😉
I saw a famous tech blog post a month ago that shocked me, it claimed DALL·E 2 is still one of the best AI image tools in 2026. Clearly, the article was generated with outdated information and a not-so-smart agent. 😂There’s nothing wrong with AI-generated blogs or articles, as long as the information is accurate, especially in the fast-moving world of AI.
That’s why I’m very proud of this publication. ✨Every issue is personally tested by me, with real hands-on experience, honest insights, and practical workflows I actually use. I’m truly grateful to all of you who keep checking in and reading. Even though I’m currently swamped with client work, I promise to keep sending out these newsletters. Thank you again for joining me on this journey!💜
I built my perfect purple smart-casual wardrobe with AI in Adobe Firefly Boards.
From cozy daytime home looks to beautiful nighttime outfits, I created everything using the “Make a fashion design moodboard” Workflow Quick Guide.
It is so easy. Check out my workflow.
Big thanks to Adobe for sponsoring this post 💜
Interested in reaching a community of AI-forward creators?
You can find all the details on sponsoring a future issue right here.
I’m presenting at Substack’s first AI Summit ✨
I wanted you to hear this from me first. 😉On March 27, I’ll be presenting at Cozora’s AI Summit — the first-ever AI summit held on Substack. Eight hours of live sessions with some of the sharpest AI creators and authors on the platform.
We’re covering the real stuff: how AI is changing work, creativity, strategy, and leadership — with live Q&A so you can ask anything.
This is free and streaming live on Substack. I’d love to see you there.🤗
Suno v5.5 introduces powerful new personalization features, including voice cloning that lets Pro and Premier users verify and use their own singing voice for generations, custom models trained on up to three of their personal tracks to capture unique styles, and the “My Taste” system that learns individual genre and mood preferences over time. These tools build on the core v5.5 model to make music creation far more intuitive and user-specific, with enhanced prompt adherence for complex ideas and seamless integration of personal audio elements.
As a result, the music achieves noticeably higher quality through greater expressiveness in vocals and instruments, more natural phrasing, dynamic emotional depth, and outputs that feel unmistakably authentic and professional rather than generic. Check out one of my trials with v5.5, this is perfect for you to listen to while reading this issue. 😉👇Uni-1 is Luma AI’s new unified model that combines understanding and generation in one autoregressive transformer. It interleaves text and visuals, reasons internally, and produces more coherent, intent-driven outputs than traditional diffusion tools. I tested it on a moody A24-style forest reveal by comparing with Midjourney v8, and the frame-to-frame consistency is impressive, it finally feels like the whole sequence was planned as one unified piece.
Finally, Seedance 2.0 is officially launched on Dreamina AI and CapCut! As their creative partner, I’ve been testing and using the model since February under NDA, so I couldn’t share any details or outputs until now. Here’s my very first generation with Seedance 2.0 back in February, no cherry-picking, just one image uploaded + a simple prompt. The result? A super cinematic cut of a boy riding his bike toward a massive volcano, with dramatic motion and sweeping camera work. I was genuinely impressed by how much life and energy the model injected on the very first try. What blew me away even in those early tests was how Seedance 2.0 turns a single reference image into something far more dynamic and controllable than most video models at the time.
Why Seedance 2.0 stands out (key features I’ve been loving):
True Multimodal Power: Combine text prompts with up to 9 images, 3 short videos (max 15s total), and 3 audio clips in one project. Use the handy reference system to precisely tell the model what each asset controls — character/style from images, motion/camera from videos, rhythm/mood from audio.
Director-Level Control: Full command over performance, lighting, shadows, camera movements, and scene composition. It feels like having a virtual cinematographer that understands your vision.
Exceptional Consistency & Motion Stability: Characters, objects, and style stay coherent across shots. Motions feel natural with realistic physics, great for action, travel, or narrative sequences like my volcano ride.
Multi-Shot Storytelling: Seamless transitions and narrative flow, perfect for short cinematic clips (up to 15s) that feel polished and professional.
Hailuo AI Light Studio is straight-up unreal! Super easy: pick from tons of cinematic templates, drag lights around to adjust angle, intensity, and color temperature, then generate. You can layer multiple lights and apply 20 pro presets with one click.It turns basic AI shots into moody, cinematic gold without any gear or hassle. Check out my demo here. 👇
I got early access to Pika AI Selves and have been using it to auto-post on Instagram and X since last month. I also use it to chat with my friends and family on WhatsApp. Pika Selves can generate images and videos for social media, create reports and drop them straight into Slack for review, and even engage with people in Discord communities. I’ve used it to generate a full website too! You can connect it with the apps and tools you already use, and let your AI Self work for you. They just opened it to the public, go try it now at pika.me! Set up is super easy, just follow the easy steps on the website.
