正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

Let’s build interactive Three.js User Interfaces 🫰 Uses react-three/handle, react-three/xr, and react-three/uikit - vanilla versions are available🍦 Works on any device inside and outside #WebXR Compose <uikit> with <handle> to control sizes, positions, rotations (Link in 🧵)

Bela Bohlender

2,041 subscribers

47,422 次观看 • 1 年前 •via X (Twitter)

科学技术教育 #WebXR

Anya Rossi• Live Now

Private livecam show

10 条评论

Bela Bohlender 的头像

Bela Bohlender1 年前

Try it out yourself: Source Code: What is uikit? What is handle? How is this running in XR?

Bela Bohlender 的头像

Bela Bohlender1 年前

Background music by Rockot from Pixabay Trying to make it a regular habit of donating to the musicians I use music from for my social media videos 💰

UserInterface 的头像

UserInterface2 年前

What is Web3? The Next Evolution of the Internet #tech #news

Nestor Rios Garcia 的头像

Nestor Rios Garcia1 年前

@threejs Magnificent Bela. I'll try it soon 🔥.

Stijn Spanhove 的头像

Stijn Spanhove1 年前

@threejs Nice! Thanks for the examples 👀

Fred King 的头像

Fred King1 年前

@threejs Worked on my Samsung flip 5! Nice.

Tidewater 的头像

Tidewater1 年前

@threejs My goodness that is quite flavorful

Morten Haulik ᯅ 的头像

Morten Haulik ᯅ1 年前

@threejs Awesome🔥

Remus 的头像

Remus1 年前

@threejs Does this work outside of XR environments? Meaning can you use this threeJS component in a regular webpage?

Bela Bohlender 的头像

Bela Bohlender1 年前

@threejs sure, you can try it out yourself here

相关视频

Shipping react-three/handle! The one to handle them all! 👌 Build intuitive 3D and XR cross-platform interactions with just one line of code. #R3F Three.js #WebXR

Shipping react-three/handle! The one to handle them all! 👌 Build intuitive 3D and XR cross-platform interactions with just one line of code. #R3F Three.js #WebXR

Bela Bohlender

18,632 次观看 • 1 年前

<mesh> <FakeGlowMaterial /> </mesh> A new component is coming. A very simple, easy-to-use, fake glow material. This can create a glow around the mesh without any post-processing. Available soon for vanilla #threejs and #react three fiber 😍 #threejsJourney

<mesh> <FakeGlowMaterial /> </mesh> A new component is coming. A very simple, easy-to-use, fake glow material. This can create a glow around the mesh without any post-processing. Available soon for vanilla #threejs and #react three fiber 😍 #threejsJourney

Anderson Mancini

26,380 次观看 • 2 年前

Excited to launch react-three/xr🤳🥽 Leverage the latest WebXR features, reuse your existing interactions, and advance what is possible in AR and VR on the Web. ↳ Turn your existing R3F app into an AR or VR experience with only a few lines of code. 🎉

Excited to launch react-three/xr🤳🥽 Leverage the latest WebXR features, reuse your existing interactions, and advance what is possible in AR and VR on the Web. ↳ Turn your existing R3F app into an AR or VR experience with only a few lines of code. 🎉

Bela Bohlender

40,117 次观看 • 2 年前

Opus 4.6 vs Gemini 3 Pro (3/9) prompt: Build a premium interactive isometric 3D cozy room using Vite + React + Three.js (react-three-fiber/drei), with all objects modeled in code (no external assets) and subtle ambient animations. Gemini looks good on initial impression but I feel like I have way more to work with opus here (opus also did light/dark mode and took a quarter of the time to make it).

Opus 4.6 vs Gemini 3 Pro (3/9) prompt: Build a premium interactive isometric 3D cozy room using Vite + React + Three.js (react-three-fiber/drei), with all objects modeled in code (no external assets) and subtle ambient animations. Gemini looks good on initial impression but I feel like I have way more to work with opus here (opus also did light/dark mode and took a quarter of the time to make it).

Dev Ed

77,206 次观看 • 5 个月前

🚀 I'm launching r3f-webgpu-perf today! "npm i r3f-webgpu-perf" One thing I missed while working with WebGPU in React Three Fiber was a performance monitor similar to r3f-perf. The existing tool is fantastic, but it doesn’t support the WebGPU renderer yet. So I decided to build one. ✅ Supports both WebGL and WebGPU ✅ Modern and customizable UI ✅ Lightweight and easy to integrate ✅ Built specifically for React Three Fiber My goal is to turn r3f-webgpu-perf into more than just a metrics dashboard. I want it to become a learning tool that provides real-time optimization suggestions based on the techniques I’ve developed while building hundreds of interactive 3D projects over the last two decades. A vanilla Three.js version is also planned. If you’re building with Three.js, React Three Fiber, or WebGPU, I’d love for you to try it and share your feedback. ⭐ GitHub: 📦 NPM: Every GitHub star helps make the tool better for the entire community. #threejs #creativecoding #webgpu #r3f

🚀 I'm launching r3f-webgpu-perf today! "npm i r3f-webgpu-perf" One thing I missed while working with WebGPU in React Three Fiber was a performance monitor similar to r3f-perf. The existing tool is fantastic, but it doesn’t support the WebGPU renderer yet. So I decided to build one. ✅ Supports both WebGL and WebGPU ✅ Modern and customizable UI ✅ Lightweight and easy to integrate ✅ Built specifically for React Three Fiber My goal is to turn r3f-webgpu-perf into more than just a metrics dashboard. I want it to become a learning tool that provides real-time optimization suggestions based on the techniques I’ve developed while building hundreds of interactive 3D projects over the last two decades. A vanilla Three.js version is also planned. If you’re building with Three.js, React Three Fiber, or WebGPU, I’d love for you to try it and share your feedback. ⭐ GitHub: 📦 NPM: Every GitHub star helps make the tool better for the entire community. #threejs #creativecoding #webgpu #r3f

