Creativity AI #41: The great AI assistant divide: ChatGPT goes broad while Claude goes deep
Plus Midjourney V7.1 updates, ElevenLabs Studio 3.0, and that hilarious talking rat from Kling AI
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's latest Office Hours (Sep 17th, 2025) kept things realistic about timelines. V7.1 is still in development and is expected to provide clarity within 1-2 weeks, so don't hold your breath for an immediate release. It might serve as a bridge, integrating V8 features into V7's framework.
Other things:
The Style Explorer doubled its style options for more variety.
No video features until after V8, though.
The team's focusing on what matters: style exploration, personalization, and that advanced creation mode.
Improved profiles are actively being developed to integrate with custom style creation.
V8 promises better personalization upper bounds and overall performance. It feels like they're taking their time to get it right, rather than rushing out half-baked features.
ChatGPT vs Claude usage patterns reveal fascinating differences. OpenAI's report shows ChatGPT dominates in practical guidance, information seeking, and writing, with coding at just 4.2% of messages. Meanwhile, Anthropic's data shows Claude crushing it in coding (36% of usage!), plus growing educational and scientific tasks.
My experience matches this perfectly. ChatGPT seems to slack off during long sessions, particularly with lengthy outputs, likely saving computational energy. You can fix it with prompting, but you'll need to monitor it closely. Claude has issues too, but sudden quality drops aren't as common. ChatGPT is becoming the general-purpose AI assistant, while Claude is the specialized technical powerhouse.
What's your go-to model? Please share with me in the comments.
Claude's browser extension for Chrome entered research preview for 10,000 Max users. This experimental feature lets Claude see what you see and take actions directly in your browser.
Capabilities include:
Real-time page summarization and explanation
Automating calendar management and expense reports
Clicking buttons and filling forms
Slash commands for saving common workflows
It won't work perfectly on all sites yet, that's why it's experimental. But imagine having Claude handle tedious browser tasks while you focus on what matters. If you're on Max plan, join the waitlist and help shape this feature.
Perplexity now connects to your apps for searching across private sources. Not exactly cutting-edge, as other LLMs have had this feature for a while, but it's useful when you want searches to consider your own resources.
Hopefully, it provides a foundation for Perplexity to develop more features, such as automated export search reports to Google Drive and other enhancements for general users.
ElevenLabs Studio 3.0 dropped massive upgrades. Full video support in their AI audio editor. You can now generate voiceovers, music, and sound effects directly in the editor.
The AI Script Generator creates story scripts from prompts in seconds.
Speech Correction lets you fix mistakes by editing text with no re-recording needed.
Voice Isolator cleans background noise like magic.
One-click captions boost accessibility.
Multiplayer editing lets teammates collaborate with time-stamped feedback.
Finally, you don't need to export everything and stitch it together elsewhere. Whether you're creating videos, podcasts, or audiobooks, everything's in one place.
Link to video
Kling AI's Avatar feature lets you upload an image and audio to bring characters to life. The movement's limited, but it's ridiculously easy to use.
I generated a video below. See how the rat mimicking speech, it's hilarious! Sometimes, simple features that just work beat complex ones that frustrate. Perfect for quick character animations, although the character motion is quite limited, and you can't add speech to a video clip yet.
Link to video
Update from Geeky Curiosity
This week's AI updates came fast and furious. Don't stress about trying everything. Just pick what looks useful for your actual work and experiment. Have fun playing with the new tools,
Link to music
Refreshed Midjourney Style Resources
The Midjourney Style Resources #19 has been updated to V7 aesthetics.
Inspiration
Scale distortion works incredibly well for grabbing attention. I've noticed our community has a soft spot for miniature worlds: there's something irresistible about tiny detailed scenes. Here are some experiments I've been working on.
tiny ballerinas dancing on a spinning vinyl record, motion blur and spotlighting, deep blacks with delicate pastels, surreal elegance --ar 16:9 --profile uw4q9o7 --v 7
a miniature man grappling with a full-size domestic cat, intense action pose, high detail fur texture, soft living room background, muted tones with sharp contrast --ar 16:9 --profile uw4q9o7 --v 7
toy-sized police cars in high-speed chase over a giant open hand, dramatic lighting, high-speed motion blur, gritty realism with a surreal twist --ar 16:9 --profile uw4q9o7 --v 7
a porcelain teacup used as a luxurious bathtub, human lounging inside with steam rising, golden faucet, ornate tiles, ambient candlelight, pastel and gold hues --ar 16:9 --profile uw4q9o7 --v 7
a giant open book transformed into a cityscape, pages standing like skyscrapers, tiny people walking its spine, warm sepia tones, soft ambient lighting --ar 16:9 --profile uw4q9o7 --v 7
Recent articles
Related articles
Free Geeky Animals' Sref code: 2659100389
Prompts are created using MJ Simple Prompt Generator (+ my brief idea)
. --sref 2659100389 --profile uw4q9o7 --sv 6 --v 7
. --sref 2659100389 --profile uw4q9o7 --sv 6 --v 7
. --sref 2659100389 --profile uw4q9o7 --sv 6 --v 7
a magician Queen and her cat uses magic to fight off the alien monster's attack --ar 16:9 --sref 2659100389 --profile uw4q9o7 --sv 6 --v 7
Cover prompt: a surreal forest glade with a massive apple half-buried in moss, woman gazing up at it, early morning mist, rich greens and deep reds, fantasy lighting --ar 16:9 --profile uw4q9o7 --v 7
I hope you like this article!
Thank you for reading and happy creating!