Dheepan Ratnam (@Dheepanratnam) header
Dheepan Ratnam (@Dheepanratnam) profile picture

AI storyteller and digital artist | Weaving whimsical tales and vibrant artwork from the wild heart of AI. CPP of @higgsfield | From the island of Sri Lanka🇱🇰

London
Joined February 2025
Photo by Dheepan Ratnam (@Dheepanratnam)Photo by Dheepan Ratnam (@Dheepanratnam)Photo by Dheepan Ratnam (@Dheepanratnam)Photo by Dheepan Ratnam (@Dheepanratnam)Photo by Dheepan Ratnam (@Dheepanratnam)Photo by Dheepan Ratnam (@Dheepanratnam)Photo by Dheepan Ratnam (@Dheepanratnam)Photo by Dheepan Ratnam (@Dheepanratnam)Photo by Dheepan Ratnam (@Dheepanratnam)Photo by Dheepan Ratnam (@Dheepanratnam)
BITCRUSH 碎像素 Premium 16-bit pixel art (modern indie fighting game) MinimaxH3 image -> Seedance 2.5 Great for Creating game assets animation work 4k at low cost @MiniMax_AI
4
1
17
518
AI text rendering just crossed the finish line 🤯 Tested Seedream 5.5 on dense 2K infographic layouts—the text clarity, structural hierarchy, and render quality took an insane leap forward. Zoom in on these results more examples ⤵️
5
10
442
Seedreem 5.5 @ 2k resolution The Anatomy of a Specialty Coffee Extraction A vertical technical breakdown of an espresso machine group head, illustrating water temperature curves, 9-bar pressure profiles, grind distribution curves, and channel prevention micro-physics. Styled with rich crema browns, brushed stainless steel accents, and minimalist espresso-bar typography.
1
81
4K images in 10 seconds. And anyone can download the model 🤯 7 AI drops you missed today 🧵 follow for daily genAi news updates
5
17
721
1/7 MiniMax H3-Image sharp 4K pictures in about 10 seconds, and the weights are open. Free to run yourself. That's rare for this quality. xitter-newx.xyz-mikolabs.workers.dev/MiniMax_AI/status/2120…
4K image generation in 10 seconds! ⚡ Meet MiniMax H3-Image. 16 million pixel resolution. Stunning realism. Precise prompt following & editing. Up to 16 reference images. All in 10 seconds. And it's open-weight. 🚀 Available now via our API, MiniMax Code, and Hailuo AI — with 50% off 16MP+ generations during launch week! Can't wait to see what you create. 🔥 Hugging Face: huggingface.co/MiniMaxAI/Min… API: platform.minimax.io/docs/gui… MiniMax Code: code.minimax.io/about
1
2
106
Every designer has this type of client “Make it night.” “My wife prefers green.” “Add our dog.” “Now the menu. Prices too 8 jobs from the client from hell 😃 now you can deal with dave MiniMax H3 Image can make super fast images in 4k @MiniMax_AI #H3Image
Meet Minimax H3 image 🔥 its fast cheap and can do a great job
4
13
630
Meet Minimax H3 image 🔥 its fast cheap and can do a great job
4K image generation in 10 seconds! ⚡ Meet MiniMax H3-Image. 16 million pixel resolution. Stunning realism. Precise prompt following & editing. Up to 16 reference images. All in 10 seconds. And it's open-weight. 🚀 Available now via our API, MiniMax Code, and Hailuo AI — with 50% off 16MP+ generations during launch week! Can't wait to see what you create. 🔥 Hugging Face: huggingface.co/MiniMaxAI/Min… API: platform.minimax.io/docs/gui… MiniMax Code: code.minimax.io/about
5
1
22
1,149
Gemini 4 Carbon is here🔥 i Asked 2 questions to compare with fable 5.2 it gave me something Bottom line Stick with me (Gemini 4 Carbon) when your work mixes code with video/audio/dense visual files, involves security auditing and rapid tool-heavy automation, or needs frontier reasoning at a fraction of the token burn (tip: bump my effort setting up from Low to High if you throw a gnarly architecture or debugging problem at me). Reach for Claude Fable 5.2 when you want a second frontier opinion on an open-ended math/science problem, or are kicking off a multi-hour, text-and-code refactoring session where squeezing out every last percentage point on long-horizon coding benchmarks matters more than token cost or multimodal input.
