Loading video...

Video Failed to Load

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

Building a Linux Desktop from Scratch Using a self-review loop, GLM-5.1 spent 8 hours autonomously refining features, styling, and interactions to build a functional desktop environment.

Z.ai

158,476 subscribers

145,003 views • 5 months ago •via X (Twitter)

Education Science & Technology

Anya Rossi• Live Now

Private livecam show

26 Comments

Z.ai5 months ago

SOTA on SWE-Bench Pro (58.4): GLM-5.1 delivers significant leaps in coding and agentic performance.

Z.ai5 months ago

Introducing GLM-5.1: The Next Level of Open Source - Top-Tier Performance: #1 in open source and #3 globally across SWE-Bench Pro, Terminal-Bench, and NL2Repo. - Built for Long-Horizon Tasks: Runs autonomously for 8 hours, refining strategies through thousands of iterations. Blog: Weights: API: Coding Plan: Coming to in the next few days.

Z.ai5 months ago

Vector-DB-Bench: 6x Performance Boost In high-performance database optimization, GLM-5.1 reached 21.5k QPS over 600+ iterations and 6,000+ tool calls. This is 6x the performance of a standard 50-turn session.

Z.ai5 months ago

Special thanks to our launch partners, AI gateways, and inference providers. Access GLM-5.1 now: - OpenRouter: - Vercel: - Requesty:

Saint5 months ago

yeah everything's fine till this one. Main one as orchestrator + 3 subagents and they just face with requests limit on the max plan

Zero Void5 months ago

8 hours of self-review to build a linux desktop is kind of insane. anime training arc but for window managers

Jacob Rothfield5 months ago

What method or harness are you using for the self-review loop?

tolgaozisik5 months ago

which self review loop? Could you specify this?

CEO del Socialismo de Mercado🌹🕊️ 市场社会主义CEO5 months ago

How much $$$ in tokens for that? 👀

Dustin Ogle5 months ago

What harness was this using? Like cloud code or open code or droid or something like that?

Mowg115 months ago

can somebody tell me what's the app where is glm 5.1 used in this video? Glm has a dedicated desktop app? or ? ty

Joilson Rabelo5 months ago

Wtf this is a new distro or what?

Chii5 months ago

Noe can you make auto claw work with linux?

Rudra Sarker5 months ago

@Zendiyy really impressive 🔥🔥

MrGenius5 months ago

Wow, this is the Most goted video of a model release

am.will5 months ago

#fail it built macOS 😅 jk

Alp Arslan Mirasow5 months ago

Is it really ? I don't believe it

Saquib Mehmood5 months ago

This is more than awesome.

wholecricket5 months ago

this is cool

DjPara965 months ago

@theo 0:32 🙌🙌

0xBeast5 months ago

wt @grok is this real ?

JoeyMonki5 months ago

what is a self-review loop? can you share more about this ?

arkheτ.hl5 months ago

Wow.

Kevin Lewis5 months ago

8 hours of autonomous iteration to build a full desktop is actually wild 👀 feels like we’re getting closer to AI handling real end-to-end systems, not just snippets or demos. Curious how polished it actually is in daily use 🔥

Mohamed Khedr5 months ago

wow

Shawn McAllister5 months ago

Let's see the code

Related Videos

Framework Laptop 12 unbox. Wayland and AI have made Linux desktop a lot easier to use.

Framework Laptop 12 unbox. Wayland and AI have made Linux desktop a lot easier to use.

