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。

wrote an open-source ngrok alternative which creates instant ephemeral HTTP tunnels to localhost. It uses userspace wireguard network, so 1) no root permissions needed to run the server 2) client side doesn't need custom software to connect

Karan Sharma

5,645 subscribers

13,059 次观看 • 1 年前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

How to find viable targets for client-side desync attacks: 1️⃣ Open Burp Suite and intercept requests. 2️⃣ Choose an endpoint that wouldn't usually expect a POST request (e.g GET) and send it to repeater. 3️⃣ Go to Inspector > Request Attributes > Protocol field > Upgrade to HTTP/2. 4️⃣ Send the Modified Request and look for this error message: "Server ALPN does not advertise HTTP/2 support"... 5️⃣ Enable ALPN Override in request settings and send again. If you see "Stream failed to close correctly" then you've confirmed that the server does not support HTTP/2 and is a valid target for desync testing! Want to learn more about desync attacks? On August 6, at Black Hat USA, James Kettle from PortSwigger Research will reveal new classes of desync attack that enabled him to compromise multiple CDNs and kick off the desync endgame! 😲 Stay up to date here:

How to find viable targets for client-side desync attacks: 1️⃣ Open Burp Suite and intercept requests. 2️⃣ Choose an endpoint that wouldn't usually expect a POST request (e.g GET) and send it to repeater. 3️⃣ Go to Inspector > Request Attributes > Protocol field > Upgrade to HTTP/2. 4️⃣ Send the Modified Request and look for this error message: "Server ALPN does not advertise HTTP/2 support"... 5️⃣ Enable ALPN Override in request settings and send again. If you see "Stream failed to close correctly" then you've confirmed that the server does not support HTTP/2 and is a valid target for desync testing! Want to learn more about desync attacks? On August 6, at Black Hat USA, James Kettle from PortSwigger Research will reveal new classes of desync attack that enabled him to compromise multiple CDNs and kick off the desync endgame! 😲 Stay up to date here:

Web Security Academy

16,569 次观看 • 1 年前

So... I literally oneshotted this code with Claude Sonnet 3.7 for replicating the exact same browser sandbox runtime that Manus uses. And I am going to open-source it, welcome contributions for building out the React VNC client, integrating to browser use, agent loop, etc. But we need a name first, should we call it... - Autonomous Neural Universal System - Magnus - or ?

So... I literally oneshotted this code with Claude Sonnet 3.7 for replicating the exact same browser sandbox runtime that Manus uses. And I am going to open-source it, welcome contributions for building out the React VNC client, integrating to browser use, agent loop, etc. But we need a name first, should we call it... - Autonomous Neural Universal System - Magnus - or ?

jian

419,367 次观看 • 1 年前

I’m making big changes to the AI hedge fund. Multi ticker trading. Here is how it works: • you specify tickers • analysts research all tickers • portfolio manager creates trading signal Before: we only supported 1 ticker. Now: you can add as many as you like. No coding experience required to run. Everything is open source.

I’m making big changes to the AI hedge fund. Multi ticker trading. Here is how it works: • you specify tickers • analysts research all tickers • portfolio manager creates trading signal Before: we only supported 1 ticker. Now: you can add as many as you like. No coding experience required to run. Everything is open source.

Virat Singh

29,648 次观看 • 1 年前

PixVerse V6 (left) vs Seedance 2 (right). Seedance mispronounced a name (a persistent problem in the model - he says "Castellani" but I wrote Castellano), however it is otherwise far superior to PixVerse. I hope another company is able to dethrone Seedance 2 in the very near future - we need proper competition in AI video, as well as an alternative to the totalitarian censorship adopted by Bytedance since SD2 first appeared - but at the moment, no other models come close to it.

PixVerse V6 (left) vs Seedance 2 (right). Seedance mispronounced a name (a persistent problem in the model - he says "Castellani" but I wrote Castellano), however it is otherwise far superior to PixVerse. I hope another company is able to dethrone Seedance 2 in the very near future - we need proper competition in AI video, as well as an alternative to the totalitarian censorship adopted by Bytedance since SD2 first appeared - but at the moment, no other models come close to it.

