I’m trying out two different styles for object descriptions.... 🔹 Left: black box, Disco Elysium style 🔹 Right: clean text, line animation Which one feels better for the game’s atmosphere? 🔍show more

Working with animation styles in Runway I love how... it generated two completely different shots off the same prompt. On the left, we get a nice zoom out and a change in focus. On the right, it expands the world in front of the character. H/t Curious Refuge for the prompt: animation shot, cute girl with amazed expression, looking around in a magical candy land.show more

Here’s a video of a McDonald's pneumatic ketchup/mustard dispenser... that never made it to restaurants (post below). For decades we had two separate dispensers: one for ketchup and one for mustard. We’re looking to make our kitchens faster, one second at a time. They’re pretty efficient, but you had to pick up two different dispensers. Now there’s a single dual-purpose version. It dispenses both at once or either one, depending on how you move the black handle (left for mustard, center for both, right for ketchup). Yes, it’s probably over-engineered. Yes, it works. Yes, it saves time.show more

Save this for the next time your typography feels... weird and you can’t explain why. Nine times out of ten, it’s line-height - the vertical space between lines of text that decides whether type feels confident or clumsy, readable or tiring. Headlines Compress them. ~1.05–1.25 is usually right. Headlines aren’t meant to be read carefully line by line. They’re meant to land as a shape. Tighter leading pulls words into a single visual block and makes the headline feel confident and intentional. One hard rule: ascenders and descenders should never touch. If your t’s and l’s are colliding with g’s and y’s, you’ve gone too far. Body text Do the opposite. Give it air. ~1.4–1.7, depending on the typeface, its x-height, and the line length. Body text is about endurance. Your reader shouldn’t notice the typography at all. If they’re aware of the type, something’s off. The goal is that they just keep going, paragraph after paragraph, without friction. In Framer, you can find the Line-height settings in the Text Panel > Line. And yes, this applies even if you’re using the same font for headlines and body text! Same font, same page, but they have completely different jobs - and different line height. Sometimes, choosing a good type isn’t enough - you need to know how much space to give letters so they can do their work.show more

004/100 Buttons. A bit of the process on building... an animation. When looking at a finished animation or in this example a finished button, it can look quite complex inside the CSS. But when building it, it’s more like a lot of simple steps, one after another. Here I had the idea to make some kind of text animation like the footer logo on the Osmo site. I try to add the base animation with no complex easing, for example transition: translate 0.4s ease. Starting with just moving the one text from bottom to top and the other text to top. Adding a stagger, play around with it. Searching for a way to make it more circular. On the research I found the sin() function inside CSS which can build a more smooth non linear curve for the stagger which creates this circular effect. And step by step adding more complexity like, different easing for hover/hover-out, opacity, 3D transform and more. I use also the sin() function to rotate the letters, so the middle ones are getting more rotated than the outer ones. Another thing which helps is to add a small delay on hover, for example 0.05s or 0.1s, you don’t really see the difference, but when you hover pretty fast on and out it doesn’t get that jumpy. I’m using here GSAP’s SplitText to split every char into spans. And then I’m adding a CSS index variable to every span, starting from the center. SplitText can provide CSS index variables, but you cannot tell it from which direction. For the sin() it’s also important to have a max length, so I add another CSS variable with the max char number on it. Crafting 100 Buttons with Osmo ⏳ Total time: 63hshow more

