Someone's building a camera free vitals tracker on the... same $8 chip philosophy ESP32 + M5Stack CoreS3, reading WiFi signal reflections (CSI) to pull heart rate, respiration, and full 3D skeletal pose, no camera involved. Same core idea as the swarm below cheap, distributed hardware doing what used to require dedicated expensive sensors.show more

APPLE JUST GOT HUMILIATED BY AN $8 JACKET BADGE... A Japanese maker looked at Vision Pro's $3,499 Persona feature and thought, "I can build that for the price of a pizza." So he did. Same real-time face mirroring. Same natural expressions and blinks. Same instant reaction when you smile. Apple threw an entire headset at this problem. M2 chip. R1 processor. 12 cameras. LiDAR. He threw one ESP32, a tiny camera, and MediaPipe tracking 468 facial landmarks, running entirely on-device. Total cost: eight dollars. This is what happens when a weekend hacker outbuilds a trillion-dollar company's flagship feature with parts from a hobby drawer. The hardware moat didn't shrink. It vanished. Full build in the video below. 👇show more

This is absolutely insane. There's a new ultra viral... Github repo that uses Wifi to spy on people with extreme accuracy • every room you've ever been in is already filled with radio waves bouncing off your body • when you move, breathe, or even just sit still, those waves scatter differently • ruvnet built the thing that reads the scatter • it maps your full body into 17 keypoints, through walls, in total darkness, no camera required • measures your breathing and heart rate while you sleep without touching you • uses your neighbors' routers as extra radar. they have no idea • $9 chip. 55kb ai. no cloud. no cameras. no consent required. Github repo:show more

Korean outfit style board GPT Image 2.5 and Seedance... 2.5 on Renoise AI 日本公式 Prompt: A clean, premium 4-second fashion lookbook animation starting from the first frame. The entire layout stays exactly as shown: same background, same cream item panel, same captions, same crown icons, same wordmark, same framing. The model on the right turns a full 360° on the spot the way a real fashion model does on set: she steps her feet around naturally, shifting her weight from one leg to the other as she turns, her arms moving freely and her shoulders relaxed. As she comes around to the back she glances over her shoulder, and as she returns to face the camera she settles into a new, different confident model pose. Natural posture, realistic cloth physics; hair, hems and accessories swing and settle with the motion. She stays in the same spot and at the same size in frame throughout. At the same time, every product cutout inside the item panel rotates in place on its own vertical axis, like a luxury product display render, revealing front, side and back. Each item stays centered in its own slot, keeps its size, and completes exactly one full rotation in sync with the model's turn, ending in its original orientation. The panel, divider lines, captions, crown icons and wordmark stay perfectly fixed; only the model moves and the products rotate. Camera locked off, no zoom, no pan. Soft studio lighting with realistic shadows under the model and the products. Smooth, natural motion, clean commercial fashion-ad aesthetic, ultra-sharp details, luxury catalog presentation.show more

Wow! Send A File By Just The Camera! A... vibe coder used AI to build a file transfer system that sends data between two phones using only a screen and a camera. One phone displays a new type of animated QR codes while the other scans them to rebuild the file, with no Wi-Fi, Bluetooth, or cables needed. It is fully optical and local. The system uses fountain codes that create each QR frame as a random mix of file data. This keeps transfers working even if some frames are missed, reaching speeds of about 129 KB/s for a 2 MB image. The entire project was built in one night and released as open source. The idea came from a music project where the developer wanted to share MP3 files without streaming or using the same network. Animated QR codes became the solution, showing a creative new way to transfer files with everyday phone hardware. GitHub link:show more

After Aroosa Khan now Iqrar Ul Hassan’s second wife... has entered the scene. I honestly don’t understand what their mission is or why they keep doing such things like they made the family Vlogging into a publicity and attention seeking instrument. Iqrar ul Hassan himself used to criticize others for this kind of behavior, and now his own wives seem to be doing the exact same things. As they say, tit for tat, that’s exactly what seems to be happening with them. With the camera constantly focused on body, it’s hard to understand what they’re actually trying to achieve. Are they genuinely trying to make a point, or is this all just for creating content that will gain them more attention?show more

Robots can now reconstruct 3D scenes in real time... from a single RGB camera. [📍 Projects page + paper] No depth sensor. No retraining. 30 FPS. Researchers at the Imperial College London introduced KV-Tracker, a training-free method that makes heavy models like π³ and Depth Anything 3 fast enough for real-time tracking. The idea is simple. These models use global self-attention, which is powerful but computationally expensive. KV-Tracker caches the key and value pairs from selected keyframes and reuses them for new frames. That cache becomes an implicit scene representation. Result: • Up to 30 FPS • 10 to 15x speedup • Accurate 6-DoF tracking on benchmarks like TUM RGB-D and 7-Scenes • Works with monocular RGB only It also supports object-level tracking with masks and allows saving the KV-cache for later reuse. For robotics, this reduces hardware constraints and moves real-time 3D perception closer to practical deployment. Credit to Marwan Taher (Marwan Taher) at Imperial’s Dyson Robotics Lab and many others who contributed to this! 📍 Save projects page + paper for later: Video: ——- if it matters in AI or Robotics you'll read it here first:show more

SOMEONE BUILT A WAY TO SEND FILES BETWEEN TWO... PHONES WITH NO WIFI, NO BLUETOOTH AND NO NETWORK AT ALL it works with just a screen flashing qr codes at a camera one device plays an endless stream of animated qr codes, the other one points its camera at the screen and rebuilds the file out of the light no pairing, no app, no shared network, nothing but the camera > its pulling about 129 kb/s, a full 2mb image transfers in seconds > theres no back channel, the receiver can never ask for a missed frame, so it uses fountain codes, every frame is a scrambled mix of the file rather than one piece of it > that means the receiver just collects any frames it can catch in any order, and once it has enough it can reconstruct the whole file > dropping frames costs you a little time but never breaks the transfer > you can point the camera at a stream thats already running and it just locks on mid flight he was building a cached mp3 player because he was sick of ads and bad streaming quality. he wanted a way to send tracks phone to phone without both phones being on the same network, and realized flashing qr codes could carry the data he built the whole working prototype in one night with claude code its open source and already at 375 starsshow more

