Till Felippi

859 posts

Till Felippi

@t1llmann

Building @lumenfall_ai / Writing about everything related to AI / Focussed on media models / Engineer / Serial founder / Father

Schweiz شامل ہوئے Kasım 2010

231 فالونگ183 فالوورز

پن کیا گیا ٹویٹ

Till Felippi@t1llmann·27 Oca

Today we are launching @lumenfall_ai, built by @gianwirth and me over the last 2 months 🚀 If you’re shipping anything with AI images, you’ve felt it: different APIs, inconsistent outputs, weird edge cases, outages. Lumenfall is our answer: an AI media model gateway for developers. Starting with image generation + editing: → top models in one place 🖼️ → all major providers + automatic failover 🔀 → OpenAI-compatible API 🔌 → zero markup, pay what providers charge 💵 fal and replicate are great. We’re not competing with them, we can route to them. We focus on making it boring and reliable: normalize behavior, reduce edge cases, keep your existing SDKs working. $1 free credit on signup to test anything 🪙 lumenfall.ai What feature do you want next?

English

688

Till Felippi@t1llmann·1h

Claude Code getting closer to 🦞 every day. You can now control it via Telegram and Discord

Thariq@trq212

We just released Claude Code channels, which allows you to control your Claude Code session through select MCPs, starting with Telegram and Discord. Use this to message Claude Code directly from your phone.

English

Till Felippi@t1llmann·3h

@Angaisb_ Impressive! Need to try that combo as well. Maybe it's also great with the impeccable style skills; something I can also recommend for frontend work (not affiliated)

English

Angel ❄️@Angaisb_·7h

"Make a website presenting yourself (Composer 2 Fast)" + Claude's frontend-design skill I really like this result, I've tried it a few other times and I've liked almost all of them

Angel ❄️@Angaisb_

Composer 2 is out! I’ve been using it for a few days and it’s a very capable model in my testing, and it’s also super fast If you have a Cursor sub you should really give it a try

English

1.8K

Till Felippi@t1llmann·3h

We recently launched a new image model comparison that is based on blind-voting on various prompt challenges. Could be interesting for you to check how Grok performed in there: lumenfall.ai/arena/text-to-… And if you want to do a few votes and try to see if you can spot Grok: lumenfall.ai/arena/vote/gro… We are still early, so every vote is appreciated

English

Gaurav@gauravisnotme·2d

Apart from cracking the top at the leaderboards, I am seeing a plethora of impressive creative workflows where Grok Imagine is now the primary driver for all image and video use cases. Would love to know more about how you see the current models cater to your process and what kind of improvements would you like to see the most.

Arena.ai@arena

Today we’re launching the Video Edit Arena to evaluate the frontier capability of video models! - #1 Grok-Imagine-Video, @xAI - #2 Kling-o3-pro, @Kling_ai - #3 Kling-o1-pro, @Kling_ai - #4 Gen4-aleph, @Runwayml The leaderboard is powered by thousands of real-world community votes. Click the Edit button in Video Arena to edit any video and compare top model outputs. More models coming soon!

English

2.6K

Till Felippi@t1llmann·5h

@1ssve Yes! Every time the same story. Apple releases a new watch, their introduction video really excites me, I get a watch and wear it every day, after a few weeks I slowly stop wearing it and then it‘s over

English

3.7K

S.🎧@1ssve·9h

one day you wake up and just stop wearing your Apple Watch

English

206

370

4.7K

602K

Till Felippi@t1llmann·5h

BREAKING: Cursor releases their Composer 2 model that is specialized in coding. In their CursorBench benchmark it is better than Opus 4.6 and not far away from GPT5.4 medium&high

Cursor@cursor_ai

Composer 2 is now available in Cursor.

English

Till Felippi@t1llmann·9h

@tgod34748 I just got my M5 Pro as a replacement for my old M2 Air. While I love the screen and other things of the Pro, I really struggle with how big and almost outdated it feels compared to holding the Air

English

203

Tech God@tgod34748·22h

I feel like people forget about the 15-inch MacBook Air. While the MacBook Neo and 13-inch MacBook Air have stolen the show over the past year, and the MacBook Pro also gets a ton of attention, the 15-inch is just kinda there. Kinda the sleeper hit of the lineup though.

English

381

38.3K

Till Felippi@t1llmann·9h

@garrytan and for the other coding tasks, would you prefer Opus 4.6?