Chun

45,518 views • 4 months ago

I didn't touch TouchDesigner myself. Hermes agent learned it from scratch and built this: → navigated my desktop with computer use → figured out how to connect to TouchDesigner → read my reference images → iterated on the art with me in a self-learning loop → then saved what it learned as a reusable skill for the next image all powered by MiniMax (official) M3 × Hermes Desktop Agent Nous Research Here's a full breakdown 📽️

I didn't touch TouchDesigner myself. Hermes agent learned it from scratch and built this: → navigated my desktop with computer use → figured out how to connect to TouchDesigner → read my reference images → iterated on the art with me in a self-learning loop → then saved what it learned as a reusable skill for the next image all powered by MiniMax (official) M3 × Hermes Desktop Agent Nous Research Here's a full breakdown 📽️

amber shen

120,162 views • 3 months ago

LEGO engineers spent over 13,000 hours and used 1 million pieces to build a fully functional, life-size Bugatti Chiron.

LEGO engineers spent over 13,000 hours and used 1 million pieces to build a fully functional, life-size Bugatti Chiron.

Massimo

101,812 views • 2 months ago

Neel built Builddy to go from a text prompt to a deployed web app in one go using GLM-5.1. The system coordinates 6 different agents to handle everything from system design to final deployment, all running on GLM-5.1's coding logic.

Neel built Builddy to go from a text prompt to a deployed web app in one go using GLM-5.1. The system coordinates 6 different agents to handle everything from system design to final deployment, all running on GLM-5.1's coding logic.

Z.ai

40,572 views • 5 months ago

Running a full desktop environment in a microVM, and accessing it remotely over LAN (Omarchy, ping DHH and vaxry 😊) QVM now supports ultra low latency desktop streaming, for purposes such as VDI / Virtual Desktop Infrastructure, RBI / Remote Browser Isolation and agent sandboxes Supported host-OS:es and hypervisors include both Linux/KVM, macOS/HVF and Windows/WHPX VirGL-acceleration is available for guests, and both X11 and Wayland based desktop environments are supported Hardware-accelerated H264 encoding is performed on the host side, avoiding the need to expose direct GPU access to the guest

Running a full desktop environment in a microVM, and accessing it remotely over LAN (Omarchy, ping DHH and vaxry 😊) QVM now supports ultra low latency desktop streaming, for purposes such as VDI / Virtual Desktop Infrastructure, RBI / Remote Browser Isolation and agent sandboxes Supported host-OS:es and hypervisors include both Linux/KVM, macOS/HVF and Windows/WHPX VirGL-acceleration is available for guests, and both X11 and Wayland based desktop environments are supported Hardware-accelerated H264 encoding is performed on the host side, avoiding the need to expose direct GPU access to the guest

Joel Eriksson

23,583 views • 28 days ago

Here's a quick demo using the Claude desktop app, where we've configured MCP: Watch Claude connect directly to GitHub, create a new repo, and make a PR through a simple MCP integration. Once MCP was set up in Claude desktop, building this integration took less than an hour.

Here's a quick demo using the Claude desktop app, where we've configured MCP: Watch Claude connect directly to GitHub, create a new repo, and make a PR through a simple MCP integration. Once MCP was set up in Claude desktop, building this integration took less than an hour.

Alex Albert

260,481 views • 1 year ago

MCP in ~2 min In ~2 min I try to explain what it is, build a MCP server from scratch, connect it to @windsurf, Anthropic, Cursor desktop app, show it working. All code and longer vid below ...

MCP in ~2 min In ~2 min I try to explain what it is, build a MCP server from scratch, connect it to @windsurf, Anthropic, Cursor desktop app, show it working. All code and longer vid below ...

Lance Martin

40,461 views • 1 year ago

sharing berd, a new open source desktop app to build with 🪿 we built this with a v small team, design + engineering working really close together, and designers building berd using berd! v proud of the fun and useful this is, and excited to keep building it out💫

sharing berd, a new open source desktop app to build with 🪿 we built this with a v small team, design + engineering working really close together, and designers building berd using berd! v proud of the fun and useful this is, and excited to keep building it out💫

Cynthia

126,346 views • 27 days ago

Bro... it's kind of crazy. You build a Linux SoC from scratch on an FPGA, get Linux running on it, and your original goal was simply to run a C64 emulator on your own Linux system. KianV StealthV.

