Kevin Kuipers

Co-Founder & Partner @ Galion.exe (early-stage fund) & the TTY community. Also co-founded Gamekult, SensCritique, Molotov·tv, and former EiR at daphni.

[Johan Peitz - picoCAD 2 is out! @johanpeitz]

I had already quit my day job when I started working on #picoCAD2, and now it is on sale on Steam and itch io! #3d model, #pixelart, and animate - all in one cosy retro package!
Links in replies! 👇

[Dan Greenheck @dangreenheck]

"AAA graphics aren't possible in the browser"

Hold my beer 🍺

[Yakov @yak32]

I finally finished my Gaussian Splat based FPS demo.

It's a @playcanvas project, runs in a browser. On a real photoscan. With physics, baked lighting, pathfinding NPCs.

Here's how 👇

[Indie Game Joe @IndieGameJoe]

This solo indie dev spent 10 years making an RPG game that looks 3D… but is actually 2D.

- Camera shifts from orthographic to perspective views
- Depth is simulated using parallax & warped sprites
- Hand-painted semi-open world

Would you play this? It’s called Bridgebourn.

[Mishig ᠮᠢᠰᠾᠢᠭ @mishig25]

The era of fine-tuned models is here! 🚀
We’re seeing three trendlines come together:

1️⃣ Fantastic open base models (Qwen, Kimi, ...) to fork from
2️⃣ Great inference providers (Fireworks, Groq, Cerebras, ...)
3️⃣ Awesome tooling (TRL, vLLM, ...)

Exciting times ahead! 🔥

[Kevin Kuipers @nanark_]

My weekly take on tech #finetuning
> After years on the sidelines, fine-tuning may be entering a renaissance. New tools, new approaches, and a growing desire for control could spark a shift from generic to bespoke AI.

https://t.co/OumKJdAOgF

[Kevin Kuipers @nanark_]

I'd love to stay on @qoder_ai_ide but it keeps hitting 100+ % CPU after 10-30 minutes leading to super slow chat 😭 (M1, 32 Go RAM, Sequoia)

[Kevin Kuipers @nanark_]

@DarioDiCarlo2 @supabase Amazing job bro!

[kyutai @kyutai_labs]

🚀 Thrilled to announce Helium 1, our new 2B-parameter LLM, now available alongside dactory, an open-source pipeline to reproduce its training dataset covering all 24 EU official languages. Helium sets new standards within its size class on European languages!

[Kevin Kuipers @nanark_]

Congrats @amtellezfdez !

[Alba María Téllez Fernández @leader_arm]

We built a urdf visualizer for fun one weekend :) Tryna actually make something useful for the robotics homies ! so fr… what tool would you love to have? 👀

[Kevin Kuipers @nanark_]

Congrats! 🔥🔥🔥

[Alba María Téllez Fernández @leader_arm]

🏆 We just won 1st place at the largest hackathon in Europe (@huggingface x @MistralAI) !!! Meet SAND-E !!!! ❤️✨

[Kevin Kuipers @nanark_]

Attention, please! @kr0niker is about to speak at 10 AM PST (7 PM CET) on how to benchmark LLMs the smart way. As a brilliant professor, doctor, and speaker, @kr0niker has prepared plenty of great content (and memes). Register now!
https://t.co/k4Nx8nKBRj

[Kevin Kuipers @nanark_]

Our third SOTA session is about to begin. Laurent Mazare from @kyutai_labs will share his exploration of Speech-to-Speech AI models, featuring Moshi and Hibiki. Join us now!

https://t.co/EPfx9oQ0hQ

[OpenAI Developers @OpenAIDevs]

MCP 🤝 OpenAI Agents SDK

You can now connect your Model Context Protocol servers to Agents: https://t.co/6jvLt10Qh7

We’re also working on MCP support for the OpenAI API and ChatGPT desktop app—we’ll share some more news in the coming months.

[Kevin Kuipers @nanark_]

Can't wait!

[Xuan-Son Nguyen @ngxson]

On Monday, the 24th, I'm proud to give a talk at @SotaFamily 's webinar.

My main talk will last for an hour to deep dive into the current state of on-device LLMs, exploring their advantages, performance trade-offs, and limitations.

The session will conclude with an open Q&A, where you can ask me anything related to this subject.

[Shubham Saboo @Saboo_Shubham_]

Figma now has an MCP server to talk to Cursor.

This MCP plugin allows Cursor to communicate with Figma to read UI designs and modify them programmatically.

100% free and opensource.

[Kevin Kuipers @nanark_]

On Monday, the 24th, we at @SotaFamily are proud to host our second webinar. This time, @ngxson from @huggingface will deep dive into the current state of on-device LLMs
👉👉👉 Register here: https://t.co/qvLlX4tCsn

[Kevin Kuipers @nanark_]

What are you doing today? At Sota, we are hosting a webinar with @hcompany_ai, @Linkup_platform, @meilisearch and https://t.co/bvdNdg7bCn to share their xp with @AnthropicAI MCP.

It's free at, and it's at
🇺🇸 11 AM SF, 8 AM NYC
🇫🇷 17h Paris
🇬🇧 4 PM London

https://t.co/dabhWf0sAL

[Kevin Kuipers @nanark_]

@mishig25 @galionexe Thank you, @mishig25, for joining our party with our Discord Reg.exe tech-only community! Bringing together all these talented individuals from across Europe and connecting them in our space was truly a delightful and valuable evening. 🔥🔥🔥

[Kevin Kuipers @nanark_]

This is impressive! Great job @kyutai_labs! 🔥🔥🔥
(🎧 listen in stereo)

[kyutai @kyutai_labs]

Meet Hibiki, our simultaneous speech-to-speech translation model, currently supporting 🇫🇷➡️🇬🇧.
Hibiki produces spoken and text translations of the input speech in real-time, while preserving the speaker’s voice and optimally adapting its pace based on the semantic content of the source speech.
Based on objective and human evaluations, Hibiki outperforms previous systems for quality, naturalness and speaker similarity and approaches human interpreters. 🧵