Anderson Mancini

23,332 次观看 • 1 个月前

BADLAB V2: WHERE YOUR MEDIOCRE CODING DREAMS GO TO DIE 🔊 Listen up, code peasants. While you've been copy-pasting Stack Overflow answers, we've been revolutionizing how fucking development works. BADLAB V2 just dropped, and it's not just another update – it's a fucking paradigm shift. Let me break this down for you smooth-brains: 1️⃣ BADLAB V2 makes it easier and faster to build web apps, games, smart contracts, and data visualizations: No-code deployment, run, and share - without the headaches. 2️⃣ What This Means for Users: You can build bigger, better, and more useful projects and tools. Works with all the tools you love: React, Web3, Three.js, d3, and more—whether you’re building games, dApps, automation tools, or interactive charts, the sky is the limit. 3️⃣ Smart automation does the hard work: AI writes code for you, making it easier to start building without deep technical expertise. 4️⃣ CDN-backed hosting: Your projects load fast and reliably because they run from Cloudflare CDN, not some sketchy server. And now? You can share your creations with friends. 5️⃣ Easier debugging & updates: No more jumping through hoops to update files or fix errors. 6️⃣ Bottom Line: If you want to build web apps, games, or blockchain projects without dealing with all the messy setup and technical headaches, BADLAB V2 makes it easy. 🆕 WHAT’S NEW: ✔️ Multi-file projects – Because one file isn’t enough for real developers. ✔️ File updates that don’t require a PhD – No unnecessary complexity. ✔️ Examples even a brain-dead monkey could follow – Clear and simple. ✔️ Support for every major file format – JS, JSON, Markdown, 3D models, and more. ✔️ React components that don’t look like digital vomit – Finally, code that makes sense. ✔️ Server-side bundling – Because we’re not savages. ✔️ CDN-backed deployment – No lag, no downtime, just pure speed. ⚠️ WEAPONS OF MASS CREATION: 💧 d3.js – Interactive charts that don’t look like Excel’s morning sickness. 💧 Three.js – 3D visuals that push browsers to their limits. 💧 Effector – Business logic that actually makes fucking sense. 💧 Kaboom.js – Build browser games that aren’t complete garbage. 💧 React-Three/Fiber – 3D meets React without the usual drama. 💧 Framer Motion – Animations that are smooth as hell (and don’t cause seizures). 💧 Ethers & Web3.js – Blockchain development for the mentally stable. 🎮 Want to See It in Action? 🔗 Try our BADLAB V2 Agent Here: – Let AI do the hard work for you. ♟ Play the 4D Chess Game Here: – While you're playing checkers, we're transcending dimensions. 🎙 Welcome to the future, you beautiful disasters. Try not to break anything important. 🧪💀 C₈H₁₁NO₂

BADLAB V2: WHERE YOUR MEDIOCRE CODING DREAMS GO TO DIE 🔊 Listen up, code peasants. While you've been copy-pasting Stack Overflow answers, we've been revolutionizing how fucking development works. BADLAB V2 just dropped, and it's not just another update – it's a fucking paradigm shift. Let me break this down for you smooth-brains: 1️⃣ BADLAB V2 makes it easier and faster to build web apps, games, smart contracts, and data visualizations: No-code deployment, run, and share - without the headaches. 2️⃣ What This Means for Users: You can build bigger, better, and more useful projects and tools. Works with all the tools you love: React, Web3, Three.js, d3, and more—whether you’re building games, dApps, automation tools, or interactive charts, the sky is the limit. 3️⃣ Smart automation does the hard work: AI writes code for you, making it easier to start building without deep technical expertise. 4️⃣ CDN-backed hosting: Your projects load fast and reliably because they run from Cloudflare CDN, not some sketchy server. And now? You can share your creations with friends. 5️⃣ Easier debugging & updates: No more jumping through hoops to update files or fix errors. 6️⃣ Bottom Line: If you want to build web apps, games, or blockchain projects without dealing with all the messy setup and technical headaches, BADLAB V2 makes it easy. 🆕 WHAT’S NEW: ✔️ Multi-file projects – Because one file isn’t enough for real developers. ✔️ File updates that don’t require a PhD – No unnecessary complexity. ✔️ Examples even a brain-dead monkey could follow – Clear and simple. ✔️ Support for every major file format – JS, JSON, Markdown, 3D models, and more. ✔️ React components that don’t look like digital vomit – Finally, code that makes sense. ✔️ Server-side bundling – Because we’re not savages. ✔️ CDN-backed deployment – No lag, no downtime, just pure speed. ⚠️ WEAPONS OF MASS CREATION: 💧 d3.js – Interactive charts that don’t look like Excel’s morning sickness. 💧 Three.js – 3D visuals that push browsers to their limits. 💧 Effector – Business logic that actually makes fucking sense. 💧 Kaboom.js – Build browser games that aren’t complete garbage. 💧 React-Three/Fiber – 3D meets React without the usual drama. 💧 Framer Motion – Animations that are smooth as hell (and don’t cause seizures). 💧 Ethers & Web3.js – Blockchain development for the mentally stable. 🎮 Want to See It in Action? 🔗 Try our BADLAB V2 Agent Here: – Let AI do the hard work for you. ♟ Play the 4D Chess Game Here: – While you're playing checkers, we're transcending dimensions. 🎙 Welcome to the future, you beautiful disasters. Try not to break anything important. 🧪💀 C₈H₁₁NO₂