Seedance 2.0 on Higgsfield AI The quality and creative... control here are incredible. This feels much closer to real music video production than ever before. Full open-sourced prompts & assets below: @ Image1 is the lead performer — preserve her exact identity in every shot: cornrow braids, septum ring, statement earrings, sculptural glossy white wrapped top, dark indigo denim. @ Audio1 is the music track — entire edit cut to it; no invented music, no voiceover, no lip-sync. Music-video route: dance/performance, brutalist stage show. Track: industrial hip-hop, 95 BPM half-time, distorted 808s, luxury-noir. Director thesis: architecture as choreography — dancers staged on concrete plinths at different heights, a human sculpture garden that activates around the lead. Visual world: brutalist concrete atrium with massive raw concrete plinths and steps at staggered heights, shafts of cold daylight from a skylight, dust motes in the light shafts, dark shadow pockets. Background dancers: 6 dancers in oatmeal and bone-toned knitwear and tailoring, sculptural silhouettes, like museum figures. Palette: bone, oatmeal, concrete grey, deep indigo accent on the lead. No neon, no particles beyond natural dust. Cut map synced to @ Audio1: 0.0–2.0s THE BREATH — static ultra-wide symmetrical: six dancers frozen in sculptural poses on six plinths at different heights, the lead standing on the floor at center like a visitor. Total stillness, museum silence. 2.0s THE DROP — 808 kick: every statue snaps into a new pose simultaneously on the hit, heads turning to her; hard cut to a fast push-in on the lead. 2.0–4.0s she struts between the plinths, camera tracking laterally through foreground columns creating parallax wipes; on each kick one statue above her snaps a new angular pose. 4.0s snare — whip-pan: full freeze, she hits a profile vogue arm frame between two plinths, statues frozen mid-rotation above. 4.0–8.0s sculpture canon: her sharp head isolations and New Way vogue lines on the snares trigger cascading pose-changes across the plinths, a mechanical wave of bodies jumping pose to pose like stop-motion statues; torso rolls on the 808 slides travel up the plinth heights from low to high. Camera slowly cranes upward through the atrium. 8.0s hi-hat triple roll — 3 rapid macro inserts: her earring in a light shaft, a dancer's hand rotating on a plinth edge, dust exploding off concrete as a boot lands. 8.5–12.0s escalation: accelerating orbital-crane move spiraling around the whole sculpture garden as all six dancers now move continuously in their own orbits on the plinths, her arm work at the center speeding up; speed-ramp stretch then snap back on the kick. 12.0s bass slide — the dancers step down off the plinths in unison and sink into deep lunges around her as she glides through a sideways level drop, camera dutch and low across the floor. 13.5–15.0s FINAL BUTTON — hard cut to the opening ultra-wide: she now stands ON the tallest central plinth in a frozen vogue pose one arm high, the six dancers frozen as statues on the floor below looking up; micro push-in, hold. Loopable. Performance rules: lead stoic, dominant; dancers precise, sculptural, expressionless. Continuity: same seven people, same wardrobe, same atrium. Audio intent: every cut and pose-snap on a hit of @ Audio1; faint foley only (steps, fabric, room tone). Quality bar: expensive editorial music video, museum-grade art direction, no AI gloss, no glow.show more