Gemini 4 Carbon is here it's our strongest model with frontier performance in cybersecurity capabilities and long-horizon agentic workflows across complex knowledge work, software engineering, and scientific reasoning Gemini 4 Carbon is now live via the Gemini API and AI Studio for paid API and Google AI Ultra subscribers try it today: ai.studio/carbon
6
10
494
I got early access to Minimax H3 image created a anime character - Put into seedance 2.5 the result is insane🔥 👇 anime key frame, original character, perfectly symmetrical one-point-perspective composition. Japanese 2D hand-drawn battle anime crossed with Gothic stained-glass window art @MiniMax_AI #H3Image
14
2
33
1,015
One image. Pulled apart into layers. Blazer swapped, set changed, light rebuilt, and her face never moves. Every frame here is Seedream 5.5 🤯 Unlimited on AZ8 right now with any Pro, Ultra, Max or Extreme plan. Create. Edit. Repeat. Link in the comments ⤵👇
Seedream 5.5 Unlimited is now live on AZ8. Purchase any AZ8 Pro, Ultra, Max, or Extreme plan and get up to 30 days of unlimited access, depending on your plan. More creative freedom. Less detail loss. >Precise local edits >Consistent details across repeated edits >Transparent assets, ready to composite >Lifelike lighting, materials, and skin >Clearer small text Create. Edit. Repeat—with Seedream 5.5 Unlimited on AZ8.
11
1
35
959
Fable 5.2 is out, so I tested it on something stupid and fun Make any photo out of particles( fable 5.2 medium ) Runs in your browser, nothing uploaded 🔥🔥 You can swipe it like sand, blow it apart with your mic, and spin it in 3D. 4 modes have fun :) link 👇
Introducing Claude Fable 5.2 and Claude Mythos 5.2, designed to tackle the most daunting technical challenges. They’re capable collaborators on ambitious projects in many fields, and show improvements in areas like coding, reasoning, and vision.
9
3
1
30
1,590
Build faster by seeing your UI before you code. With Opus 5.5 + Gemini Omni Flash 1.2, simple text-to-video prompts generate UI variations in minutes perfect for inspiration, iteration @FlowbyGoogle prompts ⤵️
9
1
19
813
FLORA — Garden design platform “Let the Landscape Grow.” Make a UI animation for FLORA, a living garden-planning brand. Enter GARDEN from the FLORA homepage and unfold a topographic landscape from a sheet of textured paper. Drag SHADE beneath a tree to reveal suitable plants, draw WATER through the terrain and place a planting cluster that settles naturally into the soil. Expand SEASON to move the garden from spring to autumn, then combine WILD with QUIET to balance dense planting and open space. Select GROW: roots spread beneath the interface, vegetation rises around the user and the completed landscape returns to the FLORA brand mark. Chalky earth, sage green, weathered stone, watercolor shadows and late-afternoon sunlight. Use organic growth, soil displacement and gentle seasonal transitions. Avoid agricultural dashboards, plant-shopping cards, bright eco-tech graphics, artificial green neon and numerical growth counters. KILN — Architectural lighting studio “Shape the Quiet Hour.” Make a UI animation for KILN, an atmospheric interior lighting brand. Enter ROOM from the KILN homepage and raise a minimalist architectural volume from a slab of unglazed travertine. Drag DUSK across the floor to cast long, diffused shadows, draw WARMTH along a plaster wall and place a recessed light source that softly illuminates the grain of the stone. Expand HOUR to shift the space from late afternoon to midnight, then combine EMBER with STILLNESS to balance focal glow and ambient shadow. Select ILLUMINATE: light washes across the threshold, architectural edges dissolve into warm atmospheric haze and the space settles back into the KILN brand mark. Raw plaster, burnt sienna, smoked bronze, soft alabaster and twilight amber. Use subtle light falloff, tactile surface reflections and gradual tonal shifts. Avoid smart-home sliders, wattage gauges, color-wheel pickers, neon RGB strips and technical lumen charts. SILL — Bespoke fragrance atelier “Memory, Distilled.” Make a UI animation for SILL, a botanical perfumery and scent-design platform. Enter VESSEL from the SILL homepage and ripple open a translucent glass basin from a sheet of heavy cotton paper. Drag RESIN into the base to anchor the composition, draw VAPOR upward through the frame and place a botanical note that dissolves like ink in warm water. Expand CLIMATE to evolve the scent profile from crisp morning rain to dry desert heat, then combine SMOKE with PETAL to balance earthy depth and delicate lift. Select INFUSE: golden oils swirl together, condensation forms softly on the glass surface and the refractive droplets coalesce into the SILL brand mark. Amber glass, pressed cedar, raw linen, smoky quartz and diffused morning mist. Use fluid dispersion, delicate refraction and slow evaporative transitions. Avoid chemical formulas, percentage pie charts, e-commerce bottle carousels, spray mist graphics and clinical lab interfaces. LOOM — Tactile acoustic architecture “Listen to the Material.” Make a UI animation