Creativity AI #31: What's happening in the world of AI Art?
News about AI art, Changes in Geeky Curiosity, Prompt Play, Inspiration
Creativity AI is a free weekly publication that brings together interesting articles and recent developments in AI art and creativity.
You can also read this newsletter on the website.
News about AI art
Midjourney Office Hours
What's launching soon
The Moodboard feature got a quiet upgrade without an official announcement. Its performance improvements are quite significant compared to before
Midjourney is considering a dedicated moodboard tab for easier discovery (but still undecided)
Style Explorer (simple version) 1 targets next release with enhanced style control options
Medium-term video and feature upgrades
Start and end frames plus looping functionality finished development, now in testing phase
Video turbo modes will cut generation time from 60-68 seconds down to 20 seconds for professional users
Midjourney video may get an upgrade from 480p to 720p resolution (likely mega plan exclusive). Commentary: This could be the first real benefit for mega subscribers since they currently get the same privileges as pro users, just more Fast hours
Text-to-video options are being explored as an alternative to the current image-to-video workflow
Audio layer development for non-speech sound effects under consideration
Video draft modes and upgraded community features are still in the discussion phase
Long-term infrastructure and pricing changes
Server migration underway involving tens of thousands of machines, taking about 30 days
Migration causes short-term research slowdowns but promises faster performance and lower costs
Video pricing is expected to drop while keeping $60 relaxed mode permanent
Midjourney Pro plan is the minimum required for video relaxed mode access
The V8 model was delayed due to server migration, roughly one month after the migration completes
Midjourney promises bigger releases once infrastructure work finishes
Other development progress
API demand surveys are planned to understand commercial usage patterns and user segment needs
Several secret projects in development with potential demos coming in upcoming months
Hugging Face just dropped something exciting: the Reachy Mini desktop robot for $299 (or $499 with WiFi). This little guy could democratize robotics, letting everyday people build their own helpful droids at home without needing a fancy lab setup.
Whether this sparks a real robotics revolution or just creates temporary buzz remains to be seen. But the possibilities are genuinely exciting.
This got me thinking about design possibilities. If you could create your own desktop robot, what would it look like? Personally, I'd go full nostalgic and aim for something Doraemon-inspired, who wouldn't want a helpful blue robot companion with magical gadgets?
I've been experimenting with Vidu's updated Q1 model, and while the concept sounds amazing on paper, the reality is a bit different. They now support up to seven reference images for creating videos with multiple characters and consistent object placement. Cool idea, right? Unfortunately, when I tested it out, subjects kept appearing and disappearing randomly throughout the video. It feels like they're onto something big with this "Reference to Video" feature, but it's clearly still in the early development stages.
Google's Veo 3 just became available in over 159 countries, but only if you have a Gemini AI Pro subscription. You can now create videos using Midjourney images and text prompts, with the addition of voice and sound effects generation. It's definitely one of the more advanced AI video models out there, but that premium pricing makes it clear this isn't exactly built for casual experimentation yet.
Google's Gemini has been making waves lately, but not necessarily in a good way. It can now access other Android apps automatically, including your private WhatsApp messages, unless you manually disable it. That's a privacy red flag many people aren't comfortable with.
Kling AI introduced a new Jelly Jelly effect that is currently trending. Convert any image to jelly form, then animate it. Everything is pre-set. Simply click and click to generate it. I turned my fish-like alien into jelly and used the pre-set to create the video.
YouTube has just introduced some significant changes to its Partner Program, which have caught many creators off guard. Starting July 15th, they're cracking down hard on what they call "mass-produced" and "repetitious" content, which basically means AI-generated videos that don't add much human creativity. It's probably their strongest move yet against low-effort AI content.
Update from Geeky Curiosity
Revised AI toolbox & buyer’s guide
I've updated my AI toolbox. I no longer recommend the unlimited plan for Runway because Midjourney video is good enough to take over. For me, Runway's use case is now confined to Gen-4 references to make up for Midjourney's Omni-reference limitations.
I also added the “Buyer’s guide” sections. Hope that’s helpful.
Refreshed Midjourney Style Resources
The Midjourney Style Resources #10 has been updated to V7 aesthetics.
Take a survey
Your curious mind wanted. Help shape Geeky Curiosity.
I'd love to hear what's sparking your curiosity these days and how I can make this newsletter even better for you.