Christopher Gwinn | Grindhouse Glitch

13,341 次观看 • 6 个月前

Inviting early testers and contributors to Project Devika - The open-source alternative to Devin. 👩‍💻 As of now, Devika is far from the capabilities of Devin... but we'll eventually get there. So I am calling the open-source community to join forces! ❤️ Features: - 12 Agentic models that can interact with each other in a feedback loop to understand, browse, research, code, document, and make decisions according to the user's query to complete a project. - Supports Claude 3, GPT-4, GPT-3.5, and Local LLMs via ollama. - Devika can run the code she writes and fix/patch the code herself if she encounters any errors without user intervention. - Devika can deploy static websites she creates on Netlify. (Experimental) - And much more... Will be doing an official launch after intensive testing and bug fixes. 🙌 I've created a Discord server for the early testers and contributors. If you're interested in joining the team, reply to this tweet and I will DM you the invite link. #buildinpublic

Inviting early testers and contributors to Project Devika - The open-source alternative to Devin. 👩‍💻 As of now, Devika is far from the capabilities of Devin... but we'll eventually get there. So I am calling the open-source community to join forces! ❤️ Features: - 12 Agentic models that can interact with each other in a feedback loop to understand, browse, research, code, document, and make decisions according to the user's query to complete a project. - Supports Claude 3, GPT-4, GPT-3.5, and Local LLMs via ollama. - Devika can run the code she writes and fix/patch the code herself if she encounters any errors without user intervention. - Devika can deploy static websites she creates on Netlify. (Experimental) - And much more... Will be doing an official launch after intensive testing and bug fixes. 🙌 I've created a Discord server for the early testers and contributors. If you're interested in joining the team, reply to this tweet and I will DM you the invite link. #buildinpublic

mufeed vh

155,097 次观看 • 2 年前

Despite being told no, I'm open-sourcing Chatpack. You can now deploy production-ready chat to any app with 30+ native chat features in a single command using npx chatpack@latest I wanted to build a simple chat layer where anyone could add 24/7 live messaging to their app without spending weeks rebuilding the same infrastructure from scratch. Connect your existing auth and database, and Chatpack handles everything underneath. So today, we're open-sourcing Chatpack. > 24/7 real-time messaging between your users > typing, presence, read receipts, permissions, and conversation history > works with your existing auth + database, so your users and data stay yours Run it as a CLI or library. Fully open source. Repo is MIT licensed. If I were starting an AI company today, I'd clone this, pick a market, and begin shipping with Claude Code. Truly excited to see what comes out of this.

Despite being told no, I'm open-sourcing Chatpack. You can now deploy production-ready chat to any app with 30+ native chat features in a single command using npx chatpack@latest I wanted to build a simple chat layer where anyone could add 24/7 live messaging to their app without spending weeks rebuilding the same infrastructure from scratch. Connect your existing auth and database, and Chatpack handles everything underneath. So today, we're open-sourcing Chatpack. > 24/7 real-time messaging between your users > typing, presence, read receipts, permissions, and conversation history > works with your existing auth + database, so your users and data stay yours Run it as a CLI or library. Fully open source. Repo is MIT licensed. If I were starting an AI company today, I'd clone this, pick a market, and begin shipping with Claude Code. Truly excited to see what comes out of this.

Daniel Ch

82,639 次观看 • 1 个月前

Litter iOS just got a beta (TestFlight) release so now you can install it and try it out. It's a remote for codex! - Finds all the computers on your local network (or tailscale) that run codex app-server or ssh - Lets you continue any session or start a new one - Experimental local codex with oauth support Free, fully open-source, fully native, with android release coming soon! Thanks to Maky drumroll.dev ☄️ PincentΞ for the contributions! TestFlight ⬇️

Litter iOS just got a beta (TestFlight) release so now you can install it and try it out. It's a remote for codex! - Finds all the computers on your local network (or tailscale) that run codex app-server or ssh - Lets you continue any session or start a new one - Experimental local codex with oauth support Free, fully open-source, fully native, with android release coming soon! Thanks to Maky drumroll.dev ☄️ PincentΞ for the contributions! TestFlight ⬇️