Pika@pika_labs
WE INTERRUPT YOUR REGULAR PROGRAMMING to say that Pika AI Selves are officially in public beta! Everyone can now give birth to an agentic extension of themselves at Pika dot me, or on our new iOS app.
5:38 PM · Mar 26, 2026 · 14.2K Views
25 Replies · 31 Reposts · 129 Likes
Midjourney v8 brings a fresh aesthetic with enhanced detail, coherence, and a more refined visual style. Especially when blending multiple --sref (style references), it creates beautifully harmonious and artistic results that feel truly elevated compared to previous versions.I know a lot of people are complaining about v8, but personally, I love it. 💜 The nicest aesthetics come from cleverly blending sref codes, though to get the absolute best and most consistent results, I always pair them with an image reference to better anchor the style. It’s also noticeably faster, with significantly better prompt adherence and much improved text generation. Try it here on their alpha page.
Pruna AI launched P-Video, delivering incredibly fast and affordable video generation with high speed: it creates 5 seconds of 720p video in just 10 seconds and 1080p in 18 seconds. It also includes a fast draft mode that generates a 2.5-second preview for a 5-second 720p video, along with strong creative control and multi-modal input support (text, image, or audio) for up to 15-second clips with native audio output. Check out my video created with an image and an audio file uploaded as reference.👇(super cute!)
ElevenLabs just launched their Music Marketplace in ElevenCreative, which means we can now easily generate our own AI music tracks, publish them directly on the platform, and actually make money every time someone downloads or remixes them! This new feature lets us create, share, and earn royalties from AI-generated music without dealing with complicated licensing, sync fees, or negotiations. Are you selling your AI music?
I love to hear from you! Feel free to reply or leave a comment here. 💜
Today I am sharing some of my Seedance 2.0 Prompts here. Enjoy!
Prompt: high-tension survival fight, fast action cuts, maintain female fighter’s exact face body clothing consistency, derelict warehouse flickering lights sparks blood sweat realistic physics.
[0-3s] Wide shot chaos: she parries knife thrust with machetes clash sparks explode, attacker snarls push back, handheld shake brutal force.
[cut] [3-6s] Close-up her gritted teeth sweat flies as she headbutts attacker nose crunch blood spray, quick zoom her fierce eyes.
I am using Seedance 2.0 Omni here.
Prompt: Premium compact espresso machine (upload the image of the coffee machine) in sophisticated matte black and polished chrome, every angle catching light with exquisite specular highlights, minimalist luxury design in perfect macro focus, slow floating parallax arc slider gliding around the machine, creating depth and dynamic reflections across its flawless surfaces. Smooth arcing pull-back reveals the machine elegantly positioned on a high-end open-concept kitchen island with marble veining and soft ambient luxury lighting. Cut to poised young woman in the luxurious space, cup in hand, face lighting up with authentic joy and peaceful fulfillment, subtle blissful smile spreading as she takes a slow, contented sip, emotional warmth and sophistication, rich cinematic grade, subtle lens glow, seamless luxury motion.
Prompt: Boots showing realistic leather grain texture, high-shine metallic reflections, natural heel flex during powerful stride. Young stylish east asian woman ascending subway stairs in golden-hour urban station, wearing boots, ultra-detailed serene confident face with natural darker skin glow, subtle smile, realistic hair flow, crisp blazer fabric details. Deep black-gold cinematic grade, high contrast, subtle volumetric lens flares, consistent upper-left key lighting with soft rim shadows. Extreme low-angle Dutch tilt close-up on heel impact, smooth transition to straight upward dolly tracking camera following stride, pan to high-angle reveal of woman striding confidently at top with elegant ease-in/ease-out motion, natural fluid gait. Photorealistic high-fidelity materials, seamless integration, zero artifacts, ultra-premium fashion ad polish
✨Thanks for reading and exploring the world of AI with me!
I’ve spent the last few years watching how tools come and go, but I’ve also seen how creative teams struggle to keep up. Most departments are currently “tool-rich but workflow-poor”, they have the subscriptions, but their process is still slow and expensive.
To help solve this, I am now offering the Creative Team AI Audit.
Instead of just adding more noise, I look at exactly what your team is currently using and where the bottlenecks are. I’ll help you optimize your existing stack, cut out the fluff, and suggest the high-performance pipelines that will actually save your company time and money.
If you want to bring your department’s creative output into 2026 and turn your team into an AI-powered powerhouse, just hit reply to this email to start the conversation.
✨Make some AI magic and tag me on X, Instagram or Pinterest.
If you find value in what I’m building here at PicAisso, please consider supporting my work here. Your support allows me to stay in the trenches testing these tools every day so you don’t have to. You are all amazing. ✨
Have fun creating with AI,
Zeng 💜
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.