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。

been building an agent-native, node-based, real-time environment. as a test, I gave gpt-5.6 sol a prompt to plan out its own setup and go into a loop of creative exploration until it felt that it was done. it invented every node, every data type, every connection and all of... show more

Spencer Sterling

6,590 subscribers

22,863 次观看 • 3 个月前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

25 条评论

Finn McKenty 的头像

Finn McKenty3 个月前

Love to see people building weird shit just because they can! That’s always where the coolest ideas come from

andrew pignanelli 的头像

andrew pignanelli3 个月前

touchdesigner but you have to be even more autistic

Jeffrey Emanuel 的头像

Jeffrey Emanuel3 个月前

Very cool

🧪Morgen 的头像

🧪Morgen3 个月前

so glad to run into your account again Cers! I gotta hop in the discord soon! Miss seeing you guys and your bad ass creations!! Hope you are well!!

Tinseltown Pixels 的头像

Tinseltown Pixels3 个月前

Amazing Spencer keep building

𝖒𝖆𝖈𝖇𝖊𝖙𝖍 的头像

𝖒𝖆𝖈𝖇𝖊𝖙𝖍3 个月前

Yesssssssssss

srolkk 的头像

srolkk3 个月前

RIP

Stas 的头像

Stas3 个月前

Remember hearing you at the comfy meet up in nyc a couple years ago and was impressed back then but this is 🤯

Evan 的头像

Evan3 个月前

Wizardry 🤯

Stern - uɹǝʇS 的头像

Stern - uɹǝʇS3 个月前

Killing it Spencer

Rohan Arun 的头像

Rohan Arun3 个月前

Hey Super, build a single-page browser tool called 'Agent Node Studio' where an autonomous planning agent designs a creative node-graph pipeline that visitors steer in real time. The first viewport shows a dark Konva.js canvas with a left-side prompt dock ("What should the agent build?") and a right-side agent reasoning panel ("Thinking… proposing source node"). The visitor types a creative intent like "a generative ambient music sequencer that responds to text mood" and clicks Compose. The agent immediately begins spawning nodes — Input(text), Analyzer(mood_score), Mapper(scale/tempo), Sequencer(drone/arpeggio), Mixer, Output(audio_graph) — with bezier connections animating as they wire. Each node is a draggable rounded rectangle with editable parameters (sliders, type selectors, value fields), color-coded by category (teal input, violet transform, amber combine, rose output, slate control). While the agent plans, the user can pause, reject a proposal, redirect with a steering prompt ("make it more glitchy"), or jump in and edit any node's parameters — the agent then adapts its next proposal to compensate. The graph runs live: audio nodes use Tone.js for real playback, visual nodes draw to the canvas, data nodes show live charts. Pin Konva.js v9 as the canvas engine and Tone.js v15 for any audio nodes. Persist the current graph and agent state to localStorage so a refresh restores the workspace, with a visible Reset button and iteration rollback. Handle empty prompts by auto-suggesting three starter intents the user can click to launch the agent. On mobile, collapse the side panels into a bottom sheet so the graph stays the primary interaction surface with full touch-drag support. When the visitor clicks Done, the canvas reveals an export panel with a clean, human-readable JSON recipe of the entire graph, ready to copy or download. The agent's planning loop is a deterministic local state machine driven by the prompt's detected domain (audio, visual, data, text) — no external API keys, no cloud calls.

Ronny Khalil 的头像

Ronny Khalil3 个月前

holy

Mr. Decentralized 的头像

Mr. Decentralized3 个月前

Is there appropriate chaos? Is there a controller? Is there blockchain based snapshots? Can a snapshot be bidded on?

Meta-Gnome 的头像

Meta-Gnome3 个月前

Insane!

Surf 的头像

Surf3 个月前

Could you please put this on github? I’m interested in the visual aspects of it. What kind of hardware does it need or is it webgpu? Thanks

Jordan Thompson 的头像

Jordan Thompson2 个月前

Cool. What the fuck is it? This looks like a whole lot of broken scalar noise layered up and randomly masked by other scalar. Can't any 3D DCC do this today? I think many have MCP hooks too now.

Gaurav Joshi 的头像

Gaurav Joshi3 个月前

Wow! Looks amazing.

