Creativity AI #9: Your weekly dose of updates on creativity and productivity
Midjourney office hours, Perplexity's Comet browser, Claude 3.7 Sonnet, Bird watching, ElevenLabs + Spotify
Creativity AI is a new publication that brings together interesting articles and recent developments in the AI world related to creativity and productivity.
Midjourney Office Hours (2025-02-26)
The developers are currently working on improving the selection of the web editor and layers. The improved editor will be available soon.
The big batches feature (more images per grid) may be released after V7, as it requires more R&D.
The Serf codes from V6 will be transferred to V7 and made forward-compatible. However, it may appear slightly different.
The default V7 version is probably three times faster.
There are two V7 versions: the "slow version" improves prompt understanding over V6 but requires more tuning for high resolution and omni-reference. The "real-time" (or "fast") version, which integrates well with omni-reference, will enable the creation of cool interfaces in the coming months.
The Midjourney video will be released by... Midjourney. The (smaller) video model now works well internally and matches the Midjourney aesthetic, as opposed to other generic-looking videos.
According to David: "Video is expensive by default, you gotta fight to make it cheap." So the goal isn't to create a $1,000 video plan, but to make video accessible to all users.
There is still debate over which model to release first: text-to-video (which is already working well) or image-to-video (which presents some challenges).
The 3D feature is delayed.
Midjourney is developing hardware products. Giant structures are being built with "thousands of gallons of stuff" and wired throughout. They plan to build a large number of prototypes in the coming months.
"Midjourney is an imagination tool," said David.
Some people questioned why David was repeating himself during office hours. David commented that as a leader, repeating the message is critical. And, because not everyone attends office hours every week, it is helpful to explain to newcomers what the team is working on.
Perplexity
Perplexity is developing an AI-powered web browser dubbed "Comet" for agentic search. You can join the waitlist at Comet Browser by Perplexity.
The Claudine 3.7 Sonnet is now available on Perplexity. The new model replaces Claude 3.5 Sonnet on the platform.
Perplexity has also released a new voice mode for the iOS app.
Why does this matter?
It is unclear what the agentic search browser will look like. I hope it outperforms current browsers and boosts productivity.
The Claudine 3.7 Sonnet is a pleasant addition. However, it lacks the "extended thinking mode" (described below) found on the Claude Pro plan.
Claude's 3.7 Sonnet
The new model combines the ordinary LLM and a reasoning model. In standard mode, Claude 3.7 Sonnet is an updated version of Claude 3.5 Sonnet. The "extended thinking mode" is only available for the paid Pro plan and causes the bot to think longer before responding. Its self-reflection improves its performance in math, physics, instruction following, coding, and a variety of other business-related tasks. Claude 3.7 Sonnet is now available at Claude.ai, through the Claude API, and on Perplexity.
Why does this matter?
Claude is one of the best-performing AI models for a variety of tasks. Its reasoning model provides another option for conducting a thorough analysis of a topic of interest that does not require Internet access (Depending on your use case, limiting information sources is not always a bad thing.)
Bird watching.. wait a minute... are those birds talking about me?
Scientists are using artificial intelligence to investigate how crows communicate with one another by recording their vocalizations and movements. They discovered that in Northern Spain, where they form family groups, offspring stay with parents for up to four years, as opposed to the typically solitary lifestyle of crows elsewhere.
The scientists categorize thousands of crow sounds in the hopes of one day understanding the meaning of the birds' vocalizations and possibly trying to speak their language.
Watch the video here
ElevenLabs
Spotify now accepts audiobooks created with ElevenLabs. Creators can submit their audiobooks to Findaway Voices (by Spotify), sell them to millions of listeners, and earn money. Read the entire announcement here.
ElevenLabs added a speed slider for its voice generator. Make the AI voices speak faster or slower.
Why does this matter?
If you are a book writer, consider using audiobooks to reach a larger audience. Use the ElevenLabs Studio to record voices for multiple characters. Use special sound effects to bring your story to life. Earn more money using AI voices.
Other news
Ideogram introduced a Team plan for businesses to manage multiple users on the platform. Its collaborative feature is said to be "coming soon."
Runway released Describe Image and Custom Image Styles, which allow users to describe an image and generate images with a similar aesthetic style. It generates a detailed and longer Describe prompt that differs from Midjourney's. The Runway’s Describe prompt has three parts: subject, scene, and style.
Google Veo 2 is now available through Freepik. According to Forbes, Freepik's annual subscription costs $69 and includes up to 84 Veo 2 video generations, which is insufficient if you want to try out and learn Veo 2 seriously. The Veo 2 is also available at Fal.ai. According to its website, the 5s video generation will cost $1.25 for a "limited time only." Every additional second costs $0.25. It's quite expensive and most likely reserved for AI video superfans.
Inspiration
According to David Osborn's article in The Phoblographer, an interesting photograph is one that makes people think. When a photograph presents an aspect that is not what the viewer expects to see, it piques the viewer's interest.
The key is to disrupt human nature's desire "not to think" or prediction of familiarity. Make the viewer do some work. Don't give them what they expect to see.
These are my three attempts at creating unusual kitchen scenes to trigger viewers' curiosity. What are your thoughts? Intrigued?
(1) A woman who puts fireflies and honey into mason jars.
(2) The suspended ice cubes.
(3) A lady is harvesting glowing mushrooms from her kitchen.
Recent articles
Free Geeky Animals' Sref code: 983305922
Cover prompt: A biomechanical woman merging with data streams, her eyes filled with boundless intelligence, fluid cyberpunk aesthetics, glowing golden veins, ultra-detailed textures, surreal futuristic ambiance --ar 16:9 --sref 983305922 --v 6.1
Invitation to publish on Creativity AI
Please contact me if you want to publish on Creativity AI.
This publication will be distributed through Substack and Medium platforms, reaching thousands of readers.
Email: animalsgeeky@gmail.com