript. SPX_Omega recently approached me regarding the GEX Profile study that I had created a few years back. He stated that he had a group of like minded traders that were very interested in this study and asked if I would partner up and develop additional features. When the GEX was first created it was designed within the parameters that a customer had requested, it was then delivered to this trader and not really updated much since that time. Given the renewed interest from SPX_Omega I reexamined the study, optimized the code and added a ton of new features. The current version which is going to be rebranded as “Exposure Profile” and released shortly can now plot 6 different exposure profiles (Gamma, Delta, Vega, Charm, Vanna and Vomma). Remember these are exposure profiles so we are not just plotting straight gamma or delta, we are applying them to open interest and getting exposure level data with individual strike level IV calculations. Additionally, I also created a companion study that bypasses some thinkscript restrictions and allows me to plot the profiles of indices like SPX. Another recent addition was the ability to view the data profiles on an intraday chart. It sure has been an exciting couple of weeks breathing new life into this study. I made a short video detailing the installation and some of the features. Be kind, video editing is not my forte but there is some good information in there and I hope you check it out. Jody

SPX_Omega

24,843 views • 1 year ago

You are probably wrong about open weight models 🫣 I was too. Then I started using more open weight models, recently tried MiniMax_Agent. I honestly did not expect the quality I got from MiniMax. It built this entire 3D skateboard application with Three.js, including Awwwards style animations that too writing whole codebase in under 5 minutes. But then it spent another 30 to 45 minutes testing everything by itself in a headless Chromium browser. It took screenshots, verified interactions, and tested the application end to end before finishing. I actually love that it does this because the final output is much more reliable. I just wish there was a setting to skip browser testing when I want to iterate quickly. Its code generation is so fast that this would make it an incredible tool for rapid prototyping. Quality also surprised me the most. It generated high quality 3D geometry using nothing but math. It also created really polished animations. My prompt was simply to make them feel like something you would see on Awwwards. I never described how the animations should work. It figured that out on its own. After spending more time with MiniMax M3 and other, I think open weight models are much closer to Claude and GPT than many people realize. Or maybe better? I am still testing them, but so far I have been genuinely impressed. Also, it didn't cost anything as I got free credits on sign up. Can't say no to freebies 😂 Live: Design:

You are probably wrong about open weight models 🫣 I was too. Then I started using more open weight models, recently tried MiniMax_Agent. I honestly did not expect the quality I got from MiniMax. It built this entire 3D skateboard application with Three.js, including Awwwards style animations that too writing whole codebase in under 5 minutes. But then it spent another 30 to 45 minutes testing everything by itself in a headless Chromium browser. It took screenshots, verified interactions, and tested the application end to end before finishing. I actually love that it does this because the final output is much more reliable. I just wish there was a setting to skip browser testing when I want to iterate quickly. Its code generation is so fast that this would make it an incredible tool for rapid prototyping. Quality also surprised me the most. It generated high quality 3D geometry using nothing but math. It also created really polished animations. My prompt was simply to make them feel like something you would see on Awwwards. I never described how the animations should work. It figured that out on its own. After spending more time with MiniMax M3 and other, I think open weight models are much closer to Claude and GPT than many people realize. Or maybe better? I am still testing them, but so far I have been genuinely impressed. Also, it didn't cost anything as I got free credits on sign up. Can't say no to freebies 😂 Live: Design:

The Bugged Dev

16,348 views • 16 days ago

haha the Runway real-time characters is great. i just spent 2 minutes pitching a ghibli version of myself a bladeless blender startup it's only a 2 min demo, but the questions and tone are pretty solid. one question even threw me off a bit (the cleaning). asking which round was unnecessary. it got a little weird at the end right before it cut off, which is due to a 2 min limit when testing on site(which seems unnecessary since they're charging me for the api - why not let me talk longer?) at current pricing, this would be roughly ~$6 for a 30 min call or $12 an hour. this doesn't include the cost of running an extraction prompt in the background to store this data, but llm tokens are so cheap. very expensive compared to mean vc (free, since it's in gpt marketplace) or even a self-hosted chat agent, but it's way cheaper than an associate. while it has downsides like not being a real person, there's also benefits like having consistent and unbiased information gathering, being available 24/7, being able to talk to founders in parallel, and consistent note taking. and it will likely get cheaper (sorry my audio volume is low, i'm testing while everyone else in the house is asleep)

haha the Runway real-time characters is great. i just spent 2 minutes pitching a ghibli version of myself a bladeless blender startup it's only a 2 min demo, but the questions and tone are pretty solid. one question even threw me off a bit (the cleaning). asking which round was unnecessary. it got a little weird at the end right before it cut off, which is due to a 2 min limit when testing on site(which seems unnecessary since they're charging me for the api - why not let me talk longer?) at current pricing, this would be roughly ~$6 for a 30 min call or $12 an hour. this doesn't include the cost of running an extraction prompt in the background to store this data, but llm tokens are so cheap. very expensive compared to mean vc (free, since it's in gpt marketplace) or even a self-hosted chat agent, but it's way cheaper than an associate. while it has downsides like not being a real person, there's also benefits like having consistent and unbiased information gathering, being available 24/7, being able to talk to founders in parallel, and consistent note taking. and it will likely get cheaper (sorry my audio volume is low, i'm testing while everyone else in the house is asleep)