Bro... it's kind of crazy. You build a Linux SoC from scratch on an FPGA, get Linux running on it, and your original goal was simply to run a C64 emulator on your own Linux system. KianV StealthV.

Logic Destroyer — Proven in Silicon.

11,710 views • 1 month ago

HERMES AGENT JUST GOT A FULL NATIVE DESKTOP APP. MULTI-AGENT MANAGEMENT, PERSISTENT MEMORY AND AUTONOMOUS WORKFLOWS ALL IN ONE PLACE. No more CLI only. Run your 24/7 self-evolving AI agent from a proper desktop UI.

HERMES AGENT JUST GOT A FULL NATIVE DESKTOP APP. MULTI-AGENT MANAGEMENT, PERSISTENT MEMORY AND AUTONOMOUS WORKFLOWS ALL IN ONE PLACE. No more CLI only. Run your 24/7 self-evolving AI agent from a proper desktop UI.

0xMarioNawfal

189,401 views • 3 months ago

Using Antigravity to build a Chrome extension tab organizer, plus a background tool that tags desktop files by date. Learn how to guide the agent using /grill-me for design specs, voice prompting, and inline artifact comments.

Using Antigravity to build a Chrome extension tab organizer, plus a background tool that tags desktop files by date. Learn how to guide the agent using /grill-me for design specs, voice prompting, and inline artifact comments.

Google Antigravity

91,162 views • 1 month ago

Spent the last months building LucidClip, a clipboard manager built with Flutter. Tech Stack: • Flutter desktop • Drift (SQLite) • Supabase (auth + sync infra) • Global hotkeys • Local-first architecture Now running on macOS and Windows. Flutter desktop is seriously underrated. Curious what others are building with it. Flutter #Dart #flutter

Spent the last months building LucidClip, a clipboard manager built with Flutter. Tech Stack: • Flutter desktop • Drift (SQLite) • Supabase (auth + sync infra) • Global hotkeys • Local-first architecture Now running on macOS and Windows. Flutter desktop is seriously underrated. Curious what others are building with it. Flutter #Dart #flutter

Ethiel A.

19,908 views • 6 months ago

New on Claude Code Desktop We’re shipping new features for Claude Code on desktop that let you preview running apps, auto-review code, and auto-fix and merge PRs to help close the development loop. - Server previews: Claude starts dev servers and previews your running app in the desktop interface. It reads console logs, catches errors, and keeps iterating on its own. - Local code review: Claude examines your local diffs and leaves inline comments before you push — an immediate second set of eyes on every change. - PR monitoring: Claude tracks CI status after you open a PR. With auto-fix, it attempts to resolve failures automatically. With auto-merge, PRs land as soon as checks pass. You can move on to your next task while Claude handles the last one. - Session mobility: Move sessions from CLI to desktop, and from desktop to the cloud. Start work at your desk, pick it up from the web or your phone.

New on Claude Code Desktop We’re shipping new features for Claude Code on desktop that let you preview running apps, auto-review code, and auto-fix and merge PRs to help close the development loop. - Server previews: Claude starts dev servers and previews your running app in the desktop interface. It reads console logs, catches errors, and keeps iterating on its own. - Local code review: Claude examines your local diffs and leaves inline comments before you push — an immediate second set of eyes on every change. - PR monitoring: Claude tracks CI status after you open a PR. With auto-fix, it attempts to resolve failures automatically. With auto-merge, PRs land as soon as checks pass. You can move on to your next task while Claude handles the last one. - Session mobility: Move sessions from CLI to desktop, and from desktop to the cloud. Start work at your desk, pick it up from the web or your phone.