JQ 的头像

JQ3 个月前

👀

Evan Boehm 的头像

Evan Boehm3 个月前

soo interesting to watch new aesthetic vibes

Moon 的头像

Moon3 个月前

holy..

SpazeCraft 的头像

SpazeCraft3 个月前

Did you just say you built this in chat gpt??

Spencer Sterling 的头像

Spencer Sterling3 个月前

The entire software was built mostly with claude all in c++. The actual visuals in this example were vibecoded with chatgpt using the built in terminal and mcp!

SpazeCraft 的头像

SpazeCraft3 个月前

@cerspense I hear that, it’s just gpt & Claude is getting expensive on my “freelance not much work career” diet lol & im having way too much fun making my life’s work come to life in Claude. I’m bout to phase GPT out sadly, and I’m praying to install comfy UI this weekend. Wish me luck lol

Caidan 的头像

Caidan3 个月前

Dude this is so cool! Similar to some of the stuff I'm working on

hansfrom.space 的头像

hansfrom.space3 个月前

Building something similar. Very inspiring to see! My system is also nodebased and aims to create audiovisual organisms "truly" random. Let's have a chat if you like!

相关视频

i wrote a brainstorming skill to get AI to walk a human through a creative process when stuck. then i told it to just use it itself, gave it a simplistic generative art piece, and let it run till it felt it was done. here's a video of every image it generated.

i wrote a brainstorming skill to get AI to walk a human through a creative process when stuck. then i told it to just use it itself, gave it a simplistic generative art piece, and let it run till it felt it was done. here's a video of every image it generated.

joshua schachter

15,596 次观看 • 2 个月前

Opus 5.5 is honestly getting out of hand… I gave it one floor plan and it built the whole house from pieces LEGO already makes. > it reads the plan into walls, doors, windows and a working staircase > 3,149 existing parts, every connection checked, nothing custom moulded > Scenario MCP renders the real house, and the .ldr prices it in BrickLink Studio Every new homeowner could get their own house as a set.

Opus 5.5 is honestly getting out of hand… I gave it one floor plan and it built the whole house from pieces LEGO already makes. > it reads the plan into walls, doors, windows and a working staircase > 3,149 existing parts, every connection checked, nothing custom moulded > Scenario MCP renders the real house, and the .ldr prices it in BrickLink Studio Every new homeowner could get their own house as a set.

Emm | scenario.com

27,507 次观看 • 1 个月前

I just built a collection of 163 insane motion graphics, and every graphic has the exact prompt that was used to create it right beside it. the gallery includes work made with Opus 5.5, GPT-6 Astra, Kimi K3 and Fable 5, with every creator credited. it’s free and i will be adding more go try it out:

I just built a collection of 163 insane motion graphics, and every graphic has the exact prompt that was used to create it right beside it. the gallery includes work made with Opus 5.5, GPT-6 Astra, Kimi K3 and Fable 5, with every creator credited. it’s free and i will be adding more go try it out:

MIKE

26,642 次观看 • 21 天前

i gave 5.6 sol access to my camera roll and had it extract pictures of every piece of clothing i own from my photos then, told it to find new outfits for me and render them on me with gpt-image! its kinda cool to see your entire wardrobe in a collection like this

i gave 5.6 sol access to my camera roll and had it extract pictures of every piece of clothing i own from my photos then, told it to find new outfits for me and render them on me with gpt-image! its kinda cool to see your entire wardrobe in a collection like this

thijs

7,254,727 次观看 • 3 个月前

Satya Nadella reveals his coding agent pulls every hyperscaler and neoclouds' SEC filings into a dashboard that refreshes every day for real-time ROIC "And this is the other aspect of it, which is the enterprise context combined with the world's context. In fact, I go to the SEC filings of every cloud provider, hyperscaler, each of these neoclouds. It's in real time." "I have a data runner in Fabric that brings all that data, puts it into a semantic model that then gets read by my coding agent and then surfaces it as a dashboard. And every day it's fresh." "So I have the entirety of every SEC filing that goes out there, plus all of my internal analysis constantly coming together, giving me real-time ROIC by layer."