Yohei

19,803 views • 5 months ago

Most AI app builders can generate a prototype. The real question is: Can they generate something you can actually launch? I recently explored MeDo, an AI-powered full-stack app generation platform that goes beyond just creating a frontend. With a single prompt, it can generate the frontend UI, backend logic, database structure, and plugin integrations. All in one workflow, without requiring coding or deployment experience. What stood out to me is the focus on building real products rather than simple demos. Whether you're creating SaaS tools, internal dashboards, e-commerce applications, games, or business automation tools. the goal is to move from idea → working product as quickly as possible. A few notable features: ✅ Full-stack generation from a single prompt ✅ Plugin ecosystem with integrations like Stripe and AI tools ✅ Auto bug-fixing capabilities ✅ Website cloning from screenshots ✅ Built-in UI enhancement options ✅ Marketplace of ready-made apps you can customize The AI app builder space is becoming increasingly competitive, and platforms are starting to differentiate themselves based on how close they get to a production-ready result rather than just generating code. For builders, founders, and creators exploring no-code AI development, it's an interesting category worth watching. Try it today and launch your next project with just a prompt:

Most AI app builders can generate a prototype. The real question is: Can they generate something you can actually launch? I recently explored MeDo, an AI-powered full-stack app generation platform that goes beyond just creating a frontend. With a single prompt, it can generate the frontend UI, backend logic, database structure, and plugin integrations. All in one workflow, without requiring coding or deployment experience. What stood out to me is the focus on building real products rather than simple demos. Whether you're creating SaaS tools, internal dashboards, e-commerce applications, games, or business automation tools. the goal is to move from idea → working product as quickly as possible. A few notable features: ✅ Full-stack generation from a single prompt ✅ Plugin ecosystem with integrations like Stripe and AI tools ✅ Auto bug-fixing capabilities ✅ Website cloning from screenshots ✅ Built-in UI enhancement options ✅ Marketplace of ready-made apps you can customize The AI app builder space is becoming increasingly competitive, and platforms are starting to differentiate themselves based on how close they get to a production-ready result rather than just generating code. For builders, founders, and creators exploring no-code AI development, it's an interesting category worth watching. Try it today and launch your next project with just a prompt:

Markandey Sharma

49,777 views • 2 months ago

Day 05 of building the space game I saw in a dream. 🌙🚀 3D spacecraft made COMPLETELY FROM SCRATCH by Fable 5 + Grok 4.5, not a single downloaded asset. The entire ship is generated by Python code in Blender: geometry, UVs, textures, rigging, even the marketing renders. The display showcase is Three.js Meet the Meteor V2 "Fable" — a 12.5 m single-seat light interceptor from a fictional manufacturer called Kizuna Technologies: 🛠️ 86k tris, 6 PBR texture sets baked at 4K, AI-generated surface detail — panel lines, rivets, ceramic heat tiles, weathered decals + full brand livery Complete FPV cockpit interior watertight from every head position (VR-ready), lit MFD avionics screens, switch panels, warning placards, night backlighting, ejection seat... Fully rigged as game modules: retractable sliding canopy, folding landing gear with rolling wheels, elevons, rudders, 20-petal variable engine nozzle, animated air ingestion 🚀 Dynamic hardpoints — quad-tube missile loader with 8 internal reloads + heat-seeker and striker missiles, all swappable in the hangar. The output also included a full brochure for ingame lore, but I didn't include it in the video. And the testing "hangar" is its own build, a three.js staging ground with lighting (studio/day/sunset/night), weathering sliders (scratches, grime, engine soot), wireframe, true-scale dimension lines, a render album, and an annotation tool I use to give the AI visual feedback directly on the model. Zero manual modeling. I describe, review, and direct the agents measure, build, validate, and fix their own mistakes. Tomorrow - let's see how this thing flies!

Startracker 🔺

76,897 views • 1 month ago

- AssetHub officially launched - Using Metahuman Component as parts system in UEFN This is crazy! Just discovered I can use the Metahuman Component in UEFN as a general parts system! Perfect for modular characters. So I had to test it out. First step: character design. With GPT Image 2 I got some really solid characters to work with. These are perfect to test anime/ lowpoly models as well. Gavin Harvey also officially launched so I jumped on a fresh workflow to build the first one. Since I’m aiming for an anime/flat shaded style, I chose Tripo P1 to generate straight to low poly models. Results were perfect for me, UV and textured directly done. Did the main assembly in Blender. I transferred weights onto the new parts (like the jacket here) and used proportional editing to tweak vertices so everything fits together nicely. Then into UEFN / Unreal Engine: For the two-tone shadow look, I made a material that uses the sun direction as a mask. To make all the modular parts work together, I set up a blueprint with the body skeletal mesh as the base, and all the other parts attached as children. Now the really cool part: the Metahuman Component actually lets you plug in extra body parts directly. And every skeletal mesh can have its own Control Rig slot. That means you can layer procedural animations on top of your main animation setup. I used it for procedural eye blinking and hair physics. Since every part is separate, swapping variations is super easy while keeping all animations perfectly synced!