for LOOM, an acoustic textile and spatial sound brand. Enter STUDIO from the LOOM homepage and weave a three-dimensional room wireframe from fine raw linen threads. Drag FELT across a reverberant surface to soften the visual tension, draw WAVE through the open air and place a baffled canopy that settles gently overhead. Expand TIMBRE to shift the room's character from bright and airy to intimate and muted, then combine WEAVE with VOID to balance dense textile absorption and open architectural volume. Select HUSH: soundwaves visibly slow into calm ripples, woven fibers tighten into a seamless surface and the quieted space folds back into the LOOM brand mark. Undyed wool, charcoal slate, warm oak, brushed brass and muted gallery light. Use textile tension, soft fabric draping and damped ripple motion. Avoid decibel meters, equalizer bars, neon soundwave graphs, speaker hardware icons and studio mixing consoles. TIDE — Coastal bathhouse & restorative spa “Return to Still Water.” Make a UI animation for TIDE, a mineral thermal bath and wellness sanctuary. Enter BASIN from the TIDE homepage and carve a stepped limestone pool out of a calm, misty horizon. Drag MINERAL into the shallows to deepen the water's hue, draw CURRENT along the stone ledges and place a thermal spring that sends gentle warmth rippling outward. Expand TIDE to raise the water level over the weathered steps, then combine STEAM with STONE to balance enveloping mist and grounded architectural mass. Select IMMERSE: the camera sinks smoothly beneath the water's surface, light refracts in slow caustic patterns and the rising ripples resolve into the TIDE brand mark. Wet basalt, sea salt, pale driftwood, overcast sky and warm thermal vapor. Use slow water displacement, caustic light reflections and rising steam dissipation. Avoid heart-rate monitors, booking calendars, temperature dials, glowing wellness badges and hydration trackers.
4
249
byteplus announced seedreem 5.5🔥🤯 stable multi-turn editing across up to 100 rounds, native RGBA transparent assets, and more realistic lighting/materials/skin plus clearer small text. Lets check it out Testing Seedream 5.5 Thread (Continuous updates) ( all single generations no cherry picking- @ native 2k resolution Share your findings or comparisons. The thread below 👇 I’ll be posting more generations and prompts in the comments throughout the day. Feel free to join
What happens after 100+ rounds of “just one more change”? How far can you push an image before it starts to drift? Meet Seedream 5.5: precise local edits, consistency across repeated revisions, more natural visuals and native transparent assets. Explore what’s new below. 👇 byteplus.com/en/contact-us/s…
9
1
1
54
3,512
typography stress test Prompt 5 dense small text Premium magazine cover design titled “THE FUTURE OF CREATIVE TOOLS”, with a realistic portrait of a stylish creative director in the lower third. Dense small cover lines around the edges: “Issue 24”, “AI Workflows”, “Design Systems”, “Cinematic Image Making”, “Inside the Studio”, “November 2026”. Clean modern serif masthead, precise kerning, crisp letterforms, editorial grid, muted cream background, subtle paper texture, high-end print-design layout. All text must be spelled exactly as written, no distorted letters, no duplicated words, no gibberish. Test for: dense small text, spelling, kerning and layout discipline. This is one of Seedream 5.5’s headline improvements.seed.bytedance Prompt 6 multilingual signage Photorealistic street scene in a multicultural London neighbourhood at night. A shopfront sign reads exactly “OPEN LATE” in English, with a smaller hanging sign reading exactly “欢迎光临” in simplified Chinese and another small poster reading exactly “Bienvenue” in French. Realistic neon glow, legible typography, correct character shapes, no duplicated letters, no warped glyphs, natural perspective on the signs, cinematic night photography, wet pavement reflections, no extra invented text. Test for: multilingual text, complex character shapes and sign perspective. Prompt 7 packaging mock-up Minimalist skincare packaging mock-up on a pale stone pedestal. A frosted-glass serum bottle has a clean label reading exactly “LUMEN — BARRIER REPAIR SERUM”, with smaller text beneath: “30 ml / 1.0 fl oz”. Elegant sans-serif typography, perfectly straight label alignment, crisp small print, realistic glass refraction, soft studio lighting, subtle shadow, premium cosmetic advertising photography, no misspelled words, no extra branding, no watermark. Test for: small label text, straight typography and commercial product layout. Prompt 8 poster typography Bold gig poster for an underground electronic music night. Large headline reads exactly “NIGHT SIGNAL”, with supporting text: “Friday 27 November”, “Warehouse 9, London”, “Doors 10PM”. Experimental Swiss-style layout, high-contrast black, off-white and electric blue, grainy risograph texture, sharp geometric typography, intentional negative space, screen-print aesthetic. All words spelled exactly as specified, no extra text, no warped letters, no AI artefacts.