Satya Nadella reveals his coding agent pulls every hyperscaler and neoclouds' SEC filings into a dashboard that refreshes every day for real-time ROIC "And this is the other aspect of it, which is the enterprise context combined with the world's context. In fact, I go to the SEC filings of every cloud provider, hyperscaler, each of these neoclouds. It's in real time." "I have a data runner in Fabric that brings all that data, puts it into a semantic model that then gets read by my coding agent and then surfaces it as a dashboard. And every day it's fresh." "So I have the entirety of every SEC filing that goes out there, plus all of my internal analysis constantly coming together, giving me real-time ROIC by layer."

Fireside Alpha

1,271,382 次观看 • 1 个月前

Claude Opus 5.5 drew my apartment in 3D and found 3 cameras in it that are not mine... one prompt, 20 minutes, two of them were still sending audio out here is exactly how it went: → I started one agent in the terminal on a laptop sitting on my home Wi-Fi and left it alone → it swept the network first and listed every device it could see, with brand and model where it could match one → then it measured signal strength from a few spots in the flat and built a rough room layout out of that → every device got pinned onto that layout, so instead of a list of IPs I was looking at a map of my own place → anything it could not tie to something I own got flagged on its own, with what it was transmitting and roughly where it sat three of those flags were cameras. two more were audio only. I had lived there two years and walked past all five every day the whole thing ran on one prompt and I never touched the keyboard after the first line full run on video, prompt in the replies ↓

Noisy

11,305 次观看 • 5 天前

I don't think i can go back to regular software anymore exploring a crazy combo of real-time depth estimation, raymarched SDFs, StreamDiffusion and BiRefNet. The entire setup was authored by GPT 5.6 Sol and Luna in Sentinel over MCP while I steered it in real-time.

I don't think i can go back to regular software anymore exploring a crazy combo of real-time depth estimation, raymarched SDFs, StreamDiffusion and BiRefNet. The entire setup was authored by GPT 5.6 Sol and Luna in Sentinel over MCP while I steered it in real-time.

Spencer Sterling

42,902 次观看 • 3 个月前

OptimAI Core Node v2 is live. And with it, the introduction of OptimAI Claw. From nodes as infrastructure → to nodes as execution environments. With Core Node v2, every device becomes a runtime for native personal agents. 👉 Get your node up and running. No complex setup. No fragmented tooling. Just a few clicks... and your agent is live. OptimAI Claw brings persistent, node-resident agents directly into your system: • Multi-source research across web & social • Real-time summarization and signal tracking • Autonomous workflow execution • Continuous learning through reinforcement All running locally. All under your control. Supported on Windows & macOS. This is beyond prompt-based AI. Beyond isolated tools. This is personal agent infrastructure. Where every node thinks, acts, and evolves. Claw is now part of the network. And this is only the beginning. Happy Building, Making Impact! 🔥

OptimAI Core Node v2 is live. And with it, the introduction of OptimAI Claw. From nodes as infrastructure → to nodes as execution environments. With Core Node v2, every device becomes a runtime for native personal agents. 👉 Get your node up and running. No complex setup. No fragmented tooling. Just a few clicks... and your agent is live. OptimAI Claw brings persistent, node-resident agents directly into your system: • Multi-source research across web & social • Real-time summarization and signal tracking • Autonomous workflow execution • Continuous learning through reinforcement All running locally. All under your control. Supported on Windows & macOS. This is beyond prompt-based AI. Beyond isolated tools. This is personal agent infrastructure. Where every node thinks, acts, and evolves. Claw is now part of the network. And this is only the beginning. Happy Building, Making Impact! 🔥

OptimAI Network

384,824 次观看 • 6 个月前

We gave an AI agent a real Coinbase wallet — then put Sentinel in front of it. Every payment it makes is gated by policy: allow, block, or escalate. Every decision signed and tamper-evident. Watch it block a prompt-injected payment in real time.

We gave an AI agent a real Coinbase wallet — then put Sentinel in front of it. Every payment it makes is gated by policy: allow, block, or escalate. Every decision signed and tamper-evident. Watch it block a prompt-injected payment in real time.