Made with seedance 2.5 Prompt: Create ONE continuous 12-second,... 9:16, ultra-photorealistic LIVE-ACTION luxury fashion film. IMPORTANT REFERENCE MAPPING: REFERENCE Image = PRIMARY CHARACTER. Use the FIRST uploaded reference image of the woman. REFERENCE image2 = ENVIRONMENT. Use the SECOND uploaded reference image of the square. REFERENCE David Carbon = HERO OBJECT. Use the THIRD uploaded reference image of the sunglasses. Do NOT refer to these images as Image, image2 or David Carbon. Use the reference images according to their upload order. CHARACTER — REFERENCE IMAGE 1 Preserve the exact woman from REFERENCE IMAGE 1 throughout the entire video: same facial identity, facial structure, natural eyes, hairstyle, body proportions, skin appearance, clothing, footwear, jewelry, accessories and overall appearance. Do not copy the original background from REFERENCE IMAGE 1. The protagonist is the ONLY main character. ENVIRONMENT — REFERENCE IMAGE 2 Use REFERENCE IMAGE 2 only for visual language: architecture, atmosphere, warm sunset lighting, materials, monumental scale and cobblestone character. Create a NEW believable cinematic pedestrian space inspired by it. Do NOT reproduce its exact composition, exact camera angle or exact street layout. Provide enough open pedestrian space for natural walking and the collision. Respect real-world spatial logic: pedestrians remain in pedestrian areas, architecture has realistic scale and perspective, no impossible geometry. HERO OBJECT — REFERENCE IMAGE 3 The sunglasses shown in REFERENCE IMAGE 3 are the ONE and ONLY HERO OBJECT. Preserve their recognizable: shape, proportions, oversized shield silhouette, dark continuous lens, thin metallic/gold frame, temples, hinge construction, surface details, colour and material. The SAME physical sunglasses must remain recognizable: when worn by the protagonist, when released, during bullet-time macro inspection, and after landing. Do not redesign, replace, duplicate or regenerate the sunglasses. EXACT OBJECT COUNT Exactly: ONE protagonist. ONE secondary passing person. ONE Labrador. ONE leash. ONE pair of sunglasses from REFERENCE IMAGE 3. ONE takeaway paper coffee cup. ONE plastic drinking lid. ONE silver-grey OPEN clamshell flip phone. No duplicate objects or people. PERSISTENT PHONE There is exactly ONE phone in the entire video. It is a compact early-2000s silver-grey OPEN clamshell flip phone. Upper half: small illuminated display and speaker. Lower half: physical numeric keypad, navigation buttons and microphone. Visible central mechanical hinge. LOCK its appearance throughout: same silver-grey colour, same dimensions, same thickness, same hinge, same display, same keypad, same buttons, same open angle. The same phone is seen in wide, medium and macro shots. It never transforms into another phone. PERSISTENT COFFEE Exactly ONE medium disposable PAPER coffee cup with ONE lightweight plastic drinking lid and dark coffee inside. Same cup, same lid, same coffee and same proportions throughout. Before collision the lid is properly attached. Coffee remains inside until a real physical opening appears. HAND OWNERSHIP Before collision: RIGHT HAND = ONE OPEN flip phone held to the RIGHT ear. LEFT HAND = ONE coffee cup. Sunglasses = naturally worn by the protagonist. After collision: RIGHT HAND = COMPLETELY EMPTY. LEFT HAND = COMPLETELY EMPTY. Nothing replaces the phone or cup. Nothing remains attached to either palm. 0.0–2.0s — CINEMATIC CHARACTER REVEAL Begin with an extreme low close-up of the protagonist’s black leather heeled boots walking naturally across the pedestrian space. Camera performs ONE elegant continuous movement: smooth crane upward + gentle backward tracking. Reveal progressively: boots → legs → ripped light-blue baggy jeans → black corset → white tank top → jewelry/accessories → sunglasses → face → nearly full body. The protagonist wears the SAME sunglasses from REFERENCE IMAGE 3. She holds the SAME OPEN silver-grey flip phone to her RIGHT ear and the SAME coffee cup in her LEFT hand. She is actively having a phone conversation. Show subtle realistic speech: natural lip movement, jaw movement and mouth articulation. Maintain exact identity from REFERENCE IMAGE 1. Premium live-action fashion cinematography. 2.0–3.7s — WALKING / SECONDARY PERSON Maintain smooth frontal tracking. The protagonist continues walking naturally through the pedestrian area. Exactly ONE secondary person approaches from the opposite direction. That person holds exactly ONE leash attached to exactly ONE Labrador. The Labrador walks slightly ahead and gently pulls the person forward. They remain background/supporting elements. Maintain: stable protagonist identity, stable hairstyle, stable clothing, stable sunglasses, same phone, same coffee. 3.7–4.4s — ACCIDENTAL COLLISION As they pass each other, the secondary person’s ELBOW accidentally clips the protagonist’s arm/upper torso. Small, accidental, believable contact. The protagonist naturally recoils backward. This SAME collision causes THREE physical release events. SUNGLASSES The sunglasses from REFERENCE IMAGE 3 receive ONE short realistic outward impulse. They separate naturally from the protagonist’s face. They travel only a short physically believable distance. They do NOT fly toward the camera. COFFEE The SAME coffee cup leaves the LEFT hand. It receives one short outward/downward rotational impulse. The cup tilts and rotates. Because of the rotation and liquid inertia, ONE EDGE of the lightweight plastic lid partially pops loose from the cup rim. The lid does not disappear and does not explode away. A real physical gap forms between lid and rim. ONLY THEN does dark coffee begin escaping through this opening. Initial liquid movement is restrained. PHONE The SAME OPEN clamshell phone leaves the RIGHT hand. It receives one short outward/upward impulse. It keeps the SAME open angle and physical design. The instant it separates from the fingers: RIGHT HAND = EMPTY. The instant the cup separates: LEFT HAND = EMPTY. FACE / HAIR CONTINUITY Preserve the exact face from REFERENCE IMAGE 1 during the entire collision. The natural human eyes already exist behind the sunglasses. When the sunglasses separate, reveal the SAME normal eyes: stable pupils, eyelids, gaze direction, brows, nose, cheekbones and facial symmetry. Do not regenerate or alter the face. Hair remains stable with only subtle realistic inertia. No sudden hairstyle change. 4.4–5.2s — BULLET-TIME ESTABLISHMENT Time decelerates dramatically to approximately 2–3% normal speed. DO NOT immediately go into macro. FIRST show a clear medium-wide frozen view of the entire accident. The protagonist is beginning to fall backward. BOTH HANDS ARE EMPTY. Establish fixed WORLD-SPACE anchors: • sunglasses suspended near their release position; • coffee cup suspended lower and toward the LEFT-hand release side; • partially detached lid still connected to or immediately beside the cup; • restrained coffee beginning to escape through the opening; • OPEN flip phone suspended toward the RIGHT-hand release side. Objects move only microscopically. They do NOT move toward the camera. The CAMERA is the main moving element. 