- AssetHub officially launched - Using Metahuman Component as parts system in UEFN This is crazy! Just discovered I can use the Metahuman Component in UEFN as a general parts system! Perfect for modular characters. So I had to test it out. First step: character design. With GPT Image 2 I got some really solid characters to work with. These are perfect to test anime/ lowpoly models as well. Gavin Harvey also officially launched so I jumped on a fresh workflow to build the first one. Since I’m aiming for an anime/flat shaded style, I chose Tripo P1 to generate straight to low poly models. Results were perfect for me, UV and textured directly done. Did the main assembly in Blender. I transferred weights onto the new parts (like the jacket here) and used proportional editing to tweak vertices so everything fits together nicely. Then into UEFN / Unreal Engine: For the two-tone shadow look, I made a material that uses the sun direction as a mask. To make all the modular parts work together, I set up a blueprint with the body skeletal mesh as the base, and all the other parts attached as children. Now the really cool part: the Metahuman Component actually lets you plug in extra body parts directly. And every skeletal mesh can have its own Control Rig slot. That means you can layer procedural animations on top of your main animation setup. I used it for procedural eye blinking and hair physics. Since every part is separate, swapping variations is super easy while keeping all animations perfectly synced!

Jerome | InsaneUnreal

24,268 views • 2 months ago

Fast Company just published a great piece on World Labs , Fei-Fei Li , Marble, and the idea that spatial intelligence / world models may be one of the next big shifts in AI. I was happy to be quoted in the article, but I also wanted to share more context about my own experience with World Labs and Marble, and why this direction is especially interesting to me. My starting point: volumetric capture — For the past few years I’ve been exploring and using volumetric capture and reconstruction (photogrammetry, NeRFs, 3D Gaussian Splats) mostly capturing locations around Montreal. Alleys, museums, urban interiors. I love every step of it: the capture itself, the pipeline, and what can be done with the output. Turning real spaces into real-time explorable systems. I do this personally, sharing explorations here, and professionally as chief technologist, and co-founder of Dpt. Physical reality + generative manipulation — In my work I’m especially drawn to mixing physical reality with generative and digital manipulation: using physical interfaces (light, clay, ink, ... ) to drive generative AI pipelines, building mixed reality prototypes that reshape your surroundings, or starting from real captured spaces and transforming them using tools like Marble. Like many people, I saw the World Labs announcement on Twitter in September 2024, and Marble when it surfaced in early December. But by then, I already had a sense something was coming. The first conversation — As someone deep into volumetric capture and radiance fields, I obviously knew about Ben Mildenhall and his pioneering work on NeRF. To my surprise, Ben reached out to me in late June 2024. He’d been following some of my experiments and wanted to chat about my process and workflows and how I was using this “stuff” creatively. At that point he didn’t share what he was building, but we had a genuinely great conversation about radiance fields, AI, and my work. He was curious about the creative perspective, not just the technical one. When the World Labs announcement dropped a few months later, it all made sense. I understood what Ben had been working on, and why the creative angle mattered to them. Then in August 2025, he invited me to try the Marble beta, and I’ve been experimenting with it since. Experimenting with Marble — The first thing I used Marble for was materializing scene and world concepts during ideation at the studio, and seeing if and how it could fit into our production pipeline. In parallel, I dove into a series of experiments focused on world manipulation: starting from real captured spaces and transforming them using Marble. I’d already been exploring that idea using img2img diffusion with ControlNet on NeRF renders, real-time video streams, and even mixed reality using headset camera feeds. But Marble brings something different. It generates persistent, spatially cohesive 3D worlds that can be rendered in real time across a wide range of devices. That’s a real shift. Experiment 01: Parallel Realities — The first experiment, Parallel Realities, starts from a volumetric capture of a real location, reconstructed as 3D Gaussian Splats. Using Marble, I generate an alternate version of that same space, something informed by the original architecture: abandoned, nature-reclaimed, alternate era. Then, using Spark (World Labs’ 3D Gaussian Splatting renderer for THREE.js) I make both realities coexist in the same spatial coordinate system. From there, I use a portal UX mechanic to let the user step between the real reconstruction and the Marble-generated version. Experiment 02: Hidden Depth The second experiment, Hidden Depth, does not transform a space as much as expand it. A captured location has a visual boundary (a mural, a doorway, a dark corridor) and Marble generates what exists beyond it. For example: a Montreal alley has a painted mural; step through it and you’re inside a world informed by what is actually depicted there. World Labs showcased part of this work here: And in their Spark 2.0 post: The project page is here: Why this matters to me — Being able to start from a real 3D Gaussian Splat scene and manipulate it with Marble opens up a lot of ideas. The 3DGS pipeline is becoming an increasingly compelling foundation for exploration, experimentation, and storytelling. What matters most to me right now is more control. The more I can steer the generated scene or world, the more useful the tool becomes. I want more features like the already existing multiple input images and Chisel, the blockout-based approach. I would like better local control, the ability to expand a generated world more and more while preserving coherence, and the ability to directly import 3D Gaussian Splat scenes to be used as a starting point. I want more ways to shape the result, not just a “prompt and hope” approach. — It is exciting to see this field moving from research and demos toward actual creative workflows.