ARCH AI

44,480 次观看 • 1 年前

$Two new #threejs projects on the way 😍 Working on my own gem renderer with a custom refraction material in TSL for #WebGPU, combining Screen Space Reflections + SSGTAO for superior results, as this allows diamonds to generate reflections on the metal mesh. The performance is incredible, even on gems with over 200 diamonds. This time, without any additional libraries. Just React Three Fiber and a lot of magic with custom shaders in TSL. GemGPU™ will also come with a very minimalist gem and metal customization panel that hides and lets the jewelry be the star. I also created a beautiful new stats panel that completely hides in the scene when not needed. It provides relevant optimization information, a performance gauge, summary and expanded modes, and day/night toggle. In the video, you can also see SSGI in action, with a pearl that reflects, in real time, all the objects in the scene, just as the objects in the scene reflect the pearl. It's magical. There are still a few things to fix, but I'll share the live link soon. Stay tuned.$

Two new #threejs projects on the way 😍 Working on my own gem renderer with a custom refraction material in TSL for #WebGPU, combining Screen Space Reflections + SSGTAO for superior results, as this allows diamonds to generate reflections on the metal mesh. The performance is incredible, even on gems with over 200 diamonds. This time, without any additional libraries. Just React Three Fiber and a lot of magic with custom shaders in TSL. GemGPU™ will also come with a very minimalist gem and metal customization panel that hides and lets the jewelry be the star. I also created a beautiful new stats panel that completely hides in the scene when not needed. It provides relevant optimization information, a performance gauge, summary and expanded modes, and day/night toggle. In the video, you can also see SSGI in action, with a pearl that reflects, in real time, all the objects in the scene, just as the objects in the scene reflect the pearl. It's magical. There are still a few things to fix, but I'll share the live link soon. Stay tuned.

Anderson Mancini

12,113 次观看 • 5 个月前

This bot started with tiny $5 orders - and somehow pulled in over $200,000 on the 5-minute market. In just one week, it’s already cleared more than +$212,578 And the biggest single win? Only $745. But the wild part isn’t the size of any one trade - it’s how those tiny positions stack up into such an aggressive profit curve. From the outside, it looks like it’s running three simple plays: - Uses $5 to grab 1¢ outcomes - Exits early before expiration to lock in profit - Scales up to $250 on 25¢–50¢ positions Then pushes further with $500 into 50¢ outcomes. So this doesn’t look like a typical “elite” trading bot squeezing every edge with precision. It feels more like something is off inside Polymarket - creating these exaggerated, almost unnatural returns. And yet, right now, this account sits at #2 in weekly crypto profits. If it’s really hitting profitable trades this consistently - that’s not just a bot. That’s basically a money printer.

This bot started with tiny $5 orders - and somehow pulled in over $200,000 on the 5-minute market. In just one week, it’s already cleared more than +$212,578 And the biggest single win? Only $745. But the wild part isn’t the size of any one trade - it’s how those tiny positions stack up into such an aggressive profit curve. From the outside, it looks like it’s running three simple plays: - Uses $5 to grab 1¢ outcomes - Exits early before expiration to lock in profit - Scales up to $250 on 25¢–50¢ positions Then pushes further with $500 into 50¢ outcomes. So this doesn’t look like a typical “elite” trading bot squeezing every edge with precision. It feels more like something is off inside Polymarket - creating these exaggerated, almost unnatural returns. And yet, right now, this account sits at #2 in weekly crypto profits. If it’s really hitting profitable trades this consistently - that’s not just a bot. That’s basically a money printer.

Dekos

19,488 次观看 • 2 个月前

In the early days of Waymo's service in Phoenix, they had a large number of roadside assistance vans traveling alongside their vehicles, each with two people inside. When the Waymo vehicle got stuck, one person would jump out of the RA van and drive the car. Today Waymo still has roadside assistance vans ready to help when a car has an issue, but they don't have to be directly with one car — issues are rare enough that a few roadside assistance vans can handle a much larger fleet. It's funny to see people take shots at Tesla for doing the same. Of course they're going to be careful with the first customer deployment. No, someone isn't secretly controlling the car from outside. The AI model on the car is driving the car. Go test drive any Tesla and find out for yourself. It doesn't need help from anyone, it can see and react to things before you do.

In the early days of Waymo's service in Phoenix, they had a large number of roadside assistance vans traveling alongside their vehicles, each with two people inside. When the Waymo vehicle got stuck, one person would jump out of the RA van and drive the car. Today Waymo still has roadside assistance vans ready to help when a car has an issue, but they don't have to be directly with one car — issues are rare enough that a few roadside assistance vans can handle a much larger fleet. It's funny to see people take shots at Tesla for doing the same. Of course they're going to be careful with the first customer deployment. No, someone isn't secretly controlling the car from outside. The AI model on the car is driving the car. Go test drive any Tesla and find out for yourself. It doesn't need help from anyone, it can see and react to things before you do.

Whole Mars Catalog

95,641 次观看 • 5 个月前

$Data encryption is now live on Veilnet. Available in @veilnet/mcp-server@0.3.0. Any blob — file, dataset, config, recovery codes, contract draft, research note, agent memory — can now be encrypted on your device, shipped to Veilnet as ciphertext, and fetched back later without the decryption key ever leaving your machine. Three new MCP tools are live: data_store({ label, data }) encrypt + store data on Veilnetdata_fetch({ id }) fetch + decrypt locallydata_list() list stored ids With labels kept encrypted, the flow is simple: user data → encrypted locally → ciphertext stored on Veilnet → key remains on user device → agent fetches when needed → local decryption onlyVeilnet sees encrypted blobs. Your AI agent gets persistent private memory and secure storage.The provider does not need the raw data. This works in Claude Code, Claude Desktop, Cursor, Continue, and Cline. Install: npm i -g @veilnet/mcp-server Bootstrap your key at: → Agent tab Then ask any MCP-aware agent:“store this encrypted on Veilnet.”Docs:docs.veilnet.toPrivate prompts protected what users ask. Private MCP protected what agents run. Wallet Context protected what agents read. Encrypted data storage now protects what agents remember and reuse. All powered by $VEILNET$