English

154

Garry Tan@garrytan·19h

OK Codex is GOAT at finding bugs and finding plan errors

English

216

360.8K

Till Felippi@t1llmann·9h

@elonmusk @grok What was the prompt? We're adding video to our ai model comparison page and this could be a fun one to compare with other models

English

Elon Musk@elonmusk·18h

Generated with @Grok Imagine 🚬

English

5.8K

4.3K

44.1K

11M

Till Felippi@t1llmann·18h

Google updated it‘s vibe design platform Stitch. Now with Agents that help you with everything. Gotta say, the video looks pretty cool.

Google Labs@GoogleLabs

Introducing the new @stitchbygoogle, Google’s vibe design platform that transforms natural language into high-fidelity designs in one seamless flow. 🎨Create with a smarter design agent: Describe a new business concept or app vision and see it take shape on an AI-native canvas. ⚡️ Iterate quickly: Stitch screens together into interactive prototypes and manage your brand with a portable design system. 🎤 Collaborate with voice: Use hands-free voice interactions to update layouts and explore new variations in real-time. Try it now (Age 18+ only. Currently available in English and in countries where Gemini is supported.) → stitch.withgoogle.com

English

Till Felippi@t1llmann·18h

Apparently @cursor_ai will release a potentially SOTA coding model, competing with Opus and Codex

Mark Kretschmann@mark_k

Tomorrow, @cursor_ai is going to release a super strong coding model. Possibly SOTA. Stay tuned. 🫡

English

Till Felippi@t1llmann·18h

@mark_k @cursor_ai Back when composer 1 came out i was actually quite impressed with the quality vs. speed it offered. It was great for live coding small tasks. But then times changed and I mostly stopped working in an IDE. I‘m excited for this one. Though, the competition is really good

English

110

Mark Kretschmann@mark_k·1d

Tomorrow, @cursor_ai is going to release a super strong coding model. Possibly SOTA. Stay tuned. 🫡

Krista Letz@kristaletz

exciting things at cursor coming soon 🎻

English

1.2K

280K

Till Felippi@t1llmann·1d

@karpathy Very cool gift. Enjoy and do good with it 🙏

English

139

Andrej Karpathy@karpathy·1d

Thank you Jensen and NVIDIA! She’s a real beauty! I was told I’d be getting a secret gift, with a hint that it requires 20 amps. (So I knew it had to be good). She’ll make for a beautiful, spacious home for my Dobby the House Elf claw, among lots of other tinkering, thank you!!

NVIDIA AI Developer@NVIDIAAIDev

🙌 Andrej Karpathy’s lab has received the first DGX Station GB300 -- a Dell Pro Max with GB300. 💚 We can't wait to see what you’ll create @karpathy! 🔗 #dgx-station" target="_blank" rel="nofollow noopener">blogs.nvidia.com/blog/gtc-2026-… @DellTech

English

488

772

17.6K

866.1K

Till Felippi@t1llmann·1d

@minchoi 😬

QME

487

Min Choi@minchoi·1d

DLSS 5 is wild... AI maxed... Demo in comment

AK@_akhaliq

DLSS-5 anything for free app: huggingface.co/spaces/victor/…

English

408

160K

Till Felippi@t1llmann·1d

@Salmaaboukarr The results are completely different than for other tasks; especially editing

English

Till Felippi@t1llmann·1d

@Salmaaboukarr We have an interesting challenge in our image model comparison which kind of shows exactly that effect: lumenfall.ai/arena/text-to-…

English

Salma@Salmaaboukarr·1d

Midjourney, Seedream, Z image are aesthetic models. Nanobanana is an editing model rather than an aesthetic model! hope that helps! :)

English

236

16.2K

Till Felippi@t1llmann·1d

Breaking: Stripe launches open payments protocol for agents

Stripe@stripe

x.com/i/article/2034…

English

Till Felippi@t1llmann·1d

Minimax M2.7 is out! And it looks very interesting

Chubby♨️@kimmonismus

Minimax M2.7 released! And its a big one Highlights: Self-evolving - first model that helped build itself, running 100+ autonomous optimization loops during its own RL training (30% internal improvement). Strong coder - 56.2% on SWE-Pro (near Opus 4.6), 55.6% on VIBE-Pro, production debugging down to under 3 minutes. ML research agent - 66.6% medal rate on MLE Bench Lite, tying Gemini 3.1. Office work - top open-source ELO on GDPval-AA (1495), 97% skill adherence, can do end-to-end analyst workflows (reports, models, PPTs). Native multi-agent and a new open-source interactive character demo called OpenRoom.

English