5.2–6.8s — SUNGLASSES HERO INSPECTION The sunglasses remain almost completely stationary at their existing world-space position. CAMERA physically dollies toward them. The sunglasses become larger ONLY because the camera gets closer. Once close, slow the camera dramatically. HOLD. Perform a restrained 20–30 degree camera orbit/parallax around the almost-stationary sunglasses. Reveal: dark shield lens, thin metallic/gold frame, temples, hinges, edges, material thickness, surface reflections, tiny manufacturing details. Use shallow photographic macro depth of field. The protagonist and sunset environment remain softly recognizable in the background. The sunglasses must look like a real luxury fashion accessory photographed with a macro cinema lens, NOT CGI product animation. Complete the inspection before moving away. 6.8–8.3s — COFFEE / LIQUID INSPECTION Only after completing the sunglasses inspection, the CAMERA leaves them at their existing world-space position. Travel diagonally through the frozen scene toward the SAME coffee cup. THE CUP DOES NOT MOVE TOWARD THE CAMERA. Hold a photographic macro close-up. Show: same paper cup, same lightweight plastic lid, ONE partially detached lid edge, clear physical gap, dark coffee escaping ONLY through that gap. Real restrained liquid physics: thin irregular stream, imperfect liquid sheet, several differently sized droplets, realistic surface tension, natural reflections. No explosive splash. No decorative CGI liquid. No perfect artificial shapes. Coffee advances only microscopically during bullet time. Perform a small restrained lateral/parallax camera movement around the cup, lid and droplets. HOLD long enough to see paper texture, plastic texture, liquid thickness and individual droplets. 8.3–9.8s — FLIP PHONE INSPECTION Only after coffee inspection, the CAMERA travels away from the cup. The cup, lid and droplets remain at their existing world-space positions. CAMERA moves through the same frozen scene toward the SINGLE OPEN silver-grey clamshell phone. THE PHONE DOES NOT MOVE TOWARD THE CAMERA. First establish the complete phone silhouette: same silver-grey housing, same upper display, same lower keypad, same central hinge, same open angle. Then CAMERA physically approaches the existing upper display. Extreme photographic macro close-up. The illuminated display reads: BOSS with one small realistic phone number underneath. HOLD. Show: real housing material, hinge, keypad texture, small scratches, surface wear, screen reflections. The phone does NOT transform to reveal its screen. It is the SAME phone seen from the beginning. The protagonist’s RIGHT HAND remains completely EMPTY in the background. 9.8–10.4s — COMPLETE FROZEN TABLEAU After phone inspection, CAMERA performs a fast but perfectly smooth dolly backward. All objects remain at their existing world-space positions. Reveal: same protagonist falling backward, both hands empty, same sunglasses, same coffee cup, same partially detached lid, same suspended coffee droplets, same OPEN flip phone, exactly one secondary person, exactly one Labrador. All positions must remain spatially consistent with the original collision and release trajectories. No object follows the camera. No object teleports. No object regenerates. Brief HOLD on the complete frozen tableau. 10.4–12.0s — TIME RETURNS / PHYSICAL PAYOFF Time suddenly snaps back to normal speed. Gravity and existing momentum resume simultaneously. Every object continues FROM ITS EXISTING WORLD-SPACE POSITION. The protagonist completes her backward fall naturally. The sunglasses continue from their suspended position and land according to their existing momentum and gravity. The coffee cup continues from its own separate position, impacts the cobblestone ground, the partially detached lid reacts naturally, and remaining coffee spills irregularly through the real opening and impact point. The OPEN flip phone continues from its own separate position, hits the ground and makes ONE small physically believable bounce. Objects land at DIFFERENT plausible locations. They do NOT suddenly converge beside the protagonist. CAMERA rapidly tilts/drops toward ground level to follow the final impacts. Finish with a cinematic low-angle composition showing the fallen protagonist and the scattered objects within the same believable sunset environment. Hard cut. BULLET-TIME CAMERA RULE — NON-NEGOTIABLE BULLET TIME = ALMOST-FROZEN PHYSICAL WORLD + FREELY MOVING CAMERA. Objects stay in WORLD SPACE. Camera moves through WORLD SPACE. Close-up is created ONLY by reducing camera-to-object distance. Never: move objects toward the lens, artificially enlarge objects, teleport objects, regenerate macro versions. STRICT CAMERA ROUTE: FULL FROZEN TABLEAU → CAMERA DOLLIES TO SUNGLASSES → HOLD → 20–30° CAMERA ORBIT → CAMERA TRAVELS TO COFFEE → HOLD → SMALL CAMERA PARALLAX → CAMERA TRAVELS TO PHONE → ESTABLISH COMPLETE PHONE → MOVE TO DISPLAY → HOLD → CAMERA DOLLIES BACK TO COMPLETE TABLEAU → TIME RETURNS CONTINUITY / CAUSALITY Wide shot → medium shot → macro shot = SAME physical object viewed from different camera distances. Sunglasses: same object, same design, same trajectory, same world-space identity. Coffee: same cup, same lid, same coffee, same rotation, same world-space identity. Phone: same open clamshell phone, same housing, hinge, keypad, display, open angle and world-space identity. Every visible change must have a physical cause. Collision causes recoil. Collision momentum separates sunglasses. Finger release changes phone position. Hand release + collision momentum rotates cup. Cup rotation/liquid inertia partially opens lid. Coffee escapes only through the physical gap. Camera movement creates apparent object scale. Gravity resumes final fall. Treat the entire sequence as ONE continuous practical-effects event photographed with a motion-control cinema camera. IDENTITY / PROP LOCK ONE protagonist. ONE secondary person. ONE Labrador. ONE leash. ONE pair of sunglasses. ONE coffee cup. ONE lid. ONE OPEN flip phone. Before collision: RIGHT HAND = phone. LEFT HAND = coffee. Sunglasses = worn. After collision: RIGHT HAND = EMPTY. LEFT HAND = EMPTY. Sunglasses = separated. No second phone. No second cup. No second sunglasses. No additional dog. No additional person. No unexplained props. No duplication. No teleportation. No unexplained scale changes. VISUAL PRIORITY 1. exact protagonist identity from REFERENCE IMAGE 1 2. natural eyes and facial anatomy 3. stable hairstyle and body proportions 4. exact clothing and accessories 5. exact sunglasses from REFERENCE IMAGE 3 6. stable phone geometry 7. empty hands after release 8. continuous physical trajectories 9. realistic environment inspired by REFERENCE IMAGE 2 PHOTOGRAPHIC REALISM Ultra-photorealistic LIVE-ACTION. High-budget international luxury fashion campaign. Real human anatomy. Natural skin pores. Natural facial micro-expressions. Detailed realistic hair. Real manufactured sunglasses. Real paper cup. Real plastic lid. Real liquid coffee. Real early-2000s flip phone. Real object mass, inertia, gravity, momentum and surface tension. ARRI Alexa 35 cinematic character. Premium cinema optics. Motivated sunset lighting inspired by REFERENCE IMAGE 2. Controlled highlights. Natural skin tones. Realistic optical depth of field. Shallow macro depth of field during object inspections. Subtle 35mm film texture. Restrained luxury editorial colour grade. Natural colours. Sharp optical clarity. Cinematic photographic texture. Bullet-time must feel like REAL high-speed practical photography with a motion-control cinema camera, NOT CGI product animation. 4K / Ultra HD. Maximum facial detail. Maximum hair detail. Maximum material fidelity. Stable identity. Stable geometry. Stable prop continuity.show more