I’m asking for all the prayers, love, and support... you can send my way right now ❤️ About eight days ago, my husband left me. After years of battling abuse, I finally found the courage to kick him out and not take him back. Now I’m trying to pick up the pieces of my life, stand on my own two feet, and figure out what’s next. I’m left dealing with severe health issues that have landed me in the hospital from dehydration and neglect since he walked away. I have no help, I don’t know my legal rights (like whether I can change the locks), and I’m trying to find the right attorney to move forward. I feel completely helpless, but I still have a roof over my head — thank God for that. If it wasn’t for Him, I wouldn’t even be here writing this. Please keep me in your prayers as I slowly try to stand again. I love you all and just need some time and a whole lot of prayer support right now. Thank you from the bottom of my heart. 💜show more

gemini omniflash is actually f*cking cracked. you can animate/edit... any video with a text prompt. character swaps, object transforms, full environment changes without regenerating/rotoscoping. everyone using AI to to animate and edit videos right now hits the same wall. the clip comes out 90% right and you regenerate from scratch hoping the 10% fixes itself. it never does. the fix is using your video as the input. omniflash edits what's already there instead of rolling the dice again. here's what's in the system: > the two-layer premiere trick: generate the same shot twice (one with background removed), stack them, cut at one frame, instant scene change > character swap with a single reference image (plus the one line you need or the model keeps the original's features) > object transforms that leave the rest of the frame untouched: stone into glowing sphere, candles into flowers > style transfer from an image reference instead of text, way more accurate > why stacking edits in one prompt breaks everything and the exact step order that doesn't > the audio limitation nobody mentions and how to work around it i packaged every prompt, the edit sequence, and the premiere layering setup. RT + reply "OMNI" and i'll send it over.show more

RUMOR HAS IT👀: An interesting theory on why Harry... could've been at Louis' show in Berlin: (©️SchuSchu8437) •There were VIP boxes all around the arena. In the break between Pale Waves and Louis most of them were illuminated. All in the same light EXPECT for one. Which had a green light. •The VIP box right next to the green light was blacked out, by a curtain? Look at the box to the right of it. Every other dark box was very obviously just without light. • Louis kept looking in that direction, at least it looked like that. •Afterwards ,outside of the arena where there was the tour buses enter and the trucks unload. Pretty shortly fans are sent away from that side. Security said they couldn’t stand there anymore and got shooed away. Just when everyone left and no one seemed to pay attention, a completely blacked out VIP Van left the backstage area of the arena and drove by . People still kept coming out of the arena and went to that spot again. They were not actively send away again after the Van left. #louistomlinson #HDWGHBerlinshow more

No tweets last two days cuz I was doing... my day job with my coach/boss Sean Foley. First swing is what I’m trying to feel. Really get to my right side, stay there during the transition and then go thru. Reason for this is I have a tendency to pull really hard with my arms and shift too early in the transition causing the club to get too far out. Then to get it back inside, I side bend late which causes too much flipping.show more