Fast Company just published a great piece on World Labs , Fei-Fei Li , Marble, and the idea that spatial intelligence / world models may be one of the next big shifts in AI. I was happy to be quoted in the article, but I also wanted to share more context about my own experience with World Labs and Marble, and why this direction is especially interesting to me. My starting point: volumetric capture — For the past few years I’ve been exploring and using volumetric capture and reconstruction (photogrammetry, NeRFs, 3D Gaussian Splats) mostly capturing locations around Montreal. Alleys, museums, urban interiors. I love every step of it: the capture itself, the pipeline, and what can be done with the output. Turning real spaces into real-time explorable systems. I do this personally, sharing explorations here, and professionally as chief technologist, and co-founder of Dpt. Physical reality + generative manipulation — In my work I’m especially drawn to mixing physical reality with generative and digital manipulation: using physical interfaces (light, clay, ink, ... ) to drive generative AI pipelines, building mixed reality prototypes that reshape your surroundings, or starting from real captured spaces and transforming them using tools like Marble. Like many people, I saw the World Labs announcement on Twitter in September 2024, and Marble when it surfaced in early December. But by then, I already had a sense something was coming. The first conversation — As someone deep into volumetric capture and radiance fields, I obviously knew about Ben Mildenhall and his pioneering work on NeRF. To my surprise, Ben reached out to me in late June 2024. He’d been following some of my experiments and wanted to chat about my process and workflows and how I was using this “stuff” creatively. At that point he didn’t share what he was building, but we had a genuinely great conversation about radiance fields, AI, and my work. He was curious about the creative perspective, not just the technical one. When the World Labs announcement dropped a few months later, it all made sense. I understood what Ben had been working on, and why the creative angle mattered to them. Then in August 2025, he invited me to try the Marble beta, and I’ve been experimenting with it since. Experimenting with Marble — The first thing I used Marble for was materializing scene and world concepts during ideation at the studio, and seeing if and how it could fit into our production pipeline. In parallel, I dove into a series of experiments focused on world manipulation: starting from real captured spaces and transforming them using Marble. I’d already been exploring that idea using img2img diffusion with ControlNet on NeRF renders, real-time video streams, and even mixed reality using headset camera feeds. But Marble brings something different. It generates persistent, spatially cohesive 3D worlds that can be rendered in real time across a wide range of devices. That’s a real shift. Experiment 01: Parallel Realities — The first experiment, Parallel Realities, starts from a volumetric capture of a real location, reconstructed as 3D Gaussian Splats. Using Marble, I generate an alternate version of that same space, something informed by the original architecture: abandoned, nature-reclaimed, alternate era. Then, using Spark (World Labs’ 3D Gaussian Splatting renderer for THREE.js) I make both realities coexist in the same spatial coordinate system. From there, I use a portal UX mechanic to let the user step between the real reconstruction and the Marble-generated version. Experiment 02: Hidden Depth The second experiment, Hidden Depth, does not transform a space as much as expand it. A captured location has a visual boundary (a mural, a doorway, a dark corridor) and Marble generates what exists beyond it. For example: a Montreal alley has a painted mural; step through it and you’re inside a world informed by what is actually depicted there. World Labs showcased part of this work here: And in their Spark 2.0 post: The project page is here: Why this matters to me — Being able to start from a real 3D Gaussian Splat scene and manipulate it with Marble opens up a lot of ideas. The 3DGS pipeline is becoming an increasingly compelling foundation for exploration, experimentation, and storytelling. What matters most to me right now is more control. The more I can steer the generated scene or world, the more useful the tool becomes. I want more features like the already existing multiple input images and Chisel, the blockout-based approach. I would like better local control, the ability to expand a generated world more and more while preserving coherence, and the ability to directly import 3D Gaussian Splat scenes to be used as a starting point. I want more ways to shape the result, not just a “prompt and hope” approach. — It is exciting to see this field moving from research and demos toward actual creative workflows.

Hugues Bruyère

69,960 views • 2 months ago

I ditched Unreal for AI. Kaiju Engine is a completely AI written replacement. To prove it was ready for production, I tested it by recreating and porting my old game, Firefall, into Kaiju, purely from the Steam download, no source. And it worked. Real game code is messy, full of difficult details and compromises. If our engine could handle a commercially released game, it proves you can ship a game with it. We're developing 2 orders of magnitude faster (93x) by lines of code and features. We have fewer bugs, and iteration is super fast. Build times have gone from 30 mins (Unreal) to 36 seconds. Feature rate is through the roof, days instead of months. Our original engine for Firefall took 2-3 years and cost over 5M to build (adjusted). Our new engine was written in 5 months for a couple grand. We built only what we needed, with none of the Unreal bloat. And we added some features: - Modern meshlet based rendering with PBR. - Vis Buffer/Forward+ with clustered lighting. - Id Tech 5 style megatexture streaming. - Planetary sized renderer (entire solar systems possible). - Seamless flight from ground to orbit and space. - No loading screens. - Procedural planet generation with plate simulation. - Client/Server at all times. - Ozz for animation. Jolt for Physics. - Companion Blender plug-in for AI directed asset exports. - 140fps currently, 200-300 projected after optimizations. You can see Firefall's assets and levels ported over into our engine in the video. Texture resolution and pop-in are limitations of the original game (high rez CDN based textures were lost when game went offline, fog hid the original's pop-in). Our meshlet renderer will be able to do much better with LOD and already supports high rez textures. I used Grok/Codex/Claude to tag team the code. This is not just a boon for indie games, it's a real game changer for game preservation. The reaction to seeing this 10 year old discontinued game revived is very emotional for Firefall fans. But we're not just preserving Firefall, we're going beyond, creating an original game that is the spiritual successor, Em-8ER. Gliding, jumpjets... it's all coming back. Moving past Unreal let us move much faster, without the bloat, and with better performance. If you want to signup to follow news on the game, sign up is free at