4
307
She knocked the whole world out of register, then punched it back into print." 💥🎨 Introducing EP01 of MISPRINT where Japanese 2D battle anime crashes head-first into gritty, ink-heavy risograph poster art. Hit play, turn the volume up, and let me know: would you watch a full series like this? 👇 @dreamina_ai #DreaminaCPP #anime
7
25
648
Midjourney 8.1 character Start frame image prompt 16:9 anime key frame, Japanese 2D hand drawn battle anime crossed with risograph screenprint poster art. Young woman fighter: chin length bleached bob, blunt bangs, fluorescent pink tips, fierce narrow eyes, pink halftone dot patch on right cheek. Oversized cobalt blue bomber jacket worn open off one shoulder, black high neck top, black shorts over tights, chunky black boots with yellow laces. Left fist wrapped in yellow tape; right forearm in heavy black steel gauntlet with yellow hazard stripes and three rubber roller rings on knuckles. Low coiled crouch right of centre, gauntlet drawn back, left hand forward. Three colour separated afterimages fluorescent pink, cobalt blue, lemon yellow streak left, offset like misprinted poster, edges dissolving into halftone dots. Setting: cream textured paper, low horizon, three giant flat cobalt blue geometric pillars, vast negative space. Four ink colours: fluorescent pink, cobalt blue, lemon yellow, black; overprints create purple, orange, green. Visible halftone shading, misregistration, grainy riso ink, bold anime line art, cel shading. No text, no logo, no 3D.
2
157
divine series ANANTA-PADMANABHA
Divine series NATARAJA: The Ring of Cosmic Fire on the Black Mirror Sea seedance 2.5 on @higgsfield
401
Divine series NATARAJA: The Ring of Cosmic Fire on the Black Mirror Sea seedance 2.5 on @higgsfield
MAYIL-VELAN (MURUGAN) The Golden Spear That Parted the Monsoon Sky #Murugan #AIfilm seedance 2.5 on @higgsfield
3
1
1
14
763
MAYIL-VELAN (MURUGAN) The Golden Spear That Parted the Monsoon Sky #Murugan #AIfilm seedance 2.5 on @higgsfield
3
1
12
928
Three Blessings Renewal, Wisdom, Prosperity. Happy Diwali, everyone 🪷✨ MAYURA-VAHINI.. Golden Cloud Sea & Giant Peacock GAJALAKSHMI Starlit Lotus Ocean & White Elephants HIMAGIRI-GANGA ..Crystal Hamsa & Glacier-to-Gold Spring a short Diwali made with AI on @dreamina_ai #DreaminaCPP
8
4
23
1,128
X has become so simple now 😊 find a video from douyin post on x and say this ai video is an absolute masterpiece 🔥🤯
15
2
1
32
1,231
Fire! Fire! Here he comes! Thalapathee🔥 Happy Diwali! (Fan-made teaser) kollywood x hollywood remix 😊
Baahubali 3 trailer leaked? 👀 One warrior. One sword. One creature the size of a temple. Kidding. It's 15 seconds of AI made scene
2
1
1
10
645
Baahubali 3 trailer leaked? 👀 One warrior. One sword. One creature the size of a temple. Kidding. It's 15 seconds of AI made scene
5
2
1
17
1,698
KODANDA The Arrow That Parted the Ocean 🏹🌊 Created with seedance 2.5 on score with suno @dreamina_ai #DreaminaCPP
Where the mortal ends, the god begins 👑🔥 30 seconds. Character Reference to Cosmic Vishwarupa scale in one uninterrupted shot. Turn your sound UP 🔊 hanuman mythology concept art motiondesign #animation #indianmythology on @dreamina_ai #DreaminaCPP
2
15
546