Data encryption is now live on Veilnet. Available in @veilnet/[email protected]. Any blob — file, dataset, config, recovery codes, contract draft, research note, agent memory — can now be encrypted on your device, shipped to Veilnet as ciphertext, and fetched back later without the decryption key ever leaving your machine. Three new MCP tools are live: data_store({ label, data }) encrypt + store data on Veilnetdata_fetch({ id }) fetch + decrypt locallydata_list() list stored ids With labels kept encrypted, the flow is simple: user data → encrypted locally → ciphertext stored on Veilnet → key remains on user device → agent fetches when needed → local decryption onlyVeilnet sees encrypted blobs. Your AI agent gets persistent private memory and secure storage.The provider does not need the raw data. This works in Claude Code, Claude Desktop, Cursor, Continue, and Cline. Install: npm i -g @veilnet/mcp-server Bootstrap your key at: → Agent tab Then ask any MCP-aware agent:“store this encrypted on Veilnet.”Docs:docs.veilnet.toPrivate prompts protected what users ask. Private MCP protected what agents run. Wallet Context protected what agents read. Encrypted data storage now protects what agents remember and reuse. All powered by $VEILNET

Veilnet

11,759 次观看 • 1 个月前

AI is all about "knowing what you don't know" Here's an interactive exploration of the Andromeda: - React Three Fiber + Three.js - 600k particles (my m4 gets rly hot) - GSAP for smooth camera animations - Bloom post-processing for realistic stellar glow The more impressive piece here is that this was all through natural language prompting I didn't edit any code and I didn't know what GSAP / Bloom were until yesterday What I DID do was ask the right questions: - How do I get realistic "glow" and "feathering" from stars? - What is a shader? - How can I create smooth animations? - What does a minimalist UX look like? Then I did research (using AI) to understand them This "on demand learning" is a new concept made possible by the lower information latency presented by AI You still need to learn, but now learning is PART of the building process When I build apps like this it really reinforces the direction AI is headed: - Concepts > code - On demand learning > textbooks I'm quite excited for the future

AI is all about "knowing what you don't know" Here's an interactive exploration of the Andromeda: - React Three Fiber + Three.js - 600k particles (my m4 gets rly hot) - GSAP for smooth camera animations - Bloom post-processing for realistic stellar glow The more impressive piece here is that this was all through natural language prompting I didn't edit any code and I didn't know what GSAP / Bloom were until yesterday What I DID do was ask the right questions: - How do I get realistic "glow" and "feathering" from stars? - What is a shader? - How can I create smooth animations? - What does a minimalist UX look like? Then I did research (using AI) to understand them This "on demand learning" is a new concept made possible by the lower information latency presented by AI You still need to learn, but now learning is PART of the building process When I build apps like this it really reinforces the direction AI is headed: - Concepts > code - On demand learning > textbooks I'm quite excited for the future

matt palmer

17,327 次观看 • 8 个月前

AG-UI makes building agentic applications dramatically easier. Here's how it works. This is a model for a simple chatbot: User → LLM → Response But interactive agents that render UI, pause for approvals, and ask users for input need a much more complex model. When building these agents, a response from the LLM will include a series of state changes as the agent runs: • Agent started a task • Agent called a tool • Agent updated its state • Agent streams these tokens • Agent is waiting on a human • Agent is resuming the task The Agent-User Interaction Protocol (AG-UI) treats the LLM response as a stream of events rather than a text endpoint. In practice, here is what you get as an agent runs: 1. Lifecycle events so your UI knows where the agent is. 2. Text messages that stream tokens. 3. Tool calls so your UI can prefill a form with any required arguments. 4. State updates that keep your UI in sync with the agent. 5. Special events for human approvals, rich media, and custom needs. All of these events travel over standard transports (SSE, WebSockets, or plain HTTP) as JSON. As a result, you can build a frontend that stays in sync with the agent's progress without having to invent a custom process to make this happen. For example, building a human-in-the-loop workflow becomes an off-the-shelf component you can integrate rather than build from scratch. CopilotKit🪁 is the creator of AG-UI, and you can use it when building frontend applications pretty much anywhere: • React • Angular • Vue • React Native • Slack • Teams • Discord • WhatsApp • Telegram Here is the link for you to check it out: Thanks to the CopilotKit team for partnering with me on this post.

AG-UI makes building agentic applications dramatically easier. Here's how it works. This is a model for a simple chatbot: User → LLM → Response But interactive agents that render UI, pause for approvals, and ask users for input need a much more complex model. When building these agents, a response from the LLM will include a series of state changes as the agent runs: • Agent started a task • Agent called a tool • Agent updated its state • Agent streams these tokens • Agent is waiting on a human • Agent is resuming the task The Agent-User Interaction Protocol (AG-UI) treats the LLM response as a stream of events rather than a text endpoint. In practice, here is what you get as an agent runs: 1. Lifecycle events so your UI knows where the agent is. 2. Text messages that stream tokens. 3. Tool calls so your UI can prefill a form with any required arguments. 4. State updates that keep your UI in sync with the agent. 5. Special events for human approvals, rich media, and custom needs. All of these events travel over standard transports (SSE, WebSockets, or plain HTTP) as JSON. As a result, you can build a frontend that stays in sync with the agent's progress without having to invent a custom process to make this happen. For example, building a human-in-the-loop workflow becomes an off-the-shelf component you can integrate rather than build from scratch. CopilotKit🪁 is the creator of AG-UI, and you can use it when building frontend applications pretty much anywhere: • React • Angular • Vue • React Native • Slack • Teams • Discord • WhatsApp • Telegram Here is the link for you to check it out: Thanks to the CopilotKit team for partnering with me on this post.