I ditched Unreal for AI. Kaiju Engine is a completely AI written replacement. To prove it was ready for production, I tested it by recreating and porting my old game, Firefall, into Kaiju, purely from the Steam download, no source. And it worked. Real game code is messy, full of difficult details and compromises. If our engine could handle a commercially released game, it proves you can ship a game with it. We're developing 2 orders of magnitude faster (93x) by lines of code and features. We have fewer bugs, and iteration is super fast. Build times have gone from 30 mins (Unreal) to 36 seconds. Feature rate is through the roof, days instead of months. Our original engine for Firefall took 2-3 years and cost over 5M to build (adjusted). Our new engine was written in 5 months for a couple grand. We built only what we needed, with none of the Unreal bloat. And we added some features: - Modern meshlet based rendering with PBR. - Vis Buffer/Forward+ with clustered lighting. - Id Tech 5 style megatexture streaming. - Planetary sized renderer (entire solar systems possible). - Seamless flight from ground to orbit and space. - No loading screens. - Procedural planet generation with plate simulation. - Client/Server at all times. - Ozz for animation. Jolt for Physics. - Companion Blender plug-in for AI directed asset exports. - 140fps currently, 200-300 projected after optimizations. You can see Firefall's assets and levels ported over into our engine in the video. Texture resolution and pop-in are limitations of the original game (high rez CDN based textures were lost when game went offline, fog hid the original's pop-in). Our meshlet renderer will be able to do much better with LOD and already supports high rez textures. I used Grok/Codex/Claude to tag team the code. This is not just a boon for indie games, it's a real game changer for game preservation. The reaction to seeing this 10 year old discontinued game revived is very emotional for Firefall fans. But we're not just preserving Firefall, we're going beyond, creating an original game that is the spiritual successor, Em-8ER. Gliding, jumpjets... it's all coming back. Moving past Unreal let us move much faster, without the bloat, and with better performance. If you want to signup to follow news on the game, sign up is free at

Grummz

350,199 views • 8 days ago

Junkyu being open about his burnout and how he overcame that phase is something that truly amazed me. His beautiful words that encourage everyone why pacing yourself actually matters 🥺🤍 🐨: I actually went through myself a burnout recently… was it in 2025? It’s not that I started to hate music, but when it came to making music, to working on it, I hit burnout. I really did. Nothing felt fun anymore, and no matter what I did, I just couldn’t move forward. No matter what—seriously, no matter what. 🐨: I work on music on my Mac, right? and I didn’t turn it on for a whole year. A year? Maybe even a year and a half? Since sometime in late 2024? It’s not that I chose not to turn it on—I couldn’t. That thought kept coming to me, “Even if I turn it on, nothing will come out” or maybe, “Even if I turn it on, I won’t get anywhere close to what I want”. My interest just… how should I say it… completely dropped. So I couldn’t turn it on. I was scared—scared that I’d have to face that feeling again. So I kept my distance from it. 🐨: But this time, starting in Korea and then going on tour, meeting TEUMEs a lot as we move through 2025 and into 2026, without even realizing it, I felt refreshed. Like I’d been aired out and I thought, “Huh? should I try again?”, “I kind of want to do it again”. “I want to go back to when I really enjoyed this”. “I want to try again—the thing I loved back then. So I finally made up my mind and turned it on. and when I did… it had been so long that I got chills. I’d forgotten everything—the details, the keys, everything. “How did I even do this before?”, “What values did I use?” I’d forgotten it all. 🐨: I completely panicked. I thought, “Oh… is this how it ends?”, “Is this how I lose the thing I love?” I was honestly really scared. And then another thought came to me “turn a crisis into an opportunity”. Maybe this is a chance for me to find something else I love. Maybe I should let it go. I was almost halfway in a state of giving up. But somehow, my body followed through anyway. My hands kept moving—on their own. Somehow, I knew what to do, how to do it. and naturally, without stopping, the flow didn’t break. So maybe I want to stay with music for a really long time in my life. And I think this process of slowing myself down a bit was part of that. Looking back now, the pace had been way too fast, and I couldn’t control myself. I’d pushed myself to the limit of what I could create, and after that, there was nothing left. 🐨: At that point, I was kind of cruel to myself. I blamed myself for everything… It’s a really bad habit, I think, but it just happens reflexively. So I was really hard on myself. That’s when I realized I needed to take a step back, give myself some distance, and look at things over a longer period of time. Yeah… if you just move at your own pace—not faster, not slower—you can do something for a long time. So I think if you want to spend your life with something you really love—whatever it is—you can’t go too fast. If your passion just burns up too quickly, it can cause problems. That’s why pacing yourself matters. If you manage your pace, you can live your life alongside the things you love for a long, long time. I really felt it this time. And because of that, I was so happy. Being able to do what I love again brought back so many memories, and it felt like I’d returned to those days when I used to enjoy it so purely. That made me incredibly happy.