I just built an interactive companion cat fully powered... by OpenClaw🦞 and Claude She has a name: Luna At first, I used the Claude neural unit, which is the same type as those used in their adaptive robotics project. It runs on OpenClaw firmware, which handles instinct modeling, memory formation, and personality changes that develop over time. On the visual side, I added the ClaudeVision-Lite module, allowing her to recognize faces, track objects, and move autonomously around the room. I also installed SenseFur tactile fibers, so when you pet her, she can really "feel" it. The overall skeleton uses the FlexBone-X structure. The components used in the entire build include: Claude neural unit, ClaudeVision-Lite camera module, SenseFur tactile fibers, FlexBone-X skeleton, PurrSynth audio chip, MoodCell adaptive battery, RestPod wireless charging base, silicone skin casing, and TailSense tail actuator. Most of these components can actually be bought at local hardware stores, or on Taobao and eBay.show more

Jeremy Corbell’s New Saucer-Shaped UFO Debunked? According to ThomasH,... the round “flying saucer” that Jeremy Corbell (Jeremy Kenyon Lockyer Corbell) claims is the object in the video looks to be identified as just a glare from the sun. In the video below, you can see the object doing the same motion as the camera, but in reverse. This is typical behavior of what happens when a lens flare occurs. Thomas inverses the footage, showing the clouds and the correct shadows, which exposes the glare. Did Corbell tamper with the video? You have to wonder why Corbell admittedly adjusted setting in the video and even used an AI tool to enhance it. Is that because Corbell knew it was a glare and giving it an AI enhance would create sharper edges to make it appear to be something it’s not? Regardless, this is very hard to dispute and looks to be a proper debunk of a video that people like Marik vR from The Hill called “undebunkable”. #ufoX #ufotwittershow more

single RTX 3090. 24 GB VRAM. Qwen3.5-35B-A3B. 4-bit quant,... 113 tokens per second at full 262K context harnessing Claude Code locally with no API, no subscription, no proxy. told it what it is. 30 Mamba2 layers, 10 attention, 256 experts, 8 active per token. said "build something that shows off what you can do." it visualized its own architecture. interactive. tokens flowing through layers. 256 experts lighting up on routing. served in the browser from the same GPU running inference. single prompt. then i said level up. 3D. Three.js. separate files. flythrough camera. clickable layers. it planned first, scaffolded 6 files, hit one API bug, fixed it itself, then optimized for smooth framerate. two iterations to a working 3D neural network explorer. llama.cpp just merged a native Anthropic endpoint. Claude Code points at localhost. the whole setup is two commands. no LiteLLM. no proxy config. the open source models coming out of china right now are genuinely changing what's possible on consumer hardware. respect to the Qwen team. this is acceleration.show more

Weekly Outfit Showcase📺 prompt 👇 Referencing the face, hair,... and body proportions of image_1 Standalone_pink_dress_woman_img — a young Asian woman with long dark brown-black hair worn loose, delicate facial features, slim figure — as the consistent lead character throughout all 5 segments. Fixed camera, vertical 9:16 mid-long shot facing a European-style building entrance: dark green vintage double wooden doors, grey stone wall facade, two stone steps leading to a sidewalk, warm golden afternoon sunlight raking diagonally across the scene, golden-hour color grading, photorealistic handheld smartphone aesthetic. Shot 1 [00:00–00:02.5] — Monday: Hard cut open. The woman walks out through the green double doors wearing a white bandeau tube top and low-rise ripped denim micro shorts, white strappy heeled sandals, a chain mini bag on one shoulder. She descends the two stone steps and takes two to three confident strides toward camera along the sunlit sidewalk, waistline fully exposed in warm light, glancing sideways with a relaxed smile. White sans-serif caption "Monday" centered on screen. Shot 2 [00:02.5–00:05] — Tuesday: Hard cut. Same woman, same face and black hair, now wearing a black deep-V satin slip mini dress with thin spaghetti straps and strappy heeled sandals. Satin fabric catches flowing highlights in the afternoon sun. She steps out from the doorway, descends the steps, raises one hand to sweep her long hair back as she walks toward camera, hemline swaying with each stride. Caption hard-cuts to white sans-serif "Tuesday". Shot 3 [00:05–00:07] — Wednesday: Hard cut. Same woman now in a white halter-neck tie-front crop top (fully open bare back) and black high-waist micro pleated mini skirt with thigh-high socks. She exits the door talking on a phone held to her ear, walks down the steps, then pivots to reveal her entire bare back to camera before continuing forward. Caption hard-cuts to "Wednesday". Shot 4 [00:07–00:09.5] — Thursday: Hard cut. Same woman in a beige open-knit crochet halter maxi dress with diamond-shaped side cutouts at the waist and a high thigh slit on both sides, Roman lace-up flat sandals. She strolls out the door, glances back over her shoulder toward the building, the slit opening and closing rhythmically with each stride as she walks toward camera. Caption hard-cuts to "Thursday". Shot 5 [00:09.5–00:12] — Friday: Hard cut. Same woman in a red structured strapless tube mini dress with side waist cutout and a small side slit, red pointed-toe stiletto heels, dark sunglasses perched on nose, holding an iced Americano cup in her right hand. She walks confidently out the door and directly toward the lens, getting closer until she fills the lower frame, then raises her free hand to remove the sunglasses, looks straight into camera, raises one eyebrow and breaks into a bright smile. Caption hard-cuts to "Friday". Frame freezes on her smiling face. Audio: upbeat syncopated background BGM at low volume throughout; sharp high-heel clicks on stone pavement synced to footsteps; a crisp single "click" sound effect on each hard cut between outfits. No dialogue.show more

What if luxury livestream shopping ads could be generated... entirely with AI? From the influencer, to the boutique, to the product showcase, every frame was created with GPT Image 2 and Seedance on BudgetPixel AI The future of social commerce content is here. Prompt: 15s ultra-realistic vertical smartphone livestream advertisement, 16:9 aspect ratio, luxury social-commerce style. Use @[char ref] and @[storyboard ref]. Follow the scene sequence with seamless continuity and fast-paced livestream energy. 0:00–0:02 — Livestream opens. Front phone-camera perspective with slight handheld movement. The host smiles warmly while holding a premium black quilted crossbody handbag toward the lens. A realistic illuminated sign reading "TOKYO LIVE FLASH SALE" glows naturally in the background. Dialogue: "Hi everyone~ look at this beautiful bag." 0:02–0:04 — Medium close-up. She rotates the handbag slowly, showing the quilted pattern and gold hardware while maintaining friendly eye contact. Dialogue: "So elegant, right?" 0:04–0:06 — Extreme macro close-up. Her fingers gently glide across the quilted leather texture and premium clasp. Warm highlights reflect realistically across the material. Dialogue: "Soft, premium, luxury." 0:06–0:08 — Top-down product demonstration. She opens the bag and places a smartphone, wallet, lipstick, and power bank inside. Shopping notifications appear. Dialogue: "Everything fits perfectly." 0:08–0:10 — Tracking shot. She wears the bag crossbody and walks naturally through the boutique. Hair moves softly as she turns and smiles. Dialogue: "So comfortable." 0:10–0:12 — Full-body fashion shot. She poses naturally, showing how the bag complements her outfit. Viewer count rises rapidly. Dialogue: "Matches every look." 0:12–0:14 — Dynamic push-in. Realistic LED-style sale graphics illuminate behind her: TOKYO LIVE FLASH SALE 50% OFF LIMITED TIME Countdown Timer Dialogue: "Fifty percent off today!" 0:14–0:15 — Hero product close-up. She brings the handbag toward the camera with a warm smile as hearts and comments flood the screen. Dialogue: "Shop now before it's gone." Audio: Upbeat Japanese city-pop instrumental, subtle livestream notification sounds, floating-heart effects, boutique ambience. Photorealistic live action, natural facial expressions, realistic motion, luxury commercial quality, high-conversion livestream energy. Maintain the same host, same hairstyle, same makeup, same outfit, same handbag, same boutique environment, same lighting, and same sale branding throughout the entire video. No subtitles. No watermark. Strong product focus. Premium luxury fashion-commercial quality.show more

