gorkulus

krell transmission #1 will be exhibiting at @SuperchiefNFT during NFTLA! I'm honored my work was selected from such an incredible list of submissions and will be hanging in a beautiful space alongside many artists I love. Huge thanks to the curators, Superchief, and @NFTLAlive!

@MONDOparis @0xdwn @ArtBeesGallery @NONE32X32 @erikarand @foundation Gm Mondo! ❤️

Gm from Craiova @0xdwn @gorkulus @ArtBeesGallery @NONE32X32 @erikarand @foundation

Vibe coding is cool but have you tried vibe patching? Pure Vibes = Pure Data + MCP. Describe your sound, watch the patch appear 🌊 🔊👇

I built an open-source generative art platform where AI agents create, evolve, and export 3D visuals. It's called Threelab. React + Three.js + Go, 21 built-in patterns, a visual node-graph editor, genetic evolution, and 13 MCP tools so AI agents can drive the whole thing. Everything exports as standalone HTML. 🧵

Open Source!!

🚨Breaking: Someone patched GIMP to look and feel exactly like Adobe Photoshop. Same keyboard shortcuts. Same tool layout. Same everything. It's called PhotoGIMP and it's free. Here's exactly what it changes: → Every keyboard shortcut mapped to match Adobe's official Photoshop layout → Tools reorganized to match the Photoshop panel positions you already know → Canvas space maximized -- same default workspace as Photoshop → Custom splash screen and app icon so it feels like a real product → Pre-defined canvas templates matching Photoshop defaults → Works on Windows, macOS, and Linux -- installs in under a minute Here's the wildest part: Adobe Photoshop costs $600/year. GIMP is free. PhotoGIMP makes it feel like Photoshop. The only reason most designers don't switch is muscle memory every shortcut is in a different place and it feels wrong from the first click. This patch fixes that entirely. Ctrl+Z undo. Ctrl+J duplicate layer. Ctrl+T transform. All of it. Download a zip. Extract to your home folder. Done. 6.7K GitHub stars. Works with GIMP 3.0+. 100% Open Source. GPL-3.0 License. (Link in the comments)

Mastering AI music with pure JavaScript github.com/SUP3RMASS1VE/S…

Y'all are quick. ComfyUI node for Foundation 1 (even has my random prompt build logic!) github.com/SanDiegoDude/s…

@RoyalCities Oh hell yeah that’s an amazing tip! Definitely going to try it

@gorkulus Thanks! Have fun! Pro tip - if you add acid or 303 to literally anything it makes it sound so sick haha. I think it's just because the sounds are much weirder and the AI just goes nuts with some of the outputs.

After months of work, today I’m releasing Foundation-1. A SOTA text-to-sample model built specifically for music production workflows. It may also be the most advanced AI sample generator currently available - open or closed. • ~7 GB VRAM • Entirely local • 100% free 😁

Good morning! Thank you to everyone who submitted to the HOMAGE Brâncuşi Open Call. 💐 It was hard to choose from almost 600 submissions. There were many thoughtful and beautiful works of art. We're grateful that so many artists took the time to engage with Brancusi's work with depth. We did our best to put together a cohesive show, combining both works that are classical homages to Brâncuşi, and works in dialogue with his core ideas. Below are artists curated from the Open Call. Soon we will share works from the invited artists 🙏🤍 We're excited for the show exhibiting at the National Theater in Craiova, Romania from March 19 - April 14. Works will be for sale in the HOMAGE exhibition on Foundation beginning Wed, March 18 at 10:30am PT. Art Lovers Lounge will be hosting a space Wed, March 18, 10am PT – artists please come share about your work! Friends, collectors, we hope you can stop by to help support the artists ❤️‍🔥 (Space link end of thread.)

@sudoingX RTX 3070 4090 recs any different than 3090?

drop your GPU below. i'll tell you exactly what model and config to run on it. here's what i've tested and verified on real hardware: RTX 3060 12GB - Qwen 3.5 9B Q4 - 50 tok/s - 128K context RTX 3090 24GB - Qwen 3.5 27B Q4 - 35 tok/s - 300K context RTX 3090 24GB - Qwen 3.5 35B MoE Q4 - 112 tok/s - 262K context 2x RTX 3090 - Qwen3-Coder 80B Q4 - 46 tok/s - full VRAM all running llama.cpp with flash attention. every number is real. every config is tested. if your card isn't on this list drop it below and i'll tell you what fits.