Montana Labs

15,605 次观看 • 4 个月前

A pile of 3D-printed parts sat unfinished, waiting for tiny metal inserts to go in straight. Doing it by hand was slow and easy to ruin. So a maker designed his own press, printed every piece, and built it with care. When it was done, every insert went in clean and true.

A pile of 3D-printed parts sat unfinished, waiting for tiny metal inserts to go in straight. Doing it by hand was slow and easy to ruin. So a maker designed his own press, printed every piece, and built it with care. When it was done, every insert went in clean and true.

Luna

11,493 次观看 • 1 个月前

The next evolution of the OptimAI Core Node is coming, and it’s bringing one of our most powerful innovations yet: The Data Mining Agent. Become an OptimAI Node: 🔍 What is the Data Mining Agent? A fully autonomous agent running on your Core Node that: + Scans, retrieves, and indexes real-time public web data + Transforms raw information into actionable intelligence + Powers faster, smarter, adaptive Persona AI systems 💡 Why It Matters: Agentic AI is only as strong as the data it can access. Centralized APIs are slow, expensive, and limited. With the Core Node + Data Mining Agent, every contributor turns their desktop into a data supernode, accelerating AI evolution and giving our community a massive edge. ⚡ We’re not just building tools - we’re architecting the global backbone of Agentic AI, powered by you. Stay tuned for teasers, and get ready to upgrade your node. The future is decentralized. The future is agentic. And you are a key part of it. 💪 The more you contribute, the more you earn. Let’s BUIDL the future of AI together.

The next evolution of the OptimAI Core Node is coming, and it’s bringing one of our most powerful innovations yet: The Data Mining Agent. Become an OptimAI Node: 🔍 What is the Data Mining Agent? A fully autonomous agent running on your Core Node that: + Scans, retrieves, and indexes real-time public web data + Transforms raw information into actionable intelligence + Powers faster, smarter, adaptive Persona AI systems 💡 Why It Matters: Agentic AI is only as strong as the data it can access. Centralized APIs are slow, expensive, and limited. With the Core Node + Data Mining Agent, every contributor turns their desktop into a data supernode, accelerating AI evolution and giving our community a massive edge. ⚡ We’re not just building tools - we’re architecting the global backbone of Agentic AI, powered by you. Stay tuned for teasers, and get ready to upgrade your node. The future is decentralized. The future is agentic. And you are a key part of it. 💪 The more you contribute, the more you earn. Let’s BUIDL the future of AI together.

OptimAI Network

50,496 次观看 • 1 年前

An agent's identity shifts with every context. Every prompt. Every data source. So who's actually accountable for what it does? The answer isn't in the agent. It's in the human behind it. Like a surname that carries the weight of everything done under it.

An agent's identity shifts with every context. Every prompt. Every data source. So who's actually accountable for what it does? The answer isn't in the agent. It's in the human behind it. Like a surname that carries the weight of everything done under it.

Billions Network

38,330 次观看 • 7 个月前

THIS CHINESE DEVELOPER’S NEURAL NETWORK VISUALIZATION IS EXACTLY HOW YOUR OBSIDIAN VAULT SHOULD WORK every node connects to every other node and the whole thing gets smarter the more data flows through it, same way a real knowledge system should work but almost nobody builds it like that most people throw everything into folders and tags and call it a second brain but what they actually built is an archive that gets harder to use every month claude reads across everything you ever captured and finds the connection you need right when you need it here’s how to build the version that actually compounds ↓

THIS CHINESE DEVELOPER’S NEURAL NETWORK VISUALIZATION IS EXACTLY HOW YOUR OBSIDIAN VAULT SHOULD WORK every node connects to every other node and the whole thing gets smarter the more data flows through it, same way a real knowledge system should work but almost nobody builds it like that most people throw everything into folders and tags and call it a second brain but what they actually built is an archive that gets harder to use every month claude reads across everything you ever captured and finds the connection you need right when you need it here’s how to build the version that actually compounds ↓

leopardracer

231,513 次观看 • 5 个月前

The biggest marketplace on Earth won't be the assets humans own. It'll be the data AI needs. Everything that can be digitised is getting pulled in to train the next models. Every scientific dataset, every piece of research sitting in a university, every sensor reading off every John Deere tractor in a field. All of it feeds the machine, and almost none of it is visible to us. What comes out is an agentic economy running at machine speed. You plug your data into a vault, and agents monetise it on your behalf, trading with other agents in real time, in a marketplace we never see. The company sitting on proprietary data becomes part of it without even knowing. This is what the Economic Singularity actually looks like from the inside. The real economy stops being the one we can see and price, and moves somewhere we can't follow. You own a piece of it, or you get priced out of it. Jordi Visser.