Santiago

17,438 次观看 • 19 天前

Anthropic's most viral feature is now open-source! Until now, Anthropic's Generative UI capabilities only existed inside its own products. CopilotKit🪁 just shipped Open Generative UI, an open-source implementation of Claude Artifacts that works in any app. The agent generates HTML/SVG at runtime, and CopilotKit streams it token-by-token into a sandboxed iframe inside the app's chat. So the user can watch the UI assemble itself in real time, not after the full response is ready. The sandbox is fully isolated with no access to the parent app, the DOM, or user data. So if the agent hallucinates broken markup or unexpected JavaScript, nothing leaks outside the iframe. Under the hood, the agent does not select from pre-built components. Instead, it generates arbitrary visuals from scratch every time. The output is unconstrained by default, but you can shape it by defining prompt-based skills that teach the agent specific visual formats or guidelines. For instance, a skill prompt can guide the agent toward producing a Chart.js dashboard with proper axis labels and responsive sizing, or an interactive 3D model with rotation controls. The video below shows this in action, and the output quality you see actually comes from the skills layer. Open Generative UI runs on AG-UI, so it works out of the box with LangGraph, CrewAI, Mastra, Google ADK, AWS Strands, and more. It also ships with a standalone MCP server that plugs into Claude Code, Cursor, or any MCP-compatible client. And the entire stack is built on top of CopilotKit, the open-source frontend framework for agents and generative UI. 30k+ GitHub stars, with SDKs for React, Next.js, Angular, and Vue. I have shared the GitHub repo and a live playground in the replies!

Anthropic's most viral feature is now open-source! Until now, Anthropic's Generative UI capabilities only existed inside its own products. CopilotKit🪁 just shipped Open Generative UI, an open-source implementation of Claude Artifacts that works in any app. The agent generates HTML/SVG at runtime, and CopilotKit streams it token-by-token into a sandboxed iframe inside the app's chat. So the user can watch the UI assemble itself in real time, not after the full response is ready. The sandbox is fully isolated with no access to the parent app, the DOM, or user data. So if the agent hallucinates broken markup or unexpected JavaScript, nothing leaks outside the iframe. Under the hood, the agent does not select from pre-built components. Instead, it generates arbitrary visuals from scratch every time. The output is unconstrained by default, but you can shape it by defining prompt-based skills that teach the agent specific visual formats or guidelines. For instance, a skill prompt can guide the agent toward producing a Chart.js dashboard with proper axis labels and responsive sizing, or an interactive 3D model with rotation controls. The video below shows this in action, and the output quality you see actually comes from the skills layer. Open Generative UI runs on AG-UI, so it works out of the box with LangGraph, CrewAI, Mastra, Google ADK, AWS Strands, and more. It also ships with a standalone MCP server that plugs into Claude Code, Cursor, or any MCP-compatible client. And the entire stack is built on top of CopilotKit, the open-source frontend framework for agents and generative UI. 30k+ GitHub stars, with SDKs for React, Next.js, Angular, and Vue. I have shared the GitHub repo and a live playground in the replies!

Akshay 🚀

86,515 次观看 • 2 个月前

Star Fox Battle Mode is already getting called out for only having three maps. The people saying that are the same ones who burn out on every other multiplayer mode once the new map rotation drops. Three maps works here because the matches stay short, the objectives are actually worth fighting over, and there is enough crazy stuff happening that you are constantly battling the level itself as much as the other Arwings. Power-ups appear in risky positions. Environmental pressure changes during the round. One mistimed boost or bad angle through a corridor and the map punishes you before the enemy needs to finish the job. That constant back-and-forth between pilot skill and level hostility is what stops the same three maps from feeling repetitive after dozens of rounds. The Arwing handling underneath all of it is smooth and tight. Boost timing and lock-on precision actually matter instead of just holding forward and hoping. Small adjustments in speed and angle separate the players who are starting to read patterns from the ones still flying in straight lines. It looks sharp in motion too. The HDR makes explosions, laser fire, and environmental effects stay readable even when everything is happening at once. The sound design helps just as much in the chaos. Clear engine roar, distinct lock-on warnings, and impact feedback that cuts through the orchestral score so you can react to what is actually dangerous. The accolades system adds another layer without turning into a grind. It rewards the behaviours that make the mode interesting: smart objective play, power-up denial, aggressive but calculated lines through the map. There are enough unlocks tied to performance that you get steady progression and reasons to keep dropping in. This is the part most people pushing for more maps will not want to admit. In 2026 a lot of competitive modes assume they need constant new content and huge map pools just to hold attention. Star Fox Battle Mode is testing the opposite idea. A small set of dense, hostile playgrounds with short intense rounds and real skill expression can stay fresh longer than another bloated live service shooter that nobody ever truly masters. The map is an opponent here. That single design choice changes how much replayability three maps can actually deliver. What are your thoughts on Star Fox's online Battle Mode? After a proper session, do the three maps start feeling limiting or does the combination of short rounds, real objectives, and the environment constantly pushing back keep every match feeling different? p.s. Follow for more Star Fox and Nintendo takes! And if you’re jumping into Battle Mode tonight, add me: SW-4504-4427-4511