Junkyu being open about his burnout and how he overcame that phase is something that truly amazed me. His beautiful words that encourage everyone why pacing yourself actually matters 🥺🤍 🐨: I actually went through myself a burnout recently… was it in 2025? It’s not that I started to hate music, but when it came to making music, to working on it, I hit burnout. I really did. Nothing felt fun anymore, and no matter what I did, I just couldn’t move forward. No matter what—seriously, no matter what. 🐨: I work on music on my Mac, right? and I didn’t turn it on for a whole year. A year? Maybe even a year and a half? Since sometime in late 2024? It’s not that I chose not to turn it on—I couldn’t. That thought kept coming to me, “Even if I turn it on, nothing will come out” or maybe, “Even if I turn it on, I won’t get anywhere close to what I want”. My interest just… how should I say it… completely dropped. So I couldn’t turn it on. I was scared—scared that I’d have to face that feeling again. So I kept my distance from it. 🐨: But this time, starting in Korea and then going on tour, meeting TEUMEs a lot as we move through 2025 and into 2026, without even realizing it, I felt refreshed. Like I’d been aired out and I thought, “Huh? should I try again?”, “I kind of want to do it again”. “I want to go back to when I really enjoyed this”. “I want to try again—the thing I loved back then. So I finally made up my mind and turned it on. and when I did… it had been so long that I got chills. I’d forgotten everything—the details, the keys, everything. “How did I even do this before?”, “What values did I use?” I’d forgotten it all. 🐨: I completely panicked. I thought, “Oh… is this how it ends?”, “Is this how I lose the thing I love?” I was honestly really scared. And then another thought came to me “turn a crisis into an opportunity”. Maybe this is a chance for me to find something else I love. Maybe I should let it go. I was almost halfway in a state of giving up. But somehow, my body followed through anyway. My hands kept moving—on their own. Somehow, I knew what to do, how to do it. and naturally, without stopping, the flow didn’t break. So maybe I want to stay with music for a really long time in my life. And I think this process of slowing myself down a bit was part of that. Looking back now, the pace had been way too fast, and I couldn’t control myself. I’d pushed myself to the limit of what I could create, and after that, there was nothing left. 🐨: At that point, I was kind of cruel to myself. I blamed myself for everything… It’s a really bad habit, I think, but it just happens reflexively. So I was really hard on myself. That’s when I realized I needed to take a step back, give myself some distance, and look at things over a longer period of time. Yeah… if you just move at your own pace—not faster, not slower—you can do something for a long time. So I think if you want to spend your life with something you really love—whatever it is—you can’t go too fast. If your passion just burns up too quickly, it can cause problems. That’s why pacing yourself matters. If you manage your pace, you can live your life alongside the things you love for a long, long time. I really felt it this time. And because of that, I was so happy. Being able to do what I love again brought back so many memories, and it felt like I’d returned to those days when I used to enjoy it so purely. That made me incredibly happy.