The biggest marketplace on Earth won't be the assets humans own. It'll be the data AI needs. Everything that can be digitised is getting pulled in to train the next models. Every scientific dataset, every piece of research sitting in a university, every sensor reading off every John Deere tractor in a field. All of it feeds the machine, and almost none of it is visible to us. What comes out is an agentic economy running at machine speed. You plug your data into a vault, and agents monetise it on your behalf, trading with other agents in real time, in a marketplace we never see. The company sitting on proprietary data becomes part of it without even knowing. This is what the Economic Singularity actually looks like from the inside. The real economy stops being the one we can see and price, and moves somewhere we can't follow. You own a piece of it, or you get priced out of it. Jordi Visser.

Raoul Pal

79,211 次观看 • 3 个月前

Only if education could be this interactive ❤️‍🔥 I've had a looong wish to build something genuinely useful through vibe coding, and I finally did it. A 3D human anatomy application built with Three.js using GPT 5.6 Sol. It all started with a single design image that I created using GPT Image 2.0. I then used it to generate every 3D organ image, one by one. Next, I converted each of those images into 3D models using Tripo (and no, they didn't sponsor this 😄). After that, I opened Codex, wrote a master prompt based on the design, and gave it the prompt, the design image, and all the 3D models. Codex built the first version beautifully, but there was one big problem. Every single 3D model was nearly 120-150 MB. That obviously wasn't practical for the web and was giving a performance of 16fps. After a few iterations, Codex optimized each model down to roughly 2–5.5 MB while preserving the visual quality, reducing the total asset size from ~900 MB to just 28.6 MB. And each model loads on demand. Along the way, Codex also generated those anatomical illustrations showing where each organ sits in the human body, and even created the interactive hotspot markers that explain different parts of every organ. It handled all of that. The process wasn't exactly one shot, but it also wasn't difficult. You just have to do it step by step. It genuinely felt like building something that could make learning anatomy much more engaging. The inspiration came from Dilum Sanjaya's 3D animal plant cell project. I remember seeing it and thinking, "I want to build something like this one day." And I did it :D Live: Code:

Only if education could be this interactive ❤️‍🔥 I've had a looong wish to build something genuinely useful through vibe coding, and I finally did it. A 3D human anatomy application built with Three.js using GPT 5.6 Sol. It all started with a single design image that I created using GPT Image 2.0. I then used it to generate every 3D organ image, one by one. Next, I converted each of those images into 3D models using Tripo (and no, they didn't sponsor this 😄). After that, I opened Codex, wrote a master prompt based on the design, and gave it the prompt, the design image, and all the 3D models. Codex built the first version beautifully, but there was one big problem. Every single 3D model was nearly 120-150 MB. That obviously wasn't practical for the web and was giving a performance of 16fps. After a few iterations, Codex optimized each model down to roughly 2–5.5 MB while preserving the visual quality, reducing the total asset size from ~900 MB to just 28.6 MB. And each model loads on demand. Along the way, Codex also generated those anatomical illustrations showing where each organ sits in the human body, and even created the interactive hotspot markers that explain different parts of every organ. It handled all of that. The process wasn't exactly one shot, but it also wasn't difficult. You just have to do it step by step. It genuinely felt like building something that could make learning anatomy much more engaging. The inspiration came from Dilum Sanjaya's 3D animal plant cell project. I remember seeing it and thinking, "I want to build something like this one day." And I did it :D Live: Code:

The Bugged Dev

2,123,870 次观看 • 2 个月前

Introducing Algrow Desktop, a new way to research YouTube. Out today on Mac and Windows, free with every plan. And you can now try it for $1. Unlike the MCP, the agent isn't stuck inside Algrow's data. It browses YouTube itself, live. It opens a real tab, searches, scrolls the feed, clicks into channels, and you can watch it live. Each agent works in its own space with its own YouTube account, so each one learns a different corner of YouTube. One for your main channel, one for a format you want to try, one for a topic you're curious about. Run as many as you want and they never mix. The MCP can't see your feed. This can. It still has everything the MCP has. Every Algrow search, every channel, every title format but with higher limits. And every video it finds on YouTube comes with Algrow's data: how fast it grew, similar subjects, similar formats, which channels are like it, why it did better than the channel's usual. No setup. If Claude is on your machine, it's already connected. It runs on your own Claude plan, so there's no extra AI bill.

Introducing Algrow Desktop, a new way to research YouTube. Out today on Mac and Windows, free with every plan. And you can now try it for $1. Unlike the MCP, the agent isn't stuck inside Algrow's data. It browses YouTube itself, live. It opens a real tab, searches, scrolls the feed, clicks into channels, and you can watch it live. Each agent works in its own space with its own YouTube account, so each one learns a different corner of YouTube. One for your main channel, one for a format you want to try, one for a topic you're curious about. Run as many as you want and they never mix. The MCP can't see your feed. This can. It still has everything the MCP has. Every Algrow search, every channel, every title format but with higher limits. And every video it finds on YouTube comes with Algrow's data: how fast it grew, similar subjects, similar formats, which channels are like it, why it did better than the channel's usual. No setup. If Claude is on your machine, it's already connected. It runs on your own Claude plan, so there's no extra AI bill.

sam

46,645 次观看 • 2 天前

Gave UEFN its own blueprint graph. Python tool running inside the editor. Scans your level and parses your Verse code to draw a full visual map of your game logic. What it shows you: - Every device in your level as a node - Every @ editable reference between devices - Every event subscription and function call - Which Verse functions are actually being used vs sitting there doing nothing What it catches: - Broken connections where an @ editable points to nothing - Orphan devices with zero connections - Unassigned editables you forgot to wire up What you can do with it: - Click a node, and it selects the device in your viewport - Double-click to fly straight to it - Open the Verse file in VS Code with one button - Search and filter by device type - Zoom in and out - Add notes to any device - Export the whole graph as a doc

Gave UEFN its own blueprint graph. Python tool running inside the editor. Scans your level and parses your Verse code to draw a full visual map of your game logic. What it shows you: - Every device in your level as a node - Every @ editable reference between devices - Every event subscription and function call - Which Verse functions are actually being used vs sitting there doing nothing What it catches: - Broken connections where an @ editable points to nothing - Orphan devices with zero connections - Unassigned editables you forgot to wire up What you can do with it: - Click a node, and it selects the device in your viewport - Double-click to fly straight to it - Open the Verse file in VS Code with one button - Search and filter by device type - Zoom in and out - Add notes to any device - Export the whole graph as a doc

Immature

12,163 次观看 • 7 个月前

Demo 2: human + agent design loop inside Codex. Not every workflow needs the agent to drive everything. You design in Open Design. Codex reviews the output. It can inspect files, open pages, test interactions, critique the design, and turn feedback into the next prompt. Design becomes a human + agent feedback loop.

Demo 2: human + agent design loop inside Codex. Not every workflow needs the agent to drive everything. You design in Open Design. Codex reviews the output. It can inspect files, open pages, test interactions, critique the design, and turn feedback into the next prompt. Design becomes a human + agent feedback loop.

Open Design

20,103 次观看 • 5 个月前

🇺🇸 Bill Gates attempts to apologize for his connection to Epstein: "I regret every minute I spent with him, and I apologize for doing it. You know, the facts are that I only went to dinners, I've never been on the island, I've never dated any women, and so the more this comes out, the clearer it becomes that while that time was a mistake, it had nothing to do with this kind of behavior."

🇺🇸 Bill Gates attempts to apologize for his connection to Epstein: "I regret every minute I spent with him, and I apologize for doing it. You know, the facts are that I only went to dinners, I've never been on the island, I've never dated any women, and so the more this comes out, the clearer it becomes that while that time was a mistake, it had nothing to do with this kind of behavior."

Lord Bebo

39,180 次观看 • 8 个月前