Star Fox Battle Mode is already getting called out for only having three maps. The people saying that are the same ones who burn out on every other multiplayer mode once the new map rotation drops. Three maps works here because the matches stay short, the objectives are actually worth fighting over, and there is enough crazy stuff happening that you are constantly battling the level itself as much as the other Arwings. Power-ups appear in risky positions. Environmental pressure changes during the round. One mistimed boost or bad angle through a corridor and the map punishes you before the enemy needs to finish the job. That constant back-and-forth between pilot skill and level hostility is what stops the same three maps from feeling repetitive after dozens of rounds. The Arwing handling underneath all of it is smooth and tight. Boost timing and lock-on precision actually matter instead of just holding forward and hoping. Small adjustments in speed and angle separate the players who are starting to read patterns from the ones still flying in straight lines. It looks sharp in motion too. The HDR makes explosions, laser fire, and environmental effects stay readable even when everything is happening at once. The sound design helps just as much in the chaos. Clear engine roar, distinct lock-on warnings, and impact feedback that cuts through the orchestral score so you can react to what is actually dangerous. The accolades system adds another layer without turning into a grind. It rewards the behaviours that make the mode interesting: smart objective play, power-up denial, aggressive but calculated lines through the map. There are enough unlocks tied to performance that you get steady progression and reasons to keep dropping in. This is the part most people pushing for more maps will not want to admit. In 2026 a lot of competitive modes assume they need constant new content and huge map pools just to hold attention. Star Fox Battle Mode is testing the opposite idea. A small set of dense, hostile playgrounds with short intense rounds and real skill expression can stay fresh longer than another bloated live service shooter that nobody ever truly masters. The map is an opponent here. That single design choice changes how much replayability three maps can actually deliver. What are your thoughts on Star Fox's online Battle Mode? After a proper session, do the three maps start feeling limiting or does the combination of short rounds, real objectives, and the environment constantly pushing back keep every match feeling different? p.s. Follow for more Star Fox and Nintendo takes! And if you’re jumping into Battle Mode tonight, add me: SW-4504-4427-4511

AussieGamer

23,360 次观看 • 24 天前

Small Language Models (SML) are the future of AI. "Small" (SML) instead of "Large" (LLM). These small models are highly specialized models with superhuman abilities on specific tasks. Here are two techniques to build these models: • Spectrum • Model Merging I give you a short introduction in the attached video, but here is a quick summary: Spectrum helps us identify the most relevant layers to solve one specific task. We can ignore everything else and focus on fine-tuning these layers. Using Spectrum, we can fine-tune models in a heartbeat. Model Merging combines multiple models into a unique, much better model than any of the individual input models. You can also combine models specialized in different tasks and get a model with multiple abilities. This is the state of the art of productizing models. It's what Arcee.ai's platform does behind the scenes. Arcee collaborated with me on this post and is sponsoring it. There are three main steps to produce a model for your particular use case: 1. You create a dataset by uploading your data. 2. You train a model. At this step, Arcee uses Spectrum and Model Merging to produce a highly specialized model for your task. 3. You can deploy that model to any environment you want. Three important notes: • Training process is 2x faster and 2x cheaper than regular fine-tuning. • Resultant models are smaller and have higher accuracy. • They create these specialized models from open-source models. Check this site so you can fully appreciate how this works: If you want to fine-tune an open-source model, consider Arcee's platform. This is the state of the art.

Small Language Models (SML) are the future of AI. "Small" (SML) instead of "Large" (LLM). These small models are highly specialized models with superhuman abilities on specific tasks. Here are two techniques to build these models: • Spectrum • Model Merging I give you a short introduction in the attached video, but here is a quick summary: Spectrum helps us identify the most relevant layers to solve one specific task. We can ignore everything else and focus on fine-tuning these layers. Using Spectrum, we can fine-tune models in a heartbeat. Model Merging combines multiple models into a unique, much better model than any of the individual input models. You can also combine models specialized in different tasks and get a model with multiple abilities. This is the state of the art of productizing models. It's what Arcee.ai's platform does behind the scenes. Arcee collaborated with me on this post and is sponsoring it. There are three main steps to produce a model for your particular use case: 1. You create a dataset by uploading your data. 2. You train a model. At this step, Arcee uses Spectrum and Model Merging to produce a highly specialized model for your task. 3. You can deploy that model to any environment you want. Three important notes: • Training process is 2x faster and 2x cheaper than regular fine-tuning. • Resultant models are smaller and have higher accuracy. • They create these specialized models from open-source models. Check this site so you can fully appreciate how this works: If you want to fine-tune an open-source model, consider Arcee's platform. This is the state of the art.

Santiago

164,162 次观看 • 2 年前

a team of researchers just proved you don't need a bigger model, you need a smarter plan researchers from Tsinghua and South China University of Technology built a framework called Atomic Task Graph. it turned 7B-8B open-source models into GPT-4 competitors on complex agent benchmarks, beating it on two out of three. no fine-tuning. no extra training. zero parameter updates. current AI agents plan in a straight line. step 1, step 2, step 3. when step 4 fails, the whole chain breaks. and the longer the chain gets, the more the model hallucinates because it's reasoning over a ballooning text history. here's how it works. 1. instead of a linear chain, ATG breaks any complex task into a directed graph where subtask inputs and outputs are explicitly mapped 2. it recursively decomposes each subtask until every node is one atomic tool call 3. independent branches run in parallel instead of waiting in line 4. before anything executes, a lightweight "thought experiment" simulates the plan internally to catch bad dependencies and missing steps early 5. when something breaks at runtime, ATG traces the failure to the exact subgraph that caused it and repairs only that piece. validated work stays frozen. the old way meant a failure at step 5 forced a full replan from scratch. hallucinated actions piled up the longer the task ran. ReAct hit a 43% hallucination rate on household tasks. ATG on an 8B Llama model scored 63.65 on ALFWorld. GPT-4 with ReAct scored 41.24 on the same benchmark. hallucinated actions dropped to 12%. those numbers happened because someone stopped throwing compute at the problem and started thinking about how work gets organized. that's the part that gets me. the industry is spending billions on scale. this team spent time on architecture. and the architecture won.

a team of researchers just proved you don't need a bigger model, you need a smarter plan researchers from Tsinghua and South China University of Technology built a framework called Atomic Task Graph. it turned 7B-8B open-source models into GPT-4 competitors on complex agent benchmarks, beating it on two out of three. no fine-tuning. no extra training. zero parameter updates. current AI agents plan in a straight line. step 1, step 2, step 3. when step 4 fails, the whole chain breaks. and the longer the chain gets, the more the model hallucinates because it's reasoning over a ballooning text history. here's how it works. 1. instead of a linear chain, ATG breaks any complex task into a directed graph where subtask inputs and outputs are explicitly mapped 2. it recursively decomposes each subtask until every node is one atomic tool call 3. independent branches run in parallel instead of waiting in line 4. before anything executes, a lightweight "thought experiment" simulates the plan internally to catch bad dependencies and missing steps early 5. when something breaks at runtime, ATG traces the failure to the exact subgraph that caused it and repairs only that piece. validated work stays frozen. the old way meant a failure at step 5 forced a full replan from scratch. hallucinated actions piled up the longer the task ran. ReAct hit a 43% hallucination rate on household tasks. ATG on an 8B Llama model scored 63.65 on ALFWorld. GPT-4 with ReAct scored 41.24 on the same benchmark. hallucinated actions dropped to 12%. those numbers happened because someone stopped throwing compute at the problem and started thinking about how work gets organized. that's the part that gets me. the industry is spending billions on scale. this team spent time on architecture. and the architecture won.