Your thoughts help shape every article I write, so please take a moment to share what's on your mind.
Click here to share your thoughts →
Prompt Play
Which AI actually creates perfect seamless tiles?
Seamless tiles are incredibly useful for desktop wallpapers, banners, and any design where you need a pattern to repeat infinitely without visible breaks. The edges must match perfectly when tiled together.
I tested three popular AI generators to see which one actually delivers seamless results.
The testing method
I used an online seamless tile checker tool to verify whether each generated image truly tiles without visible seams. This removes any guesswork and shows exactly how the patterns connect.
Midjourney: The clear winner
Result: Excellent seamless tiles with stunning aesthetics
Midjourney consistently produces beautiful, colorful textures that pass the seamless tile test with flying colors. The results appear vivid and professional, and you have full control over parameters and personalization options.
Verdict: Works perfectly
vibrant oil emulsions with rainbow hues in fluid, looping shapes, seamless abstract design, wet reflective surface --ar 16:9 --tile --profile oaefodl --v 7
A screenshot from the seamless tile checker. It shows the tile is seamless.
Ideogram: Functional but limited
Result: Creates working seamless tiles but with major restrictions
Ideogram can generate proper seamless tiles that pass the checker test. The bad news? You're stuck using Ideogram 1.0 only - the newer models don't support tile generation.
Quality assessment: Less aesthetically pleasing than Midjourney, but functional.
vibrant oil emulsions with rainbow hues in fluid, looping shapes, seamless abstract design, wet reflective surface
ChatGPT-4o: Complete failure
Result: Doesn't create true seamless tiles
ChatGPT's image generation falls short on two things:
Poor overall image quality compared to other AI generators
Complete failure on the seamless tile checker (tested multiple times)
I repeated the test with different prompts, but ChatGPT consistently failed to create properly matching edges.
Create a seamless tile vibrant oil emulsions with rainbow hues in fluid, looping shapes, seamless abstract design, wet reflective surface at aspect ratio 16:9 
ChatGPT-4o failed the seamless test.
My take
For seamless pattern creation:
First choice: Midjourney - superior quality and reliability
Backup option: Ideogram 1.0 - works but has limited features
Skip entirely: ChatGPT - doesn't work for this task
Inspiration
This week, I'm inspired to use more "negative spaces" in my image compositions instead of trying to fill it with too many stuff! (My real-life is all about how many things I can fit into my already overcrowded calendar...).
Less could be more. And possibly more beautiful and productive!
a solitary tree atop a vast sand dune, minimal horizon line, expansive sky filled with soft gradients, high contrast between subject and emptiness, delicate wind trails in the sand --ar 16:9 --profile oaefodl --v 7
a gothic cathedral floating in black space, illuminated by ambient celestial light, intricate stone detailing, surrounding void emphasizes scale, surreal isolation and silence --ar 16:9 --profile oaefodl --v 7
a small candle burning on a stone altar inside a ruined gothic cathedral, collapsed walls open to a dark sky, shadows swallow architecture, silence emphasized by space --ar 16:9 --profile oaefodl --v 7
an astronaut floating above a barren alien planet, illuminated only by a distant star, vast negative space of dark cosmos, subject positioned off-center, atmosphere of solitude --ar 16:9 --profile oaefodl --v 7
a thin rope bridge spanning across a deep, mist-filled chasm, bridge fading into fog, minimal background, stark contrast with negative space, cinematic lighting --ar 16:9 --profile oaefodl --v 7
a lone ballerina frozen mid-leap on an immense, empty stage, single spotlight above, deep surrounding shadows, negative space enhances grace and isolation --ar 16:9 --profile oaefodl --v 7
a single hot air balloon floating in a vast, clear blue sky, no clouds or landscape distractions, small silhouette casting shadow far below, peaceful and expansive scene --ar 16:9 --profile oaefodl --v 7
Recent articles
Related articles
Free Geeky Animals' Sref code: 1989527900
Prompts are created using MJ Simple Prompt Generator (+ my brief idea)
Cover prompt: A solitary wanderer holds a flickering lantern in a crumbling city of shifting geometry, building shapes bend and twist subtly, strange whispers echo through narrow alleyways --ar 16:9 --sref 1989527900 --profile i6fyyx5 --sv 4 --v 7
I hope you like this article!
Thank you for reading and happy creating!