SIGKITTEN

68,308 次观看 • 7 个月前

MoonPay is now open to AI agents, powered by PayBox all you need is a Claude or ChatGPT account from there, you can research, predict, onramp, convert, trade, offramp, and shop just by chatting with the agent on your smartphone and computer here's how to get started: 1️⃣ copy this URL: 2️⃣ open Claude with this link and the PayBox connector arrives pre-filled, just tap Add. Make sure you're signed in to Claude in this browser first: 3️⃣ If the pre-filled window doesn't appear, add it manually: go to Customize → Connectors, add a custom connector, set Name to "PayBox", and paste the URL from step 1. 4️⃣ "Connect" next to the new PayBox connector and sign in to link your account. when you open Claude next, just ask "is my PayBox connected" and you're ready to roll learn about it at and welcome to the future of agentic payments, commerce, trading, shopping, and so much more...

MoonPay is now open to AI agents, powered by PayBox all you need is a Claude or ChatGPT account from there, you can research, predict, onramp, convert, trade, offramp, and shop just by chatting with the agent on your smartphone and computer here's how to get started: 1️⃣ copy this URL: 2️⃣ open Claude with this link and the PayBox connector arrives pre-filled, just tap Add. Make sure you're signed in to Claude in this browser first: 3️⃣ If the pre-filled window doesn't appear, add it manually: go to Customize → Connectors, add a custom connector, set Name to "PayBox", and paste the URL from step 1. 4️⃣ "Connect" next to the new PayBox connector and sign in to link your account. when you open Claude next, just ask "is my PayBox connected" and you're ready to roll learn about it at and welcome to the future of agentic payments, commerce, trading, shopping, and so much more...

MoonPay 🟣

668,362 次观看 • 1 个月前

Today our Pi main screen changed to the following which we can see there will be another Hackathon! I believe this is a great work and deployment from CT! And Hackathon should be continuing to expand to apps which can connect to all industries widely in the world. I can see some Pioneers disappointed. Because these pioneers don't pay attention to Pi Network vision and mission and not understanding white paper well. They just want to get a car, property and cash ASAP. This is a short-sight. The eco system is Pi Network blood. This is critical. I used an example to make pioneers understand well what the relationship pioneers and CT is and what our pioneers task is. CT and pioneers are like a man's two feet. CT is on the right foot and pioneers are on the left foot. CT is like 💡, pioneers is like ⚡. We need to walk together to move. And if you want a light bulb, you need to have stable electricity and this stable electricity is like GCV. So when CT advanced to the Hackathon eco hardware system, pioneers should work on a software system which includes GCV, passion, attention and influence of Pi Network. When pioneers are ready with the above, ecosystem development will be super fast and easy. It will automatically go to open the mainnet! During the time CT and developers will keep on working on ecosystem hardware to develop more good platforms! So all pioneers should focus on self improvement! Of course if some Pioneers don't have trust in Pi Network, we all respect your decision 🙏 However this is yourself losses 🙏🙏🙏 All best wishes to pioneers all community! The third video introduces HuaRun Group. We don't need to argue if HuaRun joined the Pi Network. But we believe if all pioneers work hard with CT, not only HuaRun but also all super businesses will join Pi Network! So don't be an observer to wait and complain but stand up as a brave pioneer fighting for Pi Network success! Pi Network #PiNetwork #PiNetworkLive #Piopenmainnet #PiGCV