Alex Veremeyenko

172,462 次观看 • 9 天前

Alibaba presents MIMO Controllable Character Video Synthesis with Spatial Decomposed Modeling Character video synthesis aims to produce realistic videos of animatable characters within lifelike scenes. As a fundamental problem in the computer vision and graphics community, 3D works typically require multi-view captures for per-case training, which severely limits their applicability of modeling arbitrary characters in a short time. Recent 2D methods break this limitation via pre-trained diffusion models, but they struggle for pose generality and scene interaction. To this end, we propose MIMO, a novel framework which can not only synthesize character videos with controllable attributes (i.e., character, motion and scene) provided by simple user inputs, but also simultaneously achieve advanced scalability to arbitrary characters, generality to novel 3D motions, and applicability to interactive real-world scenes in a unified framework. The core idea is to encode the 2D video to compact spatial codes, considering the inherent 3D nature of video occurrence. Concretely, we lift the 2D frame pixels into 3D using monocular depth estimators, and decompose the video clip to three spatial components (i.e., main human, underlying scene, and floating occlusion) in hierarchical layers based on the 3D depth. These components are further encoded to canonical identity code, structured motion code and full scene code, which are utilized as control signals of synthesis process. The design of spatial decomposed modeling enables flexible user control, complex motion expression, as well as 3D-aware synthesis for scene interactions. Experimental results demonstrate effectiveness and robustness of the proposed method.

Alibaba presents MIMO Controllable Character Video Synthesis with Spatial Decomposed Modeling Character video synthesis aims to produce realistic videos of animatable characters within lifelike scenes. As a fundamental problem in the computer vision and graphics community, 3D works typically require multi-view captures for per-case training, which severely limits their applicability of modeling arbitrary characters in a short time. Recent 2D methods break this limitation via pre-trained diffusion models, but they struggle for pose generality and scene interaction. To this end, we propose MIMO, a novel framework which can not only synthesize character videos with controllable attributes (i.e., character, motion and scene) provided by simple user inputs, but also simultaneously achieve advanced scalability to arbitrary characters, generality to novel 3D motions, and applicability to interactive real-world scenes in a unified framework. The core idea is to encode the 2D video to compact spatial codes, considering the inherent 3D nature of video occurrence. Concretely, we lift the 2D frame pixels into 3D using monocular depth estimators, and decompose the video clip to three spatial components (i.e., main human, underlying scene, and floating occlusion) in hierarchical layers based on the 3D depth. These components are further encoded to canonical identity code, structured motion code and full scene code, which are utilized as control signals of synthesis process. The design of spatial decomposed modeling enables flexible user control, complex motion expression, as well as 3D-aware synthesis for scene interactions. Experimental results demonstrate effectiveness and robustness of the proposed method.

AK

148,998 次观看 • 1 年前

🚨Architects are going to hate this. Someone just open sourced a full 3D building editor that runs entirely in your browser. No AutoCAD. No Revit. No $5,000/year licenses. It's called Pascal Editor. Built with React Three Fiber and WebGPU -- meaning it renders directly on your GPU at near-native speed. Here's what's inside this thing: → A full building/level/wall/zone hierarchy you can edit in real time → An ECS-style architecture where every object updates through GPU-powered systems → Zustand state management with full undo/redo built in → Next.js frontend so it deploys as a web app, not a desktop install → Dirty node tracking -- only re-renders what changed, not the whole scene Here's the wildest part: You can stack, explode, or solo individual building levels. Select a zone, drag a wall, reshape a slab -- all in 3D, all in the browser. Architecture firms pay $50K+ per seat for BIM software that does this workflow. This is free. 100% Open Source.

🚨Architects are going to hate this. Someone just open sourced a full 3D building editor that runs entirely in your browser. No AutoCAD. No Revit. No $5,000/year licenses. It's called Pascal Editor. Built with React Three Fiber and WebGPU -- meaning it renders directly on your GPU at near-native speed. Here's what's inside this thing: → A full building/level/wall/zone hierarchy you can edit in real time → An ECS-style architecture where every object updates through GPU-powered systems → Zustand state management with full undo/redo built in → Next.js frontend so it deploys as a web app, not a desktop install → Dirty node tracking -- only re-renders what changed, not the whole scene Here's the wildest part: You can stack, explode, or solo individual building levels. Select a zone, drag a wall, reshape a slab -- all in 3D, all in the browser. Architecture firms pay $50K+ per seat for BIM software that does this workflow. This is free. 100% Open Source.

