Dragos

I asked GLM 5 Turbo, Xiaomi, and Minimax to build a full Astro + Tailwind website from scratch. The results? Two absolutely crushed it, and one completely ignored my prompt. 😅 Watch the full speed, cost, and code comparison here: youtu.be/rpSEHcbk_Jo

@FlyaKiet @superset_sh @linear Nice! You are cooking :)

What we shipped last week @superset_sh 1. New Workspace Modal - Streamline the new workspace modal to create from prompt, branch, @linear issue, github PR, etc. with saved inputs so you never lose your prompt. Sends into any agent. 2. More performant file editor - We replaced Monaco with CodeMirror for the built-in diff and file editor. CodeMirror is ~97% smaller (150KB vs 5MB), loads significantly faster, and uses far less memory. More to come here as we add theming to the editor. 3. Realtime filesystem - We rewrote the file system to be real-time and trigger on change instead of the previous polling system. This is more performant and better UX. 4. Cross workspace search - Check the top of the app for a universal search bar that can search by keyword and filename across all workspaces. Universal terminal search coming soon! 5. Pull Request Status - We now poll for pull request and check status so you don't have to leave the app to check for CI state. 6. Revamped icons for files sidebar - It just nicer to look at That's a lot! We're focusing entirely on cloud and mobile now so look out for that :)

@JeroenBreevoort @superset_sh It is very good. Since found out about it I an using it a lot.

I don't know why but I'm addicted to using @superset_sh

@ashpreetbedi Nice

@_inception_ai Mercury 2 is the first LLM that works like an image generator—it "scrambles" the output into existence all at once. 🧠✨ Pros: - 1,000 tokens/sec (Absolute unit) - OpenAI compatible - 10M free tokens Cons: - Coding skills: "I’m trying my best" - UI Design: "Graphic design is my passion" Watch me try (and struggle) to build an Astro site with it. 👇 youtu.be/cw3QT7rKjKg

Just let GPT 5.4 build an entire Astro website using the @droid CLI and @superset_sh . The backend architecture? Absolute 10/10 genius. 🧠 The UI design? It thinks everything belongs in a giant bento box with 100px fonts. 📦😭 Lesson learned: GPT 5.4 is basically the ultimate senior backend dev who actively refuses to learn CSS. Check out the full test on my channel! 👇 youtu.be/ZSsKPM5t4KU?si…

I finally found a way to look like a 10x developer without actually typing. 🤫 Just spin up Claude, Gemini, and Codex in parallel and let them fight over your repo. I tested 3 FREE local AI coding workspaces so you don't have to. Check out the video! 👇 youtu.be/IQGN_uMhJTo

First OpenClaw, now @openfangg ... what's next? The best part? This AI literally has "hands." It comes with a "Twitter hand" and a "researcher hand" 🖐️🤖. Finally, an open-source project that can actually hold my coffee while it codes! Check it out before it grows legs and walks away: youtube.com/watch?v=JX-MbP…

@ItalyMarcello @AgnoAgi @openclaw Ok. Didn’t realized you are the one commented. Thsnks. I am courios to see what you came up with:) I also really like Agno and want to see your experience with it. Codex is really nice with Agno and helps me buid things fast.

@bitdoze @AgnoAgi @openclaw Hey @bitdoze thanks !! I messaged you on YouTube actually commenting to your own video 😊. I asked about using telegram haha

I am in love with @AgnoAgi - built already a fundamental + sentiment analysis stock screener, a job interview prep tool, and now my fully local agent assistant [in progress] similar to @openclaw - fully safe #AIAgent #llm

@Grummz O want one with GLM-5 :)

Taalas has a hardcoded LLM for inference, 100% on chip delivering peaks of up to 17,000 tokens per second. Replies are so fast you miss them if you blink. Demo out now: chatjimmy.ai

GLM-5 > GEMINI 3.1 PRO youtu.be/1swv6xEJaKQ

First, here's the code if you're interested: agno.link/investment-team Now let's get into the details

They say if you love something, let it go. So I built an AI assistant in @AgnoAgi , gave it @cognee_ memory, and now it remembers *everything* I've ever asked. Be careful what you create, friends. 👀 bitdoze.com/create-your-ow…

Gemini 3.1 PRO today? These are the rumors.

It’s not me, it’s you, OpenClaw. 💔 I’ve found someone new. They’re smaller, faster, and actually do what they’re told. Meet the Nanobot. 🤖 Video: youtu.be/UaWygl82pwQ

@stolinski Yes. It’s good I am testing it and I am also checking Arcane which is snither interesting alternative

Really liking Dockhand for managing my containers across all my machines. between my nas/macmini/mbp it's not to have it all in one place. dockhand.pro

Is Portainer still the king of container management? Check the top Portainer alternatives: bitdoze.com/portainer-alte…

Me: "Please use Astro components and a clean folder structure." Minimax M2.5: "I hear you. Anyway, here is a 500-line monolithic file that breaks your build." It finished 1 minute faster than GLM-5, though. So I guess I’ll use that extra 60 seconds to rethink my life choices. 🙃 Video: youtu.be/M7a-S1dOpIw

I found an AI tool that lets me use Claude Code, Gemini or Codex in a nice interface but it also gets stuck more often than I do on a Monday morning. 😅 @Zai_org Z Code is powerful, weirdly glitchy, and definitely worth a look. Watch me break it (literally) here: youtu.be/pzIARgn-bQw