ギャル動画のPrompt Use the uploaded image as the first-frame reference.... Preserve the same fictional adult Japanese gyaru woman throughout the entire video. Keep the same face, Japanese gyaru appearance, long blonde curled hair with dark roots, glamorous gyaru makeup, tanned glowing skin, glossy pink lips, long eyelashes, long pink nails, delicate gold necklace, fitted white ribbed sleeveless crop tank top, and light-wash denim short shorts. Maintain the same identity, outfit, hairstyle, makeup, body proportions, and overall mood in every scene. Main subject: a fictional adult Japanese gyaru woman in her mid-20s. She feels real, casual, and familiar rather than staged. Her expression is naturally confident, relaxed, and playful, but never overly posed. She should always feel like the same woman from the reference image. Location: a quiet Japanese nightlife area at night, moving through intimate everyday spaces near the same building. Start in a dim private hallway, then shift through a small elevator, a building entrance, a vending machine corner, a convenience store front, and a quiet side street or walkway. The environment should feel local, low-key, slightly nostalgic, and believable. No crowds, no flashy club scenes, no dramatic urban spectacle. Visual Style: ultra-realistic documentary realism with a strong early-2000s home video feeling. The footage should look like it was casually recorded by a friend on an old consumer DV camcorder or early digital home video device. Emphasize imperfect realism: faded colors, soft contrast, occasional blown highlights, crushed shadows, visible low-light noise, mild digital compression artifacts, slight ghosting, inconsistent white balance, weak dynamic range, and a slightly aged consumer-video texture. The result should feel like a forgotten personal memory clip, not polished social media content. Camera Style: early-2000s consumer home video aesthetic. A friend casually records short everyday moments with no filming skill. Heavy handheld shake, unstable framing, awkward zoom distance, autofocus hunting, occasional lens breathing, delayed reframing, slight motion smear, exposure pumping when moving between dark hallway light and brighter vending machine or store light, and moments where the camera catches up too late. No stabilization. No cinematic camera moves. No slick transitions. No modern color grading. The footage should feel homemade, clumsy, and authentic. 00:00–00:02 Inside a dim private hallway at night. Start exactly from the reference image pose. She stands near the wall and looks back toward the camera with a soft, confident smile. The camera takes a moment to lock focus on her face. Warm wall light and direct camera light catch her blonde curls and white top. The framing wobbles slightly as if the person filming just raised the camera. 00:02–00:04 Hard cut to a small elevator. She casually lifts one hand to adjust her hair while glancing toward the mirror, then looks back at the camera. The camera is too close and slightly tilted. Focus briefly shifts to the wall or mirror before returning to her. The moment feels unplanned and personal. 00:04–00:06 Hard cut to the building entrance. She steps outside into the night air and pauses. A light breeze moves loose strands of hair. The exposure flares slightly while the camera adjusts from the darker interior to the brighter outside light. She glances down the quiet street and then back toward the person filming. 00:00–00:08 Hard cut to a vending machine corner on a quiet side street. She bends slightly to look at the drink choices and presses a button. Autofocus shifts uncertainly between her hand, her face, and the bright machine lights. The camera drifts off-center as if the operator is distracted for a second. 00:08–00:10 Hard cut to outside a convenience store. She stands with a small drink in hand, watching the quiet street at night. Someone off-camera says something casual. She reacts with a small smile and brushes her hair behind her ear. The shot is a loose side angle with natural drift and minor blur, like a friend filming without thinking. 00:10–00:12 Hard cut to a quiet walkway near the building. She walks slowly under streetlights holding the drink. She notices the camera, gives a brief genuine smile, then looks away and keeps walking. The framing is imperfect and slightly late, as if the camera operator nearly missed the moment. 00:12–00:15 Hard cut back near the hallway or elevator lobby. Close side profile. Someone off-camera casually calls out to her. She turns, raises one hand slightly in acknowledgment, smiles warmly, and starts to move away. The camera catches the reaction a little too late, and the recording cuts abruptly to black mid-motion as if the camcorder was switched off by accident. Audio: natural ambient sound only — faint hallway hum, elevator noise, vending machine hum, convenience store ambience, distant traffic, light footsteps, soft clothing movement, quiet night air, subtle handling noise from the camera device. No music. No narration. No added sound design. Goal: authentic Japanese nightlife fragments captured like a forgotten early-2000s personal home video — candid, clumsy, imperfect, warm, nostalgic, and believable. The same Japanese gyaru woman must remain visually consistent throughout the entire video, while the footage feels like short private memories recorded by a friend rather than a polished clip. Negative prompt: do not change her face, do not change her ethnicity, do not change her Japanese gyaru appearance, do not change her blonde curled hairstyle, do not change her outfit, do not add extra people near her, do not add text, do not add logos, do not switch to daytime, do not make it a glamorous music video, do not make it a fashion commercial, do not make it cinematic, do not make it anime, do not over-smooth skin, do not distort hands, do not distort face, do not warp body proportions, do not create extreme glamour posing, do not create nudity, do not make the image too clean or too polished.show more