$KNDX 🤖 Theres 3 big narratives that are sending... coins left right and centre rn. 🚀 #AI, #Gamefi, & #NFTs 🔹Theres 50% mindshare for #AI. 🤖 🔹#GameFi mcap is hitting ATH's with #OfftheGrid, $XBG and $SUPER making spectacular moves. 🎮 🔹NFTs and the #Metaverse are making a strong comeback with $APE up 100% over the weekend. 🐵 What if there's a project that touches all these trending narratives with groundbreaking technology to disrupt all 3 of them? 🔥 💡- That's where $KNDX comes in. -💡 Kondux is a cutting-edge Web3 SaaS platform, combining NVIDIA’s Omniverse, AI, Blockchain, and dynamic NFTs to revolutionize secure asset management across industries. 👏 Their flagship product, kNFTs, are 3D digital assets usable across Metaverse and Gaming platforms, AR/VR/XR environments, and manufacturing applications. Kondux’s scalable model opens new revenue streams by enabling effective digital asset monetization. 💰 Kondux is the first Web3 project to integrate VFX pipelines with NVIDIA’s Omniverse and bringing it onto the Blockchain. ⛓️ It is also the only Web3 project with a *Select Status Partnership* with NVIDIA, operating under NVIDIA NDAs and working with them directly for more than 2 years. About their NVIDIA Integrations: 🤖 🔹There are three areas of the Kondux tech stack that coincide with three divisions of NVIDIA: 📡GDN (Graphics Delivery Network, the backbone of GeForce Now) 💡Omniverse for 3D aspects such as, geospatial data, real world physics, lighting, and raytracing 🤖NVIDIA AI Foundation, which covers many aspects of #AI, including inference and deployment scaling. The convergence of all these components lie within .USD file format . 🔹 They are the first blockchain project to integrate NVIDIA’s Omniverse Cloud and Graphics Delivery Network (GDN) to provide high-quality 3D content accessible on any device without requiring high-end hardware. 🔹 This setup streamlines content management, democratises access to resource-intensive 3D content, and enables real-time interaction with 3D NFTs. Now, I haven’t seen any crypto project so deeply connected with NVIDIA and NVIDIA technology. GDN is a HUGE competitive advantage. With it, the need for #GPU’s basically goes out the window. 🤯 Now lets take a look at some of the other main features... 👀 OpenUSD (Universal Scene Description): 📽️ 🔹 Kondux is leveraging USD technology, developed by Pixar and used by Meta, Apple, Microsoft and other industry leaders to enhance 3D graphics and interoperability within its creative ecosystem. 🔹 Originally created for high-end film production, USD now supports a variety of applications, including gaming and virtual reality, making it a key asset for Kondux. kNFT's: 🎨 🔹 Kondux is pioneering a new category of NFTs known as kNFTs, which aim to redefine NFT utility through innovative features. 🔹 A standout feature is the upgradeable aspect provided by Kondux DNA, allowing kNFTs to transform and combine with other NFTs, creating limitless possibilities in art, gaming, and music. 🔹Through the Kondux AI portal it will be possible to communicate with kNFTs. They can learn and adapt. This AI technology is revolutionary because it makes human to kNFT interaction possible, turning it into a unique, personalized experience. Check out the clip of kNFTs in Unreal Engine 5 gameplay below. 👇 Kondux is a very obvious utility play with huge upside because it’s multi narrative. 📈 It's seriously groundbreaking stuff that they’re about to launch. 🚀 After speaking with the team there’s no doubt in my mind this will do crazy big numbers in the next months. 🤑show more

When strangers become big sisters in seconds. GPT Image... 2 + Seedance 2.0 on Renoise prompt Create a clean, professional Pixar-style 3D character sheet of a 6-year-old girl, highly detailed, vibrant colors, polished family-friendly animation style. Character Description: Adorable 6-year-old girl with soft dark wavy hair tied in a cute high ponytail with a small pink bow. Big sparkling brown eyes, round expressive face, rosy cheeks, small button nose, and a warm friendly smile. She has a healthy, energetic child body type suitable for a 6-year-old. Outfit: Bright yellow t-shirt, light pink shorts, white ankle socks, and colorful small sneakers (blue and pink accents). Character Sheet Layout: A single clean vertical character sheet with multiple views of the same girl arranged neatly: Top Left: Full-body front view, standing naturally with hands on hips, smiling confidently. Top Right: Full-body 3/4 angle view, showing depth and personality. Middle Left: Full-body side profile view (facing right). Middle Right: Full-body back view. Bottom Center: Large close-up of her face showing 4 different expressions in small circles — Happy smile, Surprised, Shy/embarrassed, and Excited/grinning. Bottom Right: Small extra details — close-up of her ponytail with bow, shoe details, and hand poses. Style & Quality: Beautiful 3D Pixar animation style, smooth rounded shapes, expressive eyes, soft realistic textures, vibrant yet soft lighting, clean white background with subtle light shadows, highly polished, professional character design sheet, perfect proportions, studio quality, sharp details, warm and wholesome feel. 16:9 aspect ratio, ultra-detailed, cinematic lighting. #RenoiseCanvasshow more

I was arrested and released on bail pending investigation... for defending my family. A man was taking pictures of my daughter and trying to enter my restaurant. Any parent would react. Things escalated and blows were exchanged, yet I’m the one facing charges based on the culprit’s complaint. What message does this send? That protecting your own family makes you the criminal? I respect the law, but the system must also respect a father’s right to act when his family feels threatened. Metropolitan Police, I hope the full truth comes out. Because this isn’t justice.show more

