AI roundup #10 👩💻
In today's roundup ⬇️
💥 OpenAI does it again: ChatGPT 'plus' plugins
OpenAI has just made a groundbreaking announcement, set to revolutionize the AI landscape.
ChatGPT, the popular language model, will soon gain unprecedented capabilities with the introduction of ChatGPT Plugins.
We’re rolling out web browsing and Plugins to all ChatGPT Plus users over the next week! Moving from alpha to beta, they allow ChatGPT to access the internet and to use 70+ third-party plugins. https://t.co/t4syFUj0fL pic.twitter.com/Mw9FMpKq91
— OpenAI (@OpenAI) May 12, 2023
Exclusive to ChatGPT Plus subscribers, this exciting update grants access to the vast world of the web along with a remarkable collection of over 70 third-party plugins.
The release of ChatGPT Plugins marks a seismic shift in the AI landscape, providing ChatGPT subscribers with unparalleled access to cutting-edge features.
For a monthly subscription of $20, users can unlock the potential of ChatGPT as a web-surfing, plugin-powered powerhouse.
Sources: Venture Beat, Mashable
🚀 Google takes center stage at I/O 2023
On May 10th, Google exceeded expectations at the I/O 2023 conference, leaving attendees in awe with a flurry of groundbreaking AI announcements.
The tech giant solidified its position as a dominant force in the industry, showcasing an impressive array of cutting-edge developments that will shape the future of artificial intelligence.
Pretty sure Google is focusing on AI at this year’s I/O. #GoogleIO pic.twitter.com/RxlFQw2l8b
— The Verge (@verge) May 10, 2023
Here's a quick rundown on the top 5 announcements:
- Google Maps: Introduced "Immersive View for Routes" feature, providing comprehensive information about traffic simulