- AI Bits
- Posts
- 💻Image generators on the move, AI in Tennis, Deciphering old tablets with AI, and Remodeled.ai!
💻Image generators on the move, AI in Tennis, Deciphering old tablets with AI, and Remodeled.ai!
Catch up with this week's A.I. news and learn about another A.I. tool.
Hey everyone!
Only 4 days left until Virtual Vision, our first AI Art competition!! Don’t miss out on the last few spots sign up below 👇
➡️Sign up here⬅️
Here’s what we’ll learn about today:
Stable Diffusion and Midjourney updates 🎨
AI reached Tennis too, AI commentators in Wimbledon 🎾
AI helps decipher 5000 years old cuneiform tablet 🗿
Remodeled AI - reimagine the interiors and exteriors of your house in 10 seconds!
Reading time: 5 minutes 10 seconds
News Reel:
Highlight 1: Midjourney and SD new features!
For all of our art aficionados! Here we got a summary of all the exciting news on AI art from the last week. Midjourney and Stable Diffusion, our beloved AI art image generators, have just leveled up their game bringing in highly requested features 💡
Midjourney v5.2 → head to Discord and try out the features we will describe!
Outpainting: Say goodbye to canvas limitations! Midjourney's "Zoom Out" feature lets you extend your artwork beyond its original borders. Non-square images? No problem! Midjourney will square them up for you🎨.
Aesthetic System: Prepare for an aesthetic revolution! Midjourney's new system offers improved aesthetics, sharper images, and better text understanding. It's like having your personal art critic cheering you on 🌟.
Prompt Share: Just change out the words in square brackets. Test it out!
Prompt: Full frame recycled mural of a [skull] made from many [flowers, buttons] and found objects, [dark sky blue and gold], recycled material murals, intricate textures, vibrant figuration.
#midjourney52… twitter.com/i/web/status/1…
— Nerbts (@Nerbts)
2:37 AM • Jun 26, 2023
High Variation Mode: Embrace the art of surprise! Midjourney's latest mode injects more variety into your image variations. Say farewell to repetition and hello to endless creative possibilities 🌈.
/shorten Command: No more beating around the bush! Use the /shorten command to analyze your prompt and get rid of unnecessary tokens. It's prompt makeover time📝!
Massively Improved Image Detail: Prepare to be awestruck by the hyper-realistic creations that this model churns out. Whether it's films, television, music, instructional videos, or industrial design, SDXL 0.9 is a game-changer in the realm of AI imagery🚀.
Stunning features in the face of this wolf
— Stability AI (@StabilityAI)
5:16 PM • Jun 22, 2023
ClipDrop Integration & API Excitement: You can access SDXL 0.9 through ClipDrop today, with the API release just around the corner. Unleash your creative potential with ease 🔍!
Research Weights Available Now: For all the AI enthusiasts and researchers out there, you can get your hands on the SDXL 0.9 research weights right away. The open release is scheduled for mid-July, so mark your calendars and get ready to explore the possibilities of this incredible technology 📚!
“XL 0.9 is the strongest multi-aspect model in the open world. It is a massive leap over any previous model when it comes to natural rep of large and small aspect content due to part to the much deeper transformers, new image conditionings, and multi-aspect specific training”
— Emad (@EMostaque)
10:58 PM • Jun 22, 2023
Amazing updates, right? Well, that is not all just yet. Some weeks ago, in our newsletter, we talked about DragGan, the end of Photoshop 😵… and it has been made open source now! You can now access this tool through run it locally through this GitHub repository or use it on HuggingFace. Drag and drop the image as you like to edit it instantly through AI! 💫
These updates are revolutionizing the world of AI art, igniting the imaginations of artists and creators worldwide. Get ready to unleash your inner Picasso and take your artistic journey to extraordinary new heights! 🏔️
Highlight 2: AI is coming to tennis too
Game, set, and AI-powered commentary! 🎾
Wimbledon is embracing artificial intelligence this year with IBM's partnership, bringing fans a new level of engagement. The tournament will feature AI-generated audio commentary and captions in online highlights videos, courtesy of IBM's Watson AI platform.
Here are the key points:
AI commentary: Fans can toggle on and off audio commentary and captions during video highlights on the Wimbledon website and app.
Complementing the human element: IBM sees AI as a complement, not a replacement, for human sports commentary. 🤝
Analyzing draws: AI will analyze shots and performances, providing insights into players' paths to the final beyond their rankings.📈
Tennis embracing technology: Wimbledon follows the trend of tennis integrating technology, such as electronic line calls set for 2025 (If you play tennis, you know the number of debates on an unclear call 🙃).
A smarter fan experience: By leveraging AI, Wimbledon aims to enhance the fan experience and provide unique insights into the tournament.
Wimbledon is serving up an exciting blend of tradition and technology. With AI-powered commentary and analysis, fans can expect a one-of-a-kind experience at this year's tournament. 🎾
If the experience will be as nice as advertised, what does this mean for commentators and journalists?
Read an article from The Guardian here.
Highlight 3: 5000-Year-Old Cuneiforms Tablets were deciphered with AI
In a groundbreaking study published in the journal PNAS Nexus by Oxford University Press, archaeologists, and computer scientists have joined forces to create an AI program that can instantly translate ancient cuneiform tablets 🗿. These tablets hold immense historical significance, documenting the political, social, economic, and scientific history of ancient Mesopotamia ⛏️.
Sample tablet used in the article - copyright The Trustees of the British Museum
Some bits on the research:
Akkadian, an ancient East Semitic language once spoken in Mesopotamia, is preserved on clay tablets dating back to 2500 BC📆.
Most of these documents remain untranslated and inaccessible due to their sheer number and the limited number of experts proficient in reading them📑.
The AI program achieves high accuracy in translating formal Akkadian texts but struggles with more literary and poetic works.
The NMT model is currently accessible through an online notebook and the source code is available on GitHub at Akkademia 🚀
The goal of the neural machine translation (NMT) into English was to create a collaborative pipeline that assists scholars and students in understanding the ancient language. They definitely succeded and it’s promising to see the combined results of AI and human expertise! 💡
Read the official paper here.
A.I. Showcase: Remodeled.ai
Interior designers are next… remodeled.ai provides quick renders of your house or rooms in any style, material, or color palette of your choice! 🎨 Here’s how it works:
Go to https://remodeled.ai/
Click on Generate your room
3. log in as you prefer
4. Click on ‘Start Remodeling’ at the bottom of the page
5. Upload the image you want (you can simply take an image of your living room from your phone and use that too!)
6. Click on Click to select themes and choose the one you like
7. Click on Click to select colour palette
8. Click on Generate
That’s it! Here are our results!
Original image used
Japanese theme with neutral colors
Japanese theme with neutral colors and mahogany wood as the primary material
Try it out here now! You get 3 free renders before it’s paid!
AI Art of the Week
We’ve got some fun ones for you this week from artist CT_Actual (discord tag: ct_actual)
With a crazy amount of detail and creativity, these images are sure to make your day! He’ll be one of the participants in our upcoming competition as well so stay tuned for more of CT_Actual’s work!
Generated by CT_Actual
Generated by CT_Actual
Generated by CT_Actual
Generated by CT_Actual
Generated by CT_Actual
See you soon!
-AI Bits
Reply