Smooth Commute Ahead on Hebbal–Nagawara Stretch 🚶‍♂️...🛵... 🚗✨ A... major clean-up and encroachment removal drive along the Hebbal–Nagawara–Allalasandra Bridge (Outer Ring Road) corridor is set to improve traffic flow, pedestrian movement and overall road safety. Over the past week, a large-scale sanitation and debris-clearing operation was carried out across this busy stretch, involving multiple civic and infrastructure agencies, along with local businesses, traders and community groups. 🔹 What was done? • Removal of construction debris and roadside waste • Clearing of encroached footpaths to restore pedestrian access • Beautification of pedestrian pathways • Elimination of long-standing garbage black spots 🔹 Key impact: • Around 1,200 tonnes of waste and debris cleared • Over 150 sanitation workers deployed daily • Use of JCBs, tractors, tippers and compactors • ~40 illegal shop extensions removed, freeing footpaths for public use 🔹 Local-level results: In Dasarahalli ward alone, 150 tonnes of waste were cleared and 80 metres of footpath was upgraded with the support of civic staff and volunteers. 🌱 Why this matters: Cleaner roads, safer footpaths and smoother traffic not only improve daily commutes but also encourage better civic habits, environmental responsibility and respect for public spaces. 📍 Location: Hebbal – Nagawara – Allalasandra Bridge (ORR) Bengaluru North City Corporationshow more

Stop guessing how much to spend. Use these two... new tools in Google Analytics to see exactly where your budget should go. 📉 𝗦𝗰𝗲𝗻𝗮𝗿𝗶𝗼 𝗣𝗹𝗮𝗻𝗻𝗲𝗿: 𝗠𝗮𝗽 𝗼𝘂𝘁 𝘆𝗼𝘂𝗿 𝗺𝗲𝗱𝗶𝗮 𝗽𝗹𝗮𝗻. Test different spend levels to see which one actually hits your revenue goals. 🧮 See how much revenue to expect 𝙗𝙚𝙛𝙤𝙧𝙚 you commit. 📊 Compare different spend plans side-by-side. 📱 Find the best way to allocate your budget between platforms. 𝗣𝗿𝗼𝗷𝗲𝗰𝘁𝗶𝗼𝗻𝘀: 𝗙𝗶𝘅 𝘆𝗼𝘂𝗿 𝘀𝗽𝗲𝗻𝗱 𝘄𝗵𝗶𝗹𝗲 𝗰𝗮𝗺𝗽𝗮𝗶𝗴𝗻𝘀 𝗮𝗿𝗲 𝗹𝗶𝘃𝗲. Stop waiting for end-of-month reports to find out where you overspent. 📈 See where your revenue is pacing in real-time. 🔍 Spot failing campaigns before they waste your budget. 💡 Shift money to top-performing channels while they’re still active. Build a better budget →show more

When the C219 CLS launched in 2004, it invented... a body style that every other manufacturer spent the next decade trying to copy. Four doors, a fastback roofline, no B-pillar, and frameless windows, none of which existed together in one car before this. The CLS 500 came with the 5.0L M113 V8, 306 HP and 460 Nm, fitted with two spark plugs per cylinder for a more even burn across the combustion cycle. It was built on the W211 E-Class platform but with completely different proportions and a lower, more aggressive seating position with 4 seats, not 5. That made it feel like a different car entirely. Most buyers chose the diesel because it was cheaper to run. The ones who chose the V8 got the engine the car was always meant to have.show more