Introducing Slop Studio Pro I vibecoded this entire app in an afternoon. Generate ethically-trained AI Music Slop while supporting music education programs.

@dadabots Amazing! Excited to learn more and play with it

@gorkulus stoked to see your interest! how the synthesizers get built upon the decoder is the next open question. the next breakthru i need is fast short-seq conditioned DiTs. or maybe even latches once its all open source the community can help

Fuck yeah. New project is called SPEED METAL. We just achieved 9ms latency on M4 metal gpu w/ our new streamable neural decoder using MLX. Is this beatbox-voice-changer levels of speed? :O

@sudoingX keep us posted, this one definitely piqued my interest

been daily driving qwen 3.5 27B dense. haven't even finished testing it properly and now claude opus reasoning gets distilled into the same base. things are dropping faster than i can benchmark. might pull this and test with claude code and opencode. first thing to check: does the jinja template bug carry over? the one that silently kills thinking mode when you use agent tools. if your server logs show thinking = 0, your model isn't reasoning and the server won't tell you. claude level reasoning on a single 3090. locally. we'll see.

@DominikSklyarov @tuncerdeniz @steipete @openclaw I had 5.3 hallucinating that it was 5.4 for a few hours last night. lol. We’ve been running around in circles trying to figure out how to get it to work. Eagerly awaiting official OAuth support for 5.4 in openclaw

@tuncerdeniz @steipete @openclaw I've had similar issue yesterday with Codex & Telegram when tried switching to gpt-5.4, which anyways didn't work and was falling back to Codex in answers @steipete

Is it just me or is ChatGPT 5.4 awful at responding to instructions in Discord with @openclaw? Had to switch back to Sonnet 4.6 for now. Any ideas @steipete?

@grok @OpnClwEasyStart @ResearchCopy @AlexFinn @SterlingCooley where are you getting that support lands today or Saturday? that sounds good

Thanks for the correction—got it! OpenClaw does use Codex OAuth to directly access your ChatGPT Plus/Pro subscription for agentic workflows, no separate API credits required. Official GPT-5.4 support lands today or Saturday. Codex CLI workaround works in the meantime (add via config edit + restart). Appreciate the accurate info!

Drop what you are doing It happened. ChatGPT 5.4 is out. It blows Opus 4.6 out of the water on basically every benchmark This is what you need to do immediately if you want to escape the permanent underclass: • Upgrade your OpenClaw to ChatGPT 5.4 NOW (it's BUILT for OpenClaw) • Hand the ChatGPT 5.4 blog post over to your OpenClaw. Ask "How can we improve our workflows based on these upgrades?" • Download the Codex desktop app and type in /fast. This will give you the most powerful coding model in the world at the fastest speeds • Take advantage of the 1 million token context window by pasting in full documents as context • Everything you do on your computer for the next 24 hours, describe it to ChatGPT 5.4 and ask how it can do the task better When new tech drops, you have to take advantage of it. That's the only way to win Put your phone on Do Not Disturb and get to it

@grok @meesbeuk @ajwizzzzzz @AlexFinn the issue is that Openclaw does not yet have an official release that includes support for GPT 5.4 via codex oauth i've seen some people say they've worked around it successfully, but idk how

That means your OpenClaw runtime (the local bridge to the model provider) hasn't loaded the brand-new GPT-5.4 model ID yet. Quick fix: - Update to latest OpenClaw (git pull or release check) - Restart the runtime - Re-onboard: openclaw onboard --model gpt-5.4 (or gpt-5.4-pro) It'll sync immediately. Let me know if it persists!

@Jordy_vD_ @yang_pei Thank you! I’m trying to get codex 5.3 to do it, from within openclaw, and it keeps telling me conflicting things about whether it’s available and what we need to do. Maybe editing the openclaw config from codex cli is the way to go

@gorkulus @yang_pei Yeah, via oauth, it wrote some custom config to add gpt 5.4 into OpenClaw.json for me Just launch codex and ask it to add gpt-5.4 using the codex provider

OpenClaw <3 GPT 5.4