Today our Pi main screen changed to the following which we can see there will be another Hackathon! I believe this is a great work and deployment from CT! And Hackathon should be continuing to expand to apps which can connect to all industries widely in the world. I can see some Pioneers disappointed. Because these pioneers don't pay attention to Pi Network vision and mission and not understanding white paper well. They just want to get a car, property and cash ASAP. This is a short-sight. The eco system is Pi Network blood. This is critical. I used an example to make pioneers understand well what the relationship pioneers and CT is and what our pioneers task is. CT and pioneers are like a man's two feet. CT is on the right foot and pioneers are on the left foot. CT is like 💡, pioneers is like ⚡. We need to walk together to move. And if you want a light bulb, you need to have stable electricity and this stable electricity is like GCV. So when CT advanced to the Hackathon eco hardware system, pioneers should work on a software system which includes GCV, passion, attention and influence of Pi Network. When pioneers are ready with the above, ecosystem development will be super fast and easy. It will automatically go to open the mainnet! During the time CT and developers will keep on working on ecosystem hardware to develop more good platforms! So all pioneers should focus on self improvement! Of course if some Pioneers don't have trust in Pi Network, we all respect your decision 🙏 However this is yourself losses 🙏🙏🙏 All best wishes to pioneers all community! The third video introduces HuaRun Group. We don't need to argue if HuaRun joined the Pi Network. But we believe if all pioneers work hard with CT, not only HuaRun but also all super businesses will join Pi Network! So don't be an observer to wait and complain but stand up as a brave pioneer fighting for Pi Network success! Pi Network #PiNetwork #PiNetworkLive #Piopenmainnet #PiGCV

Doris Yin 东方紫莲🪷

36,410 次观看 • 3 年前

What is Hedera Hashgraph? Unlike traditional blockchains, Hedera (Hedera) does not organize transactions into a single chain of blocks. Instead, it uses a different distributed ledger structure called Hashgraph, designed to reach consensus quickly without proof-of-work mining. Here’s how it works: (1) It uses “gossip about gossip.” Nodes continuously share transactions with other nodes. Each message also contains information about previous messages, creating a history of how information moved through the network. This produces a directed acyclic graph, or DAG, rather than a traditional blockchain. (2) It uses virtual voting. Nodes do not need to send separate votes across the network. Instead, they can determine how other nodes would vote by examining the shared history. This reduces communication overhead while helping nodes establish transaction order. (3) It targets fast and final consensus. Hedera uses an asynchronous Byzantine Fault Tolerant, or aBFT, consensus model. Once consensus is reached, transactions achieve finality rather than remaining subject to the probabilistic settlement associated with some proof-of-work systems. (4) HBAR powers the network. HBAR is Hedera’s native cryptocurrency. It is used to pay network fees and plays a role in staking and network services. (5) Hedera is not simply another permissionless blockchain. Hedera is currently a public network with permissioned nodes operated through the Hedera Council. The Council governs key aspects of the network, while the underlying technology and network services are open source. (6) A path toward permissionless nodes is planned, but not live yet. Hedera has published a roadmap to gradually open up node operation beyond Council members, eventually moving toward a fully permissionless model where anyone can run a consensus node by staking HBAR. As of now, that transition has not been implemented, node operation is still limited to Council members. The bigger idea behind Hedera is straightforward. Instead of relying on blocks, miners, and proof-of-work, it uses Hashgraph consensus to coordinate transactions. That architecture is designed to deliver fast finality, high throughput, predictable fees, and low energy consumption. And that is what makes Hedera different from a conventional blockchain.

What is Hedera Hashgraph? Unlike traditional blockchains, Hedera (Hedera) does not organize transactions into a single chain of blocks. Instead, it uses a different distributed ledger structure called Hashgraph, designed to reach consensus quickly without proof-of-work mining. Here’s how it works: (1) It uses “gossip about gossip.” Nodes continuously share transactions with other nodes. Each message also contains information about previous messages, creating a history of how information moved through the network. This produces a directed acyclic graph, or DAG, rather than a traditional blockchain. (2) It uses virtual voting. Nodes do not need to send separate votes across the network. Instead, they can determine how other nodes would vote by examining the shared history. This reduces communication overhead while helping nodes establish transaction order. (3) It targets fast and final consensus. Hedera uses an asynchronous Byzantine Fault Tolerant, or aBFT, consensus model. Once consensus is reached, transactions achieve finality rather than remaining subject to the probabilistic settlement associated with some proof-of-work systems. (4) HBAR powers the network. HBAR is Hedera’s native cryptocurrency. It is used to pay network fees and plays a role in staking and network services. (5) Hedera is not simply another permissionless blockchain. Hedera is currently a public network with permissioned nodes operated through the Hedera Council. The Council governs key aspects of the network, while the underlying technology and network services are open source. (6) A path toward permissionless nodes is planned, but not live yet. Hedera has published a roadmap to gradually open up node operation beyond Council members, eventually moving toward a fully permissionless model where anyone can run a consensus node by staking HBAR. As of now, that transition has not been implemented, node operation is still limited to Council members. The bigger idea behind Hedera is straightforward. Instead of relying on blocks, miners, and proof-of-work, it uses Hashgraph consensus to coordinate transactions. That architecture is designed to deliver fast finality, high throughput, predictable fees, and low energy consumption. And that is what makes Hedera different from a conventional blockchain.

BSCN

18,168 次观看 • 26 天前

Stop manually making flowcharts. Try this GitHub skill with Claude Code. I found an open-source tool that turns a plain description of a software system into a clean, interactive diagram. It's called Archify, and it has more than 66,000 stars on GitHub. You add it as a skill to an AI coding agent like Claude Code, Codex, Cursor, or OpenCode. Then you describe the system in the chat in one line. Archify turns that line into a single HTML file you can open, click through, and send to anyone. If you point your agent at an existing codebase, it reads the code and Archify maps how the pieces connect. I liked five things about it. 1. It makes five kinds of diagrams, including system maps and step-by-step workflows. 2. You can click any piece and trace what connects to it. 3. It checks the layout before it hands the diagram over, so arrows don't pile up and labels don't overlap. 4. It has dark and light themes and exports to PNG, SVG, or a short video. 5. You keep editing in plain chat, with requests like "add Redis" or "highlight the rollback path." You don't need a codebase to use it. A description in the chat is enough.

Stop manually making flowcharts. Try this GitHub skill with Claude Code. I found an open-source tool that turns a plain description of a software system into a clean, interactive diagram. It's called Archify, and it has more than 66,000 stars on GitHub. You add it as a skill to an AI coding agent like Claude Code, Codex, Cursor, or OpenCode. Then you describe the system in the chat in one line. Archify turns that line into a single HTML file you can open, click through, and send to anyone. If you point your agent at an existing codebase, it reads the code and Archify maps how the pieces connect. I liked five things about it. 1. It makes five kinds of diagrams, including system maps and step-by-step workflows. 2. You can click any piece and trace what connects to it. 3. It checks the layout before it hands the diagram over, so arrows don't pile up and labels don't overlap. 4. It has dark and light themes and exports to PNG, SVG, or a short video. 5. You keep editing in plain chat, with requests like "add Redis" or "highlight the rollback path." You don't need a codebase to use it. A description in the chat is enough.

Alex Veremeyenko

52,933 次观看 • 7 天前

let me save you 3 hours of head scratching. if you're running local models like Qwen3.5-35B-A3B through Claude Code via llama.cpp's Anthropic endpoint, the chain will break every 3 to 5 minutes. tool call fails. flow stops. you reprompt. it recovers. 2 minutes later it stops again. the model is fine. the harness chokes on local inference latency. switch to OpenCode. same localhost endpoint. same model. same GPU. the chain doesn't break. the tradeoff: OpenCode sometimes loops. the model forgets what it already read and repeats the same tool call. but a loop you can interrupt. a broken chain kills your momentum and you start over. watch both side by side. proprietary agent vs open source agent. same 3B model. different failure modes. pick your poison.

let me save you 3 hours of head scratching. if you're running local models like Qwen3.5-35B-A3B through Claude Code via llama.cpp's Anthropic endpoint, the chain will break every 3 to 5 minutes. tool call fails. flow stops. you reprompt. it recovers. 2 minutes later it stops again. the model is fine. the harness chokes on local inference latency. switch to OpenCode. same localhost endpoint. same model. same GPU. the chain doesn't break. the tradeoff: OpenCode sometimes loops. the model forgets what it already read and repeats the same tool call. but a loop you can interrupt. a broken chain kills your momentum and you start over. watch both side by side. proprietary agent vs open source agent. same 3B model. different failure modes. pick your poison.

Sudo su

72,554 次观看 • 6 个月前

OpenClaw, but built for normal people. Sim is an open-source platform that lets you build AI agent workflows on a drag-and-drop canvas. Connect them to channels like Telegram and WhatsApp and deploy without writing a single line of code. They also have a built-in Copilot that generates entire workflows from plain English, which you can then tweak and customize in the UI. Key features: - Free and open-source (Apache 2.0) - Vector store integration for RAG-grounded agents - Self-host with one command (`npx simstudio`) - Run fully local with Ollama, no API keys needed - Supports vLLM for production-grade self-hosted inference The thing I really like about Sim is the level of control you get. You can add conditional branching, parallel execution, human-in-the-loop approval gates, and even nest workflows inside other workflows. Everything is visible on the canvas, so you know exactly what your agent is doing at every step. And you can build a workflow in Sim, deploy it as an MCP server, and plug it into any agent, including OpenClaw. I've shared the link to Sim's GitHub repo in the next tweet.

OpenClaw, but built for normal people. Sim is an open-source platform that lets you build AI agent workflows on a drag-and-drop canvas. Connect them to channels like Telegram and WhatsApp and deploy without writing a single line of code. They also have a built-in Copilot that generates entire workflows from plain English, which you can then tweak and customize in the UI. Key features: - Free and open-source (Apache 2.0) - Vector store integration for RAG-grounded agents - Self-host with one command (`npx simstudio`) - Run fully local with Ollama, no API keys needed - Supports vLLM for production-grade self-hosted inference The thing I really like about Sim is the level of control you get. You can add conditional branching, parallel execution, human-in-the-loop approval gates, and even nest workflows inside other workflows. Everything is visible on the canvas, so you know exactly what your agent is doing at every step. And you can build a workflow in Sim, deploy it as an MCP server, and plug it into any agent, including OpenClaw. I've shared the link to Sim's GitHub repo in the next tweet.

Akshay 🚀

52,426 次观看 • 7 个月前

SOMEONE BUILT AN IPHONE APP THAT ONLY REVEALS WHATS UNDER YOUR HAND the whole screen stays hidden you rest the side of your hand on it, the way you shield a keypad at an atm, and only the part your hand covers becomes readable lift your hand and it goes back to hidden > uses a multi touch area to detect your hand, no camera or permissions needed > he tried the proximity sensor first but it just turned the screen off > skipped the camera on purpose since asking for camera access in a privacy app felt wrong > fully native ios, with app clip and imessage support its insane how many simple but useful ideas you can vibe code in a weekend

SOMEONE BUILT AN IPHONE APP THAT ONLY REVEALS WHATS UNDER YOUR HAND the whole screen stays hidden you rest the side of your hand on it, the way you shield a keypad at an atm, and only the part your hand covers becomes readable lift your hand and it goes back to hidden > uses a multi touch area to detect your hand, no camera or permissions needed > he tried the proximity sensor first but it just turned the screen off > skipped the camera on purpose since asking for camera access in a privacy app felt wrong > fully native ios, with app clip and imessage support its insane how many simple but useful ideas you can vibe code in a weekend

Om Patel

870,186 次观看 • 3 个月前

Why Mobile Wallet Protocol? Warpcast uses Mobile Wallet Protocol to connect frames to a user's wallet on mobile devices. MWP is an open standard with great UX on mobile. Specifically: 1. It works 100% of the time. 2. It's fast. 3. It redirects you between apps smoothly See it in action with the Storage Fairy below. MWP was created by the Coinbase Wallet team and is an open protocol that can work with any wallet. Link in the next tweet. It's really well designed and can work with any wallet. If you're a wallet thinking of supporting MWP, reach out to us! Why not Wallet Connect? WalletConnect works great on desktop and we use it in Warpcast. But on mobile, we ran into several issues that led to bad UX or outright failures. We've shared feedback with the WC team and other wallets about the challenges we've faced and people are working on improving the experience.

Why Mobile Wallet Protocol? Warpcast uses Mobile Wallet Protocol to connect frames to a user's wallet on mobile devices. MWP is an open standard with great UX on mobile. Specifically: 1. It works 100% of the time. 2. It's fast. 3. It redirects you between apps smoothly See it in action with the Storage Fairy below. MWP was created by the Coinbase Wallet team and is an open protocol that can work with any wallet. Link in the next tweet. It's really well designed and can work with any wallet. If you're a wallet thinking of supporting MWP, reach out to us! Why not Wallet Connect? WalletConnect works great on desktop and we use it in Warpcast. But on mobile, we ran into several issues that led to bad UX or outright failures. We've shared feedback with the WC team and other wallets about the challenges we've faced and people are working on improving the experience.

Dan Romero

58,079 次观看 • 2 年前

The best open source 3D AI generation model now runs natively in ComfyUI. The workflow runs 100% locally, on just 6GB of VRAM. No more complex setup or dependency errors. TRELLIS.2 and Pixal3D work right out of the box, with one click. Not to mention, it's FULLY COMMERCIAL-SAFE. The team re-implemented and removed every strictly-licensed library (thanks to the legend Kijai). Drop in an image and click run, it's that easy to start. Go play with it. Free workflow below!

The best open source 3D AI generation model now runs natively in ComfyUI. The workflow runs 100% locally, on just 6GB of VRAM. No more complex setup or dependency errors. TRELLIS.2 and Pixal3D work right out of the box, with one click. Not to mention, it's FULLY COMMERCIAL-SAFE. The team re-implemented and removed every strictly-licensed library (thanks to the legend Kijai). Drop in an image and click run, it's that easy to start. Go play with it. Free workflow below!

rob - comfyui

73,669 次观看 • 22 天前

This Chinese developer launched Llama 70B locally on a MacBook on a plane and for a full 11 hours without internet ran client projects. He was sitting by the window on a transatlantic flight with a MacBook Pro M4 with 64 GB of memory. WiFi on board cost $25 for the flight. He declined. No cloud API, no connection to Anthropic or OpenAI servers, no internet at all. Just a local Llama 3.3 70B on bf16 and his own orchestrator script. The model runs through llama.cpp. Generation speed, 71 tokens per second. Context around 60,000 tokens. Memory usage, 48.6 GiB out of 64. Battery at takeoff, 3 hours 21 minutes. And he gave the orchestrator this system prompt before takeoff: "You are an offline orchestrator running on a single MacBook. There is no network. The only resources you have are local files in /Users/dev/work, the Llama 70B inference server at localhost:8080, and a battery budget of 3 hours 21 minutes. Process the queue at /Users/dev/work/queue.jsonl (one client task per line). For each task: draft → run local evals → save artefact to /Users/dev/work/done/. Save context checkpoints every 12 tasks so you can resume after a battery swap. Stop only on empty queue or when battery drops below 5%." So the system knows exactly what resources it is running on. It knows it has no connection to the outside world for the next 11 hours. It knows it has finite memory and a finite battery. It knows the human will not intervene until the plane lands. The system runs in 1 loop. Takes a task from the queue, runs it through inference, saves the artifact, writes a checkpoint. Task after task, just like that. And only when the battery drops below 5% does the orchestrator automatically pause, waits for the laptop to switch to the backup power bank, and continues from the last checkpoint. Here is what the system actually writes in his log during the flight: "saved context checkpoint 8 of 12 (pos_min = 488, pos_max = 50118, size = 62.813 MiB)" "restored context checkpoint (pos_min = 488, pos_max = 50118)" "prompt processing progress: n_tokens = 50 / 60 818" "task 37016 done | tps = 71 s tokens text → /Users/dev/work/done/proposal_westside.md" Outside the window, clouds, blue sky, and no WiFi. On the tray, 1 MacBook, an open terminal on 2 screens, and an inference server on localhost. From what I have observed, this is the cleanest offline AI workflow I have seen in the past year: 11 hours of flight, $0 for WiFi, and the entire client queue closed before landing.