We can't quite believe it, but we've never posted... an FX Animation Breakdown before! There's been such amazing FX work done on so many of our projects, we're definitely going to show you more of how it's done. This film in particular was so FX heavy, we had to highlight the crazy work that went into it. You all know we love an action sequence. But for this one, we took a pretty different approach than we usually do! The entire thing was rendered through FX animation. The animator did their rough and tie-downs but instead of going into clean-up and BG paint, the FX team rendered the characters as well as animated the background FX (which is a super rare thing to do!). The result is this abstract representation of two magnificent powers colliding in an epic battle!!! Client: Digital Extremes Character Layout: Colin Tan 2D Animation: Shane Dering 2D Animation Lighting and Render: Simon Duclos FX Animation: Christophe Madeleine, Louve Karnas Compositing: Maxime Temple #FXAnimationBreakdownshow more

🚀 Introducing EgoExo Forge - built on top of... Rerun, Gradio, and Hugging Face hub (I’ll be in San Francisco July 21–29 — if you’re into robotics, egocentric AI, large-scale data collection, or just want to chat, DM me!) In my opinion, large-scale, diverse, and high-quality data is still the largest bottleneck for generalized robotics deployment. I believe that some version of imitation learning from human examples will be the most scalable + clean way to train humanoid robots 🤖 (similar to what Tesla did for Full Self Driving). Teleop is too expensive to collect a large enough dataset in a reasonable manner, so passive collection via egocentric (and in certain cases, exocentric) views feels like the right bet. Over the past few months, I've been trying to build out the scaffolding for this and using Rerun as my underlying infrastructure. Data being collected needs to be easily inspectable + time series and rerun provides the right tooling for this. My goal is to first build out a ground truth representative dataset from already existing open source data, generate some reasonable baselines, and then go out and collect my own data that adheres to the defined schema. 🔍 Starting with open-source datasets 1. EgoDex from Apple 2. HOCap from Nvidia and the University of Texas at Dallas 3. Assembly101 from Meta All these different datasets have different sensor configurations + annotations, so my goal with egoexo-forge is to have one consistent labeling scheme + data layout. I built a data pipeline that aligns all of the different datasets in one general schema assuming the COCO133 keypoint layout that allows for exo+ego, ego only, or exo only Since the scaffolding is already there, it becomes MUCH easier to add other datasets. So the next ones that I'll be including are HD-EPIC kitchens dataset, HOT3D, and finally my own personal iPhone + insta360 go collection method. Once I have a diverse variety of datasets, I'll double down on what I believe to be the key algorithms required to make useful data for imitation learning 📊 1. Camera Pose estimation via SLAM/SFM for ego perspective (and automatic calibration for exo) 2. Human pose estimation for both egocentric + exocentric views 3. Metric 3D reconstruction + object tracking I'll be setting up reasonable open-source baselines for each of these to validate that these datasets work, and then finally try to use the generated datasets for some imitation learning via the pi0-lerobot repo I've been working on. I plan on making a blog post + providing more info on all of this in the near future so stay tunedshow more

Household Chores Motion-Based Reference Prompt for ChatGPT Image 2.0... / Seedance 2.0 on Yapper Prompt: Create a monochrome grayscale 4×4 instructional storyboard showing a full household chore sequence. Use a clean white background, soft studio lighting, and high contrast to emphasize posture, movement, and object interaction. The character must match the provided reference image in face, skin tone, proportions, and overall likeness, with natural makeup, a soft expression, and consistent identity across all 16 panels. The character should wear a modern modest outfit: a fitted crop top with a clean neckline, high-waisted straight or slightly wide-leg jeans, and optional minimal sneakers or barefoot indoor styling. Keep fabric movement natural with subtle folds and tension. Each panel must be the same size, separated by thin black lines, and clearly numbered 1 to 16. Show a full-body pose in each frame performing a different chore in a minimal environment with only essential props. No clutter, no complex background, no extra characters. Include in every panel: top-left step number and task title, center full-body action pose, bottom-left 3–4 short instruction lines, and motion arrows or guides showing movement flow. Use this chore sequence: Make the Bed, Tidy the Room, Dust Surfaces, Vacuum Floor, Sweep Floor, Mop Floor, Do Laundry, Hang Clothes, Fold Clothes, Clean Kitchen Counter, Wash Dishes, Take Out Trash, Water Plants, Clean Bathroom, Organize Shelves, Final Room Reset. Use motion indicators appropriately: curved arrows for wiping and folding, straight arrows for movement, and circular arrows for scrubbing and mopping. Style should be highly detailed 3D, smooth grayscale shading, soft shadows, clean linework, and polished concept-art quality. No color, no revealing clothing, no extra background detail, only the subject, props, and instructional elements.show more