ᯓ★

43,569 views • 6 months ago

If you watch this ~50 minute screen recording closely (yeah, I know, it's long; there are also some times when my computer was very slow and laggy, just skip past that part. And at one point I had to run and get my 9-month-old a new bottle and left it on a boring screen, sorry!), I believe you can see real signs of the kind of runaway, recursive AI self-improvement that people have been warning of for a while (Mr. Kurzweil most notably and prophetically). Why do I say that? What's different now? Well, there's a reason my set of agent coding tooling is called the Flywheel. These tools all mutually self-reinforce each other. And they all flow directly into my ntm tool (short for "named_tmux_manager"), which acts as a sort of integration point and nerve center for the tools (this is becoming more true by the minute as I'm now seriously working on ntm). Now, ntm was something I started making to automate some aspects of my workflow, but it was the kind of thing where, until it was perfect, it sort of just slowed me down. So I didn't actually use it even though I kept working on it and trying to improve it, and suggested to users that they try it in my tutorials. Well anyway, I finally got around to "dogfooding" ntm last night, and now it's going to get very dramatically better at an alarming rate. Some of that is from applying my "idea wizard" prompt to generate more useful features and building that stuff out and addressing obvious pain points I encountered during my newfound usage of the tool. But a lot comes from my realization that, once again, ntm's true utility is not as a tool for ME, but for an agent. That is, ntm lets one instance of Claude Code or Codex act as, well, me, do the things that I had been doing manually. Do I wish I had started using ntm earlier? No, for two big reasons: 1) Doing it manually helped me build up my intuition massively, which directly led me down the path of creating useful prompt strategies and workflows; these often began as ad-hoc prompts that I realized could be generalized and made more versatile/universal. Lesson: don't prematurely automate until you have an intimate, intuitive feel for your "core value-add loop." Otherwise you'll have a fully automated system quickly that efficiently and automatically does a stupid or otherwise sub-optimal thing. 2) My eyes have been opened to the beauty and power of Skills. I'm not talking about your garden-variety skills that are just a simple markdown file. I'm talking about true tour-de-force directories of perfectly structured and organized files that are filled with good information, insights, workflows, etc., but presented in a way that is highly optimized for consumption by AI agents, with extreme attention paid to things like perfect progressive disclosure, token density, agent-ergonomics, agent-intuitiveness, etc. And also Skills that go way beyond markdown files, with full integration into Claude Code where it makes sense via hooks, sub-agents, and even Python scripts. These kinds of skills are a qualitative difference in expressive power and usefulness and a total game changer. They are also effectively composable, creating almost an algebra of skills that let you use them together in powerful ways. I'm working on a subscription service website and CLI tool now to share what I've learned here most effectively, stay tuned for that in the coming days. Anyway, I now know what to make and how to make it. So, getting back to that screen recording, what does it show that makes me claim recursive self-improvement is here? If you keep your eye on the upper left tmux pane, that's the "controller" agent. It is using ntm to control all the other panes which are also running Claude Code (but ntm fully supports other agent types like Codex and Gemini-CLI, and it's trivially easy to mix and match them if you wanted to have, say, 8 CCs and 6 Codexes for writing the code and 3 Gemini-CLIs for reviewing code.) Now, there's nothing that crazy about this much so far. But where it starts to get very cool is that as the session continues and we encounter real-world problems, things like my ridiculously overloaded computer that keeps hanging for long periods, Claude Code instances that crash and get into a frozen, unresponsive state, it can learn from that. And you can see it using my skill writing skill to refine its ntm vibe coding skill in real time. And then take that skill and refine it to be more intuitive for itself. Or use my cass tool skill to search all the session histories to look for problems that came up and strategize how to solve them. The most useful part was when, towards the end of the session, I told it to reflect on all the things we had done and problems we encountered. One way it can usefully leverage those reflections is by improving its ntm vibe coding skill to make it cover more edge cases and exigencies. But the other, more fundamental, way is for it to conceive of and design the optimal new features and functionality for ntm itself so that the tool embodies those lessons in a first-class way. This offloads cognition from its brain onto its tooling, just like how a person can lean on spellcheck or a calculator. It codifies correct, effective reasoning at the tool level, where it's more reliable and robust and repeatable. And btw, did you notice what code base it was working on the whole time? It was none other than ntm itself! So as it worked on its own tool, it had reflections and ideas about how to further improve the tool. Now, it could have just as easily gotten those insights and ideas while using ntm to work on a different project, but the fact that it was working on itself is almost gloriously meta and recursive. So by the end, after learning from tending to a big group of agent workers (btw, I have previously emphasized doing everything in a really distributed/decentralized way, where each fungible agent gets identical marching orders that tell it to use my bv tool to find the optimal bead to work on. This does work very well, but occasionally results in some contention and overlap from thundering herd, or at least wastes time/tokens/communication in avoiding that before the agents waste time duplicating work. But in this new ntm-oriented workflow, I was able to have the controller agent in the upper left use bv itself and then optimally parcel out the instructions to each agent so that we could know for sure that there's no overlap), I ended up with a ton of new beads for new features, which I had it optimize and polish a few times. Now I can swap to a new Claude Max account and have the swarm implement all those new features! It should only take a couple passes like the one shown in the screen recording to get everything implemented. Then we can rinse and repeat, having the agent read through the full session histories of each agent and its experience from its own session in sending ntm commands and seeing how they worked out in practice, to come up with the next batch of changes to both its ntm vibe coding skill AND to the ntm tool itself. Do you see how rapidly this turns into Skynet? My mistake earlier was in focusing on making myself a "faster horse" as Henry Ford used to joke about customers wanting before he showed them what they should really want (a Model T). That is, something that would make my experience nicer while doing this agent swarm based development workflow. But the obvious lesson is that you should make all your tooling agent-first because the agents are just better at this stuff. You can still watch, and of course I did add a ridiculous number of very nice human-centric features to ntm that you'll be seeing in the next day or two, but those are really kind of "for fun" to make us humans feel better about the process. All the real value-add is happening "by agents, for agents." PS: Towards the end, you can see me switch to my Mac and tell Claude to improve the skill that I made earlier today for taking the mkv screen recording files from OBS Studio and muxing them into MP4 files for sharing, while downloading songs from YouTube to serve as the background music. I made it so it can also grab the thumbnails and generate little song credit cards that show up in the lower right corner. This worked perfectly the first time! I'll include some screenshots in a response post showing how that worked, but it was awesome to witness. Skills are POWERFUL. I'll also post a link to this video on YouTube if you prefer to watch it there.