how to prompt undetectable ai shots while designing a... running scene, first think about these 3 basic questions: how does the camera move? what is it looking at? where does it stop? a good motion prompt is really just a timeline it needs to follow real-world physics, and it needs to carry story at the same time 1. start with the narrative goal of the shot camera movement is not just movement it is the storytelling so before writing the prompt, define what the shot is trying to do for example, in a 15-second one take, the goal could be: follow the female lead laterally while she runs, to build speed and tension then briefly reveal the people chasing her then let the camera hesitate for a moment and find her again that small “lose and recapture” moment adds spatial depth and makes the chase feel more intense 2. build a clear space for the camera to work in if the space is vague, the shot gets messy very fast i like breaking the scene into layers so the model knows where everything belongs foreground: passing objects, environmental motion main subject layer: the woman running midground: cafe tables, pedestrians, the people chasing her background: the vanishing point of the street, and the entrance to the pedestrian area once the space is clear, the camera has a stage to move through 3. describe motion like a physical process a good moving shot has to respect inertia if the movement feels weightless or too perfect, it instantly feels fake so instead of using broad words, describe a chain of actions the camera can actually perform what accelerates what slows down when it adjusts when it slightly overshoots when it catches itself again that little bit of imperfection is usually what makes it feel real 4. use focus as part of the storytelling in a one take, focus is one of the best ways to guide attention you can design moments where focus shifts with intention for example: focus briefly drifts from the chasers’ faces, passes beyond them, lands on the woman in the distance, then quickly pulls back again it feels like a small mistake, but that’s exactly why it works it simulates a camera operator re-evaluating the subject in the middle of a fast-moving shot that kind of temporary focus loss and recovery adds a lot of immediacy and documentary feeling 5. build the sound space with the camera sound should move with the shot when the camera turns toward the chasers, the woman’s breathing should fall deeper into the sound field, while the chasers’ footsteps and breathing move to the center when the camera finds the woman again, her breath becomes the main sound again that shift in audio perspective helps the scene feel much more immersive it’s not just about what we see it’s also about where we feel the scene from 6. use negative constraints to stop common ai mistakes this part matters a lot i usually add clear “don’ts” at the end to stop the model from breaking the shot for example: no cuts no teleporting zooms no sliding characters no body fusion no floating props the travel bag must keep believable weight and inertia these negative constraints act like guardrails they help keep the result inside a believable physical world for me, the core of a strong motion prompt is simple: organize space, camera, focus, action, and sound into one executable timeline that’s really the difference instead of prompting a vague feeling, you’re designing a physical process and that’s usually what helps ai generate a moving shot that feels coherent, grounded, and full of tensionshow more

a moonshot engineer leaked the benchmark anthropic, openai and... xai all buried the same week: kimi k3 beat opus 5, gpt-5.6 and grok 4.6 at $0.94 a task. stop paying anthropic $200 a month for opus 5 and openai $200 for gpt-5.6 when kimi does the same work for $8 the leak showed kimi k3 winning 9 of 12 categories against opus 5, gpt-5.6 and grok 4.6. within 48 hours all three labs quietly pushed pricing pages and one very specific comparison chart off their sites. nobody announced anything. they just deleted, which tells you everything the four numbers they scrubbed: cost per task · $0.94 vs $1.80 -> opus 5 charges $1.80 to finish one task. gpt-5.6 $1.04. grok 4.6 $0.61. kimi k3 $0.94 and it landed 487 of 500 clean -> anthropic is billing you double for a model that lost the benchmark it paid to promote the weights · free, sitting on huggingface right now -> the entire model is a public download. pull it, keep it, run it forever, nobody can switch it off -> a model you can hold cannot be rented at $200 a month. that single fact is what three labs deleted a chart over the switch · one line of bash -> moonshot ships an anthropic-compatible endpoint. one env variable and claude code points at kimi -> same cli, same keybindings, same /model. you change a url, opus 5 never knows it lost the seat the bill · $400 down to $8 -> opus 5 max plus gpt-5.6 pro is $400 a month. kimi runs the same daily work for $8 metered -> that is a 98% cut for output that beat both of them 9 categories to 3 here is the part they will fight me on: the frontier tax died the week this leaked and all three labs know it. once the weights are public the price has a ceiling, because anyone can serve the same model. anthropic, openai and xai are charging 2025 prices on a lead that ended in a benchmark they deleted instead of answered drop your $400/mo ai stack to $8. the run above is kimi k3 finishing the task opus 5 bills $1.80 for. the full breakdown is in the article belowshow more

more seedance 2.5 realism tests: > it always makes... the arms a bit too long > shot selection is pretty insane > audio is a bit sub par, needs more work > in aggregate, can 100% pass as real (read the article below for full guide) prompt: Create an exactly 15-second photorealistic shopping vlog in 16:9: extremely sharp 4K compact-camera detail, natural HDR, accurate skin, realistic motion and perfect identity continuity. No VHS, grain, subtitles, logos, watermarks or graphics. SUBJECT: A 21-year-old American woman with long wavy chestnut hair, light everyday makeup and an expressive, approachable face. Fitted black long-sleeve top, light-blue straight jeans, sneakers and a small black shoulder bag. Keep her face, hair, outfit, bag and jewellery identical through every cut. She talks like a normal friend, never a presenter. STORE: A large modern high-street womenswear shop resembling an upscale European fast-fashion retailer without real branding. Cream walls, black fixtures, mirrors, warm-neutral spotlights, polished floor, fitting rooms, folded denim and dense racks of blazers, tops, dresses, knitwear and trousers. Adult shoppers browse independently. No readable brands, signs, price tags or labels. CAMERA: The same compact camera stays in her hand throughout. Every frame is one-handed arm’s-length selfie, her mirror reflection holding it, or her rotating it outward to show clothes. Never place it down, prop it up, pass it to anyone or use an external angle. Sharp 20–24mm-equivalent wide lens, mild barrel distortion, deep focus, reliable eye autofocus and true 24fps with 1/48-second blur. Stabilization controls impacts but retains footstep bob, wrist sway, breathing, grip shifts, late pans and imperfect reframing. Preserve pores, flyaways, teeth, fabric and accurate reflections without beauty filtering or fake bokeh. Exposure stays face-priority but shifts subtly between mirrored aisles and the fitting-room corridor. Fast wrist turns show slight rolling-shutter skew and natural background blur while her face and garments remain sharply resolved. EDITING IS CRITICAL: Use 12 clips and 11 hard jump cuts. Most cuts remain in the same aisle or beside the same rack, with only a small jump in pose, expression, camera distance, garment position or sentence. Remove mundane pauses and bits of walking. Cut mid-gesture, during camera turns and inside sentences for the fast rhythm of an edited social vlog. Geography and identity stay coherent despite time compression. No fades, transitions, whip effects or cinematic montage. Cuts are not beat-synced; some happen for no narrative reason and simply skip a fraction of an ordinary action. CLIP 1 — 0.0–1.2: Walking into the store in loose selfie framing: “Okay, I’m literally just getting—” CLIP 2 — 1.2–2.3: Jump cut a few steps inside, nearly the same angle but her hair and bag have shifted: “—one top. That’s it.” CLIP 3 — 2.3–3.4: She rotates the camera toward a crowded rack and slides hangers with her free hand. Hangers clack; no dialogue. CLIP 4 — 3.4–4.6: Jump cut beside the same rack. In selfie view she holds a dark-red top against herself: “Wait, this is cute.” CLIP 5 — 4.6–5.5: Cut to a nearby mirror, camera visible in her hand. She checks the top. Her female friend, mostly off-camera, says, “No.” She turns: “Really?” CLIP 6 — 5.5–6.4: Close selfie micro-cut in the same spot. She laughs: “Okay, rude.” CLIP 7 — 6.4–7.6: She rotates toward a blazer rail and pulls out a charcoal jacket. Focus briefly transfers from mirror to fabric. CLIP 8 — 7.6–8.8: Jump cut in the same aisle. The top and jacket now rest over her free arm. Looking down: “Why do I already have four things?” CLIP 9 — 8.8–9.8: Near-identical angle after another micro-cut. Friend: “Because you picked them up.” She smiles: “Fair.” CLIP 10 — 9.8–11.2: Cut to her walking past another rack. She pans across dresses, passes the item she wants, then corrects back. No dialogue. CLIP 11 — 11.2–13.2: Mirror near the fitting rooms. Holding the jacket against herself: “Do I need this?” Friend: “No.” She smiles, ignoring the answer. CLIP 12 — 13.2–15.0: Close walking selfie approaching the fitting rooms, garments over her free arm: “Okay, I’m trying it on.”show more