There’s been two papers released in the past couple... months, one by Google and one by NVIDIA, that argue that ordering the documents retrieved by RAG systems can enhance performance. However, they both give two different strategies on HOW these documents should be ordered 🤔 Both papers agree on two main points: 1️⃣ There’s a fundamental issue in RAG - as more documents are retrieved, more irrelevant context (e.g., hard negatives) are introduced, which leads to confusion for the LLM and eventually degrades the quality of the generated output. This is called an inverted-U performance curve. 2️⃣ Ordering the retrieved documents is a key lever for optimizing RAG performance. Google Cloud researchers proposed ordering results based on relevance scores: The authors in this paper argue for relevance-based reordering, or ordering the retrieved chunks based on their similarity scores, so the most relevant documents are at the beginning and the end of the inputs to counter the “lost in the middle” effect. NVIDIA researchers proposed ordering results based on the original sequence of document chunks: The authors of this paper argue for Order-Preserving Reordering, or Order-Preserve RAG (OP-RAG), to maintain the logically coherent content flow of the document. So they preserved the original order of retrieved document chunks in the source text, instead of ranking them by relevance scores. So which one is right? It probably depends on the specific use case and dataset - relevance-based reordering could perform better in tasks where you need fast access to the most critical information (e.g., fact retrieval, QA systems), while order-preserving RAG might be better where you need to understand the sequential structure of information (e.g., narrative or legal documents). There are still so many uncertainties in AI - we don’t actually know what we’re doing, and it takes awhile to figure out the best strategies for most things! Excited to see more research about this.show more

When I was 16 working in a hospital cafeteria,... a 38-yr old a coworker named “Mike 7-4” slapped the shit outta me because I told my supervisor he was out back doing cocaine instead of helping me in the dish room. This is where politicians need to be working. A Sunday morning ain’t shit in most food service jobs. Work fast food on a Friday night after HS football. Or Deliver pizzas, Put Trump a man down in a two-man dish room, with the front line beggin for clean plates while Mike 7-4 out back trying to buy Vicodin off the nurses in the cardiac wing. Dish room a sweat box. 300° steam coming off the machine and here come with Mr. Wallaby cussin because the silverware aint rolled yet. Bitch It ain’t rolled because my coworker is high sir . Whichever candidate Survive a shift with Mike 7-4 at the Healthsouth Rehab Hospital in Birmingham AND can find out why he added “7-4” to his name, Black male vote locked in I guarantee.show more

Short clip but the footballing IQ from everyone involved... top: Burgess; one-two with Yalcouyé, moving back into space behind him—baiting the press. Brilliant touch to take Campbell out of the game. Plays the ball to Yalcouyé with a good weighted pass. Yalcouyé; Quickly moves into space & demands the ball, good touch into a very small space. Excellent composure on the ball to find Galbraith. Galbraith; Follows the ball to entice his marker. Sees Hamer (marker) press Yalcouyé, little step backwards into space, slightly scans ahead to see how much time and space he has got. Drives with the ball, plays a great pass to Franco, in the half-space who's in between the lines. Franco; In between Sheffield's defensive line + midfield line—Good positioning. Before receiving the ball, glances over his left shoulder (X2) to see where Ronald is. Sees Sheffield's LW move over to the left, so steps to the right to make the pass easier for Galbraith. On the half-turn, sets it with his left foot, *punches* the ball into right into Ronald's path. Very fine details but the movement to the execution is perfect. #Swansshow more