If you watch this ~50 minute screen recording closely (yeah, I know, it's long; there are also some times when my computer was very slow and laggy, just skip past that part. And at one point I had to run and get my 9-month-old a new bottle and left it on a boring screen, sorry!), I believe you can see real signs of the kind of runaway, recursive AI self-improvement that people have been warning of for a while (Mr. Kurzweil most notably and prophetically). Why do I say that? What's different now? Well, there's a reason my set of agent coding tooling is called the Flywheel. These tools all mutually self-reinforce each other. And they all flow directly into my ntm tool (short for "named_tmux_manager"), which acts as a sort of integration point and nerve center for the tools (this is becoming more true by the minute as I'm now seriously working on ntm). Now, ntm was something I started making to automate some aspects of my workflow, but it was the kind of thing where, until it was perfect, it sort of just slowed me down. So I didn't actually use it even though I kept working on it and trying to improve it, and suggested to users that they try it in my tutorials. Well anyway, I finally got around to "dogfooding" ntm last night, and now it's going to get very dramatically better at an alarming rate. Some of that is from applying my "idea wizard" prompt to generate more useful features and building that stuff out and addressing obvious pain points I encountered during my newfound usage of the tool. But a lot comes from my realization that, once again, ntm's true utility is not as a tool for ME, but for an agent. That is, ntm lets one instance of Claude Code or Codex act as, well, me, do the things that I had been doing manually. Do I wish I had started using ntm earlier? No, for two big reasons: 1) Doing it manually helped me build up my intuition massively, which directly led me down the path of creating useful prompt strategies and workflows; these often began as ad-hoc prompts that I realized could be generalized and made more versatile/universal. Lesson: don't prematurely automate until you have an intimate, intuitive feel for your "core value-add loop." Otherwise you'll have a fully automated system quickly that efficiently and automatically does a stupid or otherwise sub-optimal thing. 2) My eyes have been opened to the beauty and power of Skills. I'm not talking about your garden-variety skills that are just a simple markdown file. I'm talking about true tour-de-force directories of perfectly structured and organized files that are filled with good information, insights, workflows, etc., but presented in a way that is highly optimized for consumption by AI agents, with extreme attention paid to things like perfect progressive disclosure, token density, agent-ergonomics, agent-intuitiveness, etc. And also Skills that go way beyond markdown files, with full integration into Claude Code where it makes sense via hooks, sub-agents, and even Python scripts. These kinds of skills are a qualitative difference in expressive power and usefulness and a total game changer. They are also effectively composable, creating almost an algebra of skills that let you use them together in powerful ways. I'm working on a subscription service website and CLI tool now to share what I've learned here most effectively, stay tuned for that in the coming days. Anyway, I now know what to make and how to make it. So, getting back to that screen recording, what does it show that makes me claim recursive self-improvement is here? If you keep your eye on the upper left tmux pane, that's the "controller" agent. It is using ntm to control all the other panes which are also running Claude Code (but ntm fully supports other agent types like Codex and Gemini-CLI, and it's trivially easy to mix and match them if you wanted to have, say, 8 CCs and 6 Codexes for writing the code and 3 Gemini-CLIs for reviewing code.) Now, there's nothing that crazy about this much so far. But where it starts to get very cool is that as the session continues and we encounter real-world problems, things like my ridiculously overloaded computer that keeps hanging for long periods, Claude Code instances that crash and get into a frozen, unresponsive state, it can learn from that. And you can see it using my skill writing skill to refine its ntm vibe coding skill in real time. And then take that skill and refine it to be more intuitive for itself. Or use my cass tool skill to search all the session histories to look for problems that came up and strategize how to solve them. The most useful part was when, towards the end of the session, I told it to reflect on all the things we had done and problems we encountered. One way it can usefully leverage those reflections is by improving its ntm vibe coding skill to make it cover more edge cases and exigencies. But the other, more fundamental, way is for it to conceive of and design the optimal new features and functionality for ntm itself so that the tool embodies those lessons in a first-class way. This offloads cognition from its brain onto its tooling, just like how a person can lean on spellcheck or a calculator. It codifies correct, effective reasoning at the tool level, where it's more reliable and robust and repeatable. And btw, did you notice what code base it was working on the whole time? It was none other than ntm itself! So as it worked on its own tool, it had reflections and ideas about how to further improve the tool. Now, it could have just as easily gotten those insights and ideas while using ntm to work on a different project, but the fact that it was working on itself is almost gloriously meta and recursive. So by the end, after learning from tending to a big group of agent workers (btw, I have previously emphasized doing everything in a really distributed/decentralized way, where each fungible agent gets identical marching orders that tell it to use my bv tool to find the optimal bead to work on. This does work very well, but occasionally results in some contention and overlap from thundering herd, or at least wastes time/tokens/communication in avoiding that before the agents waste time duplicating work. But in this new ntm-oriented workflow, I was able to have the controller agent in the upper left use bv itself and then optimally parcel out the instructions to each agent so that we could know for sure that there's no overlap), I ended up with a ton of new beads for new features, which I had it optimize and polish a few times. Now I can swap to a new Claude Max account and have the swarm implement all those new features! It should only take a couple passes like the one shown in the screen recording to get everything implemented. Then we can rinse and repeat, having the agent read through the full session histories of each agent and its experience from its own session in sending ntm commands and seeing how they worked out in practice, to come up with the next batch of changes to both its ntm vibe coding skill AND to the ntm tool itself. Do you see how rapidly this turns into Skynet? My mistake earlier was in focusing on making myself a "faster horse" as Henry Ford used to joke about customers wanting before he showed them what they should really want (a Model T). That is, something that would make my experience nicer while doing this agent swarm based development workflow. But the obvious lesson is that you should make all your tooling agent-first because the agents are just better at this stuff. You can still watch, and of course I did add a ridiculous number of very nice human-centric features to ntm that you'll be seeing in the next day or two, but those are really kind of "for fun" to make us humans feel better about the process. All the real value-add is happening "by agents, for agents." PS: Towards the end, you can see me switch to my Mac and tell Claude to improve the skill that I made earlier today for taking the mkv screen recording files from OBS Studio and muxing them into MP4 files for sharing, while downloading songs from YouTube to serve as the background music. I made it so it can also grab the thumbnails and generate little song credit cards that show up in the lower right corner. This worked perfectly the first time! I'll include some screenshots in a response post showing how that worked, but it was awesome to witness. Skills are POWERFUL. I'll also post a link to this video on YouTube if you prefer to watch it there.

Jeffrey Emanuel

25,483 views • 7 months ago