Jev + GrokBot is the best GTM system I've... ever built It just made my GTM x200 CHEAPER and x400 FASTER than what 95% of teams are running... setup takes literally 9 minutes: prompt → GrokBot → Jev routes every candidate → GrokBot opens only survivors → campaign step 1 → open TypeSafe AI, create API key (keep it off chat paste) step 2 → tell GrokBot: store TYPESAFE_API_KEY in the secure field step 3 → prompt GrokBot: install typesafe-sdk on Agent Computer + smoke system_one (Choice) step 4 → point it at a target: "read end to end, pull positioning, ICP, every proof point. no copy yet" step 5 → add skill gtm-router: every candidate from X / LinkedIn / YouTube gets 6 typed questions → CHEAP SKIP / OPEN FULL / FLAG KOL step 6 → let Jev rank hook archetypes by viral rate, not by how often people post them - this is where the whole playbook comes from step 7 → GrokBot assembles the KOL shortlist and the 7-day plan - you stay in control of anything irreversible the result: → 3,412 candidates, X, LinkedIn and YouTube. → 15.7 seconds. $0.41. → me doing the same reading: 6h 12m. 8 grok bots brief it, Jev decides, GrokBot executes. no fine-tuning, no embeddings, no vector db - 6 typed questions and a schema. Setup your GrokBot JEV GTM today, then read the article below to learn how to build an agent team with JEV.show more

What would you do if a monster charged at... you in the middle of the street? Made with Nano Banana 2 + Seedance 2.0 4k on TapNow prompt Beat 1 - The Calm Opening The reference image character is a Western woman with long wavy blonde hair casually drinking a Coke in the middle of a chaotic urban street. Her outfit and cap match Image 1 exactly. Energy is calm, unfazed, and confident. Camera: grounded, cinematic, medium shot. Beat 2 - The Monster Charges A terrifying monster bull charges toward her at full speed. It smashes through debris and scatters people in panic. The street erupts into chaos around the character. Camera: dynamic tracking, wide angle to capture scale of monster and chaos. Beat 3 - The Transformation The young woman stays calm and completely unfazed. She lowers the Coke. She raises one hand. Her arm begins transforming into solid ice from the fingertips upward. Frost spreads across her skin, glowing with cold blue-white energy. Vapor rolls off the arm as the ice transformation builds. Camera: close-up on the arm transformation, slow and deliberate. Beat 4 - The Touch - Slow Motion As the monster lunges, she calmly reaches forward. She touches the monster with her frozen hand. DRAMATIC SLOW MOTION at the moment of contact. Ice rapidly spreads from the point of contact across the monster's body. The monster panics - thrashing and roaring as frost crawls over its claws, chest, face, and eyes. The monster's movement slows as the freezing effect takes over. Camera: extreme close-up on contact point, then pulls out to reveal the spreading ice. Beat 5 - Full Freeze The creature becomes completely encased in crystal-clear ice, frozen mid-attack. A tense pause holds on the frozen monster. Deep cracks spread through the frozen monster. Cracks split across its icy body with sharp crystalline sounds. Detailed cracking ice - realistic frost and crystal texture throughout. Camera: wide shot of the frozen monster, locked and still for the pause beat. Beat 6 - The Walk-Off The young woman casually takes another sip of Coke. She turns away without looking back. She walks off calmly as the frozen monster cracks behind her. Confident calm hero energy. No celebration. No reaction. Just casual. Camera: low angle tracking shot as she walks away, monster in background. Visual Effects Realistic ice transformation - frost spreading from fingertips upward. Frost vapor rolling off the arm during transformation. Cold blue-white glowing energy across frozen skin. Ice rapidly spreading across monster body from point of contact. Detailed cracking ice with sharp crystalline fracture lines. Crystal-clear ice encasing the entire monster mid-attack. Realistic freeze physics - movement slowing as ice covers the creature. Camera Style Dynamic cinematic camera movement throughout. High-action monster attack framing during the charge. Dramatic slow motion at the exact moment of hand contact. Close-up on ice transformation - fingertips to shoulder. Pull-back reveal as ice spreads across the monster's full body. Low angle walk-off shot to close the scene. One continuous take. No cuts. Camera never stops. Character appearance: a beautiful Western woman with long wavy blonde hair flowing naturally beneath the cap, maintaining the same calm, confident, and unfazed presence throughout the entire sequence.show more