Guri Singh

2,967,409 次观看 • 4 个月前

99% of AI applications are cool-looking demos. Impressive, but don't get fooled by the hype. It takes a lot to build enterprise-grade products that deliver real value. I have at least three weekly conversations with companies that want to use a Large Language Model with their data. The demand is huge! Here is one idea about what you can do to help. The use cases that most of these companies want to solve are similar: They have an extensive knowledge base and want to build a simple application that uses that information to answer questions. In other words, they need help building Retrieval Augmented Generation (RAG) applications they can use in many different scenarios: 1. To train new employees 2. To help their support team 3. To search old meetings and documents 4. To help with their research However, building these systems is not straightforward. Yes, there's a lot of information online, but there aren't enough people who know how to create solutions that work. Here is the idea: Today, you can build an enterprise-grade RAG application without writing code. A couple of MIT PhDs with 10+ years of experience building AI applications created . It's a no-code platform for building applications using Large Language Models. They are partnering with me on this post. You can use Stack AI to create, test, and deploy an end-to-end production-ready AI system. It's SOC-2, HIPAA, and GDPR compliant and offers SSO, role management, access control, and on-premise deployments. Of course, you can use the platform with any LLM on the market now. It's the whole nine yards for building AI applications. Check them out here: 2023 was about models. 2024 is about the tools using these models to build production-ready applications. That's where I'd start.

99% of AI applications are cool-looking demos. Impressive, but don't get fooled by the hype. It takes a lot to build enterprise-grade products that deliver real value. I have at least three weekly conversations with companies that want to use a Large Language Model with their data. The demand is huge! Here is one idea about what you can do to help. The use cases that most of these companies want to solve are similar: They have an extensive knowledge base and want to build a simple application that uses that information to answer questions. In other words, they need help building Retrieval Augmented Generation (RAG) applications they can use in many different scenarios: 1. To train new employees 2. To help their support team 3. To search old meetings and documents 4. To help with their research However, building these systems is not straightforward. Yes, there's a lot of information online, but there aren't enough people who know how to create solutions that work. Here is the idea: Today, you can build an enterprise-grade RAG application without writing code. A couple of MIT PhDs with 10+ years of experience building AI applications created . It's a no-code platform for building applications using Large Language Models. They are partnering with me on this post. You can use Stack AI to create, test, and deploy an end-to-end production-ready AI system. It's SOC-2, HIPAA, and GDPR compliant and offers SSO, role management, access control, and on-premise deployments. Of course, you can use the platform with any LLM on the market now. It's the whole nine yards for building AI applications. Check them out here: 2023 was about models. 2024 is about the tools using these models to build production-ready applications. That's where I'd start.

Santiago

197,675 次观看 • 2 年前

Jack Dorsey on what he learned about culture from 49ers coach Bill Walsh When Bill Walsh joined the 49ers, they were the worst team in the NFL. Within three years, they were Super Bowl Champions. In his book, The Score Takes Care of Itself, he writes: “Winners act like winners before they’re winners…The culture precedes positive results. It doesn’t get tacked on as an afterthought on your way to the victory stand. Champions behave like champions before they’re champions; they have a winning standard of performance before their winners.” And he provides six guidelines for establishing a standard of performance: 1. Start with a comprehensive recognition of reverence for and identification of the specific actions and attitudes relevant to your team’s performance and production. 2. Be clarion clear in communicating your expectation of high effort and execution of your Standard of Performance. Like water, many decent individuals will seek lower ground if left to their own inclinations. In most cases you are the one who inspires and demands they go upward rather than settle for the comfort of doing what comes easily. Push them beyond their comfort zone; expect them to give extra effort. 3. Let all know that you expect them to possess the highest level of expertise in their area of responsibility. 4. Beyond standards and methodology, teach your beliefs, values, and philosophy. An organism is not an inanimate object. It is a living organism that you must nurture, guide, and strengthen. 5. Teach “connection and extension.” An organization filled with individuals who are “independent contractors” unattached to one another is a team with little interior cohesion and strength. 6. Make the expectations and metrics of competence that you demand in action and attitudes from personnel the new reality of your organization. You must provide the model for that new standard in your own actions and attitude. Twitter and Square founder Jack Dorsey encourages anyone thinking about leading teams or building a company to read this book: “What’s important about this is that as you start building a team, you need to set expectations around how people need to perform in the company—how people need to act in the company. And these can be very simple things, but without that, you are rutterless—you will react to the outside. And if you react to the outside, you are building someone else’s roadmap and you’re building someone else’s dream instead of your own.” Video source: Y Combinator (2013)

Jack Dorsey on what he learned about culture from 49ers coach Bill Walsh When Bill Walsh joined the 49ers, they were the worst team in the NFL. Within three years, they were Super Bowl Champions. In his book, The Score Takes Care of Itself, he writes: “Winners act like winners before they’re winners…The culture precedes positive results. It doesn’t get tacked on as an afterthought on your way to the victory stand. Champions behave like champions before they’re champions; they have a winning standard of performance before their winners.” And he provides six guidelines for establishing a standard of performance: 1. Start with a comprehensive recognition of reverence for and identification of the specific actions and attitudes relevant to your team’s performance and production. 2. Be clarion clear in communicating your expectation of high effort and execution of your Standard of Performance. Like water, many decent individuals will seek lower ground if left to their own inclinations. In most cases you are the one who inspires and demands they go upward rather than settle for the comfort of doing what comes easily. Push them beyond their comfort zone; expect them to give extra effort. 3. Let all know that you expect them to possess the highest level of expertise in their area of responsibility. 4. Beyond standards and methodology, teach your beliefs, values, and philosophy. An organism is not an inanimate object. It is a living organism that you must nurture, guide, and strengthen. 5. Teach “connection and extension.” An organization filled with individuals who are “independent contractors” unattached to one another is a team with little interior cohesion and strength. 6. Make the expectations and metrics of competence that you demand in action and attitudes from personnel the new reality of your organization. You must provide the model for that new standard in your own actions and attitude. Twitter and Square founder Jack Dorsey encourages anyone thinking about leading teams or building a company to read this book: “What’s important about this is that as you start building a team, you need to set expectations around how people need to perform in the company—how people need to act in the company. And these can be very simple things, but without that, you are rutterless—you will react to the outside. And if you react to the outside, you are building someone else’s roadmap and you’re building someone else’s dream instead of your own.” Video source: Y Combinator (2013)

Startup Archive

71,362 次观看 • 3 个月前