@janusch_patas Ditched postshot entirely. Ridiculous subscription and innane res limits, they don’t push the lever. Lichtfeld does and often
English
Zwecharki of Immense AGI fundamentalism
5.6K posts
Zwecharki of Immense AGI fundamentalism
@DefaultUserName
militant AGI fanatic • decentralist & hierarchist • Mormon • anti-hedonist
Wyoming, USA Katılım Mart 2009
390 Takip Edilen404 Takipçiler
@bilawalsidhu Do you guys not turn on mip filtering?
English
3d gaussian splats suddenly looking way less splatty! high frequency details look so much better with MRNF + PPISP
Tavius Koktavy 🌐💿🔆@TaviusKoktavy
In @lichtfeldstudio the new MRNF strategy plus PPISP is great. Compare MCMC (left) to MRNF (right) on a quick 20 second scan. Full-screen recommended Fine details pop, backgrounds are much more clear, and the color space is true-to-life.
English
@nishinsu_101 Any info on MRNF technique?
English
@DuckbillStudio Its very hard to persuade me that this isnt the future of gaming and immersive VR
English
「風で揺れる」3DGSの樹木。
リアルタイムに視点変更できるインタラクティブ表現。
Splatの分布密度から樹形を解析し、自動リギングでボーンを生成。さらに、各ボーン周辺のSplatを自動スキニング、風向・風速に応じた揺れを再現。
WebGPU対応により、性能もかなり出せるようになってきました。
ダックビル＠STUDIO DUCKBILL LLC@DuckbillStudio
せきぐちあいみさん の3D Gaussian Splattingの龍に、奈良井宿の3DGS街道を飛んでもらいました！ 開発中の3DGS自動リギングで、スプラットデータを解析→ボーン自動生成→アニメーション、すべてブラウザ内で処理しています。
日本語
@signalwatchHQ @kimmonismus That's the story of AI in the last 3 years. Everyone does this and everyone blames each other
English
@kimmonismus apple might be one of the smartest distillation stories in ai. if they can use gemini’s reasoning to sharpen smaller models that run on iphone, they don’t need to win the model race outright. they just need to package intelligence better than everyone else.
English
Apple's deal with Google goes way deeper than anyone thought.
Apple doesn't just get to fine-tune Gemini, they have full access to the model inside their own data centers.
That means they can distill (and are doing so) Gemini's knowledge into smaller models purpose-built for specific tasks, some small enough to run directly on your iPhone.
Apple can access Gemini's internal reasoning process, not just its outputs. That lets their smaller models learn how Gemini thinks, not just what it says. The result is compact models that punch way above their weight class.
English
@kimmonismus Is there any indication Apple isn't doing this already with their new foundational model?
English
@LocallyAIApp There doesn’t seem to be a way to select Apple foundational model on ios
English
Rediscover Apple Foundation with the updated model on iOS and iPadOS 26.4.
The new model gives better answers and improves instructions-following.
Update your device and app now.
English
@IHateZuckSoMuch don't you have access to program in whatever you want?
English
when the tipping point is optimal, to intensely spread certain information that would high confidence result in the expected outcome.
I bet this will be used often by both opposing sides, whichever they may be
English
Bottlenecked information, the right moment, DDOS false news spreading into an uncontrollable (this time good) frenzy
I wonder if, using AI surveillance of global sentiment, you could calculate the right nudge at the right time...
English
@thehavenly My favorite thing is when you don't spam peoples homes with your "offers" all the time.
English
@IHateZuckSoMuch Truth nuke? what? you're just skyrim gen!
English
@IndieGameJoe Awesome but where do you get those "friends" things?
English
These devs are building a survival game where you and 7 friends have to escape a massive tornado. This looks insane!🚐💨
- Survive the storm
- Scavenge parts to fix your van
- Large-scale destruction
Would you play this? It’s called Funnel Runners.
English
@andrewyguo Exactly what i need. Sick of splats getting ruined in dark areas
English
📢📢📢Dark3R (#CVPR2026) recovers camera poses and point clouds—and enables novel view synthesis—from images captured in ultra-low light (image SNR < −4 dB).
Project Page: andrewguo.com/pub/dark3r
Code/Dataset: github.com/andrewyguo/dar…
Paper: arxiv.org/abs/2603.05330
English
@yukky_saito Was it more precise SFM than what others or COLMAP would provide? or Realityscan?
English
これ試してみたけどめちゃすごかった
Junyi Zhang@junyi42
𝗢𝗻𝗲 𝗺𝗲𝗺𝗼𝗿𝘆 𝗰𝗮𝗻’𝘁 𝗿𝘂𝗹𝗲 𝘁𝗵𝗲𝗺 𝗮𝗹𝗹. We present 𝗟𝗼𝗚𝗲𝗥, a new 𝗵𝘆𝗯𝗿𝗶𝗱 𝗺𝗲𝗺𝗼𝗿𝘆 architecture for long-context geometric reconstruction. LoGeR enables stable reconstruction over up to 𝟭𝟬𝗸 𝗳𝗿𝗮𝗺𝗲𝘀 / 𝗸𝗶𝗹𝗼𝗺𝗲𝘁𝗲𝗿 𝘀𝗰𝗮𝗹𝗲, with 𝗹𝗶𝗻𝗲𝗮𝗿-𝘁𝗶𝗺𝗲 𝘀𝗰𝗮𝗹𝗶𝗻𝗴 in sequence length, 𝗳𝘂𝗹𝗹𝘆 𝗳𝗲𝗲𝗱𝗳𝗼𝗿𝘄𝗮𝗿𝗱 inference, and 𝗻𝗼 𝗽𝗼𝘀𝘁-𝗼𝗽𝘁𝗶𝗺𝗶𝘇𝗮𝘁𝗶𝗼𝗻. Yet it matches or surpasses strong optimization-based pipelines. (1/5) @GoogleDeepMind @Berkeley_AI
日本語
@migero_usausagi results? I'm super excited to try it too
English
@curiosityonx Wouldn’t call it “learned” more like “leveraged the archetypic core memory that we all share to play doom”
English
@bilawalsidhu @victor_UWer Overcapture is fixable. Undercapture isn't
English
@victor_UWer i crank the shutter speed up and use burst mode -- tend to overcapture -- old habit from my lightfield days. this was 700 ish photos
English
Look at how crispy this 3d gaussian splat came out! The reflections in the painting have me floored.
Captured with a Sony a7iii camera and 18mm lens. I posed the imagery using epic's reality capture and trained it with lichtfeld studio.
Lichtfeld is my new favorite way to train splats -- it integrates all the latest research and it’s completely free.
Also what I love about 3d capture -- you can process old datasets with new algorithms and breathe new life into them.
English
@bilawalsidhu Whoever solves that white wall clouding is MY HERO!
English