Ado

71,910 views • 6 months ago

GLM-5.2 delivers a substantial leap in app development capabilities, which also represent demanding long-horizon tasks. Results: - GLM-5.1: 21/70 - GLM-5.2: 48/70 - Claude Fable 5: 56/70 That's more than a twofold improvement from GLM-5.1 to GLM-5.2. These come from an internal benchmark of 35 challenging mobile development tasks, each run twice for a total of 70 trials. We measured task completion, defined as core features working without major issues.

GLM-5.2 delivers a substantial leap in app development capabilities, which also represent demanding long-horizon tasks. Results: - GLM-5.1: 21/70 - GLM-5.2: 48/70 - Claude Fable 5: 56/70 That's more than a twofold improvement from GLM-5.1 to GLM-5.2. These come from an internal benchmark of 35 challenging mobile development tasks, each run twice for a total of 70 trials. We measured task completion, defined as core features working without major issues.

Zixuan Li

335,067 views • 2 months ago

What happens when AI codes autonomously for 10+ days? 🚀 Qwen3.8‑Max built, tested, verified, and repaired a self‑evolving harness, all in one long‑horizon autonomous coding run. Watch it build from scratch!

What happens when AI codes autonomously for 10+ days? 🚀 Qwen3.8‑Max built, tested, verified, and repaired a self‑evolving harness, all in one long‑horizon autonomous coding run. Watch it build from scratch!

Alibaba Group

2,205,763 views • 1 month ago

Claude Code Desktop can now run your app, catch errors, review code, and merge PRs automatically 🤯 This update closes the entire development loop — from writing to previewing to reviewing to fixing to merging — without you lifting a finger. Want the full guide? DM me.

Claude Code Desktop can now run your app, catch errors, review code, and merge PRs automatically 🤯 This update closes the entire development loop — from writing to previewing to reviewing to fixing to merging — without you lifting a finger. Want the full guide? DM me.

Julian Goldie SEO

12,984 views • 6 months ago

Custom Shortcuts on Magnific Desktop Generate a video, move it to a folder, all from one click in your dock How to build it on Mac 👇

Custom Shortcuts on Magnific Desktop Generate a video, move it to a folder, all from one click in your dock How to build it on Mac 👇

Magnific

308,140 views • 17 days ago

Wiidesk has been revealed! It's a community-driven project to recreate the Nintendo Wii Menu experience as an interactive desktop for PC! It aims to launch apps and websites directly from the interface. A Linux version is planned!

Wiidesk has been revealed! It's a community-driven project to recreate the Nintendo Wii Menu experience as an interactive desktop for PC! It aims to launch apps and websites directly from the interface. A Linux version is planned!

Reticent

17,972 views • 10 days ago

I’ve been running Ubuntu on my second machine for over a decade. After seeing another breakdown of Omarchy, this time from NetworkChuck, I’m fully making the switch today. He changed the entire desktop theme in seconds, used an AI agent to build a Quickshell plugin, handed a crashed process to the agent for diagnosis, ran games through Proton, and launched a Windows VM when Linux wasn’t enough. Omarchy isn’t just Arch with a nice desktop. It starts with Arch, Hyprland, Quickshell, and a keyboard-first workflow, then gives you an environment you can continue reshaping around how you work (the linux boys calls this ricing). The operating system can become another surface your agents can help you modify. Ubuntu had a great run. Omarchy loading…

I’ve been running Ubuntu on my second machine for over a decade. After seeing another breakdown of Omarchy, this time from NetworkChuck, I’m fully making the switch today. He changed the entire desktop theme in seconds, used an AI agent to build a Quickshell plugin, handed a crashed process to the agent for diagnosis, ran games through Proton, and launched a Windows VM when Linux wasn’t enough. Omarchy isn’t just Arch with a nice desktop. It starts with Arch, Hyprland, Quickshell, and a keyboard-first workflow, then gives you an environment you can continue reshaping around how you work (the linux boys calls this ricing). The operating system can become another surface your agents can help you modify. Ubuntu had a great run. Omarchy loading…

Hybrid 🏃🏾

34,818 views • 24 days ago