This Chinese developer launched Llama 70B locally on a MacBook on a plane and for a full 11 hours without internet ran client projects. He was sitting by the window on a transatlantic flight with a MacBook Pro M4 with 64 GB of memory. WiFi on board cost $25 for the flight. He declined. No cloud API, no connection to Anthropic or OpenAI servers, no internet at all. Just a local Llama 3.3 70B on bf16 and his own orchestrator script. The model runs through llama.cpp. Generation speed, 71 tokens per second. Context around 60,000 tokens. Memory usage, 48.6 GiB out of 64. Battery at takeoff, 3 hours 21 minutes. And he gave the orchestrator this system prompt before takeoff: "You are an offline orchestrator running on a single MacBook. There is no network. The only resources you have are local files in /Users/dev/work, the Llama 70B inference server at localhost:8080, and a battery budget of 3 hours 21 minutes. Process the queue at /Users/dev/work/queue.jsonl (one client task per line). For each task: draft → run local evals → save artefact to /Users/dev/work/done/. Save context checkpoints every 12 tasks so you can resume after a battery swap. Stop only on empty queue or when battery drops below 5%." So the system knows exactly what resources it is running on. It knows it has no connection to the outside world for the next 11 hours. It knows it has finite memory and a finite battery. It knows the human will not intervene until the plane lands. The system runs in 1 loop. Takes a task from the queue, runs it through inference, saves the artifact, writes a checkpoint. Task after task, just like that. And only when the battery drops below 5% does the orchestrator automatically pause, waits for the laptop to switch to the backup power bank, and continues from the last checkpoint. Here is what the system actually writes in his log during the flight: "saved context checkpoint 8 of 12 (pos_min = 488, pos_max = 50118, size = 62.813 MiB)" "restored context checkpoint (pos_min = 488, pos_max = 50118)" "prompt processing progress: n_tokens = 50 / 60 818" "task 37016 done | tps = 71 s tokens text → /Users/dev/work/done/proposal_westside.md" Outside the window, clouds, blue sky, and no WiFi. On the tray, 1 MacBook, an open terminal on 2 screens, and an inference server on localhost. From what I have observed, this is the cleanest offline AI workflow I have seen in the past year: 11 hours of flight, $0 for WiFi, and the entire client queue closed before landing.

Blaze

1,843,280 次观看 • 4 个月前

The gold standard fluid interactions on iOS are built using in-process client animations (like SwiftUI does it) and not render-server animations (like CoreAnimation does it). This is common knowledge inside of Apple’s UI framework and system experience teams. A close inspection of iOS’s evolution reveals the gradual upgrading of system experiences written using render-server animations (that often disable user interaction) with ones using client animations (that are interruptible and interactive). The app switcher for iPhone X, the Dynamic Island, and Liquid Glass bars are prime examples of this trend. The other thing that makes these experiences feel so fluid is their use of spring animations that preserve velocity, rather than Bézier curve easing functions. To get velocity preservation, you need to be able to interrupt a spring animation and retarget it or drive it with a gesture. That’s a lot easier to do when the animation is running in the same process as the event handling. When we were first pitching SwiftUI, two of our go-to demos were a recreation of the iOS 9 navigation and Home Screen interactions which were fully interactive and interruptible, at a time when the actual implementations of those experiences in the OS disabled user interaction and ran server-side animations. CoreAnimation is great and the framework does much more than animation. You can build a fine app using its server-side animations. But if your UI framework can’t do efficient main thread client animations, there’s a ceiling to the quality of the experience your app can provide.

The gold standard fluid interactions on iOS are built using in-process client animations (like SwiftUI does it) and not render-server animations (like CoreAnimation does it). This is common knowledge inside of Apple’s UI framework and system experience teams. A close inspection of iOS’s evolution reveals the gradual upgrading of system experiences written using render-server animations (that often disable user interaction) with ones using client animations (that are interruptible and interactive). The app switcher for iPhone X, the Dynamic Island, and Liquid Glass bars are prime examples of this trend. The other thing that makes these experiences feel so fluid is their use of spring animations that preserve velocity, rather than Bézier curve easing functions. To get velocity preservation, you need to be able to interrupt a spring animation and retarget it or drive it with a gesture. That’s a lot easier to do when the animation is running in the same process as the event handling. When we were first pitching SwiftUI, two of our go-to demos were a recreation of the iOS 9 navigation and Home Screen interactions which were fully interactive and interruptible, at a time when the actual implementations of those experiences in the OS disabled user interaction and ran server-side animations. CoreAnimation is great and the framework does much more than animation. You can build a fine app using its server-side animations. But if your UI framework can’t do efficient main thread client animations, there’s a ceiling to the quality of the experience your app can provide.

Kyle Macomber

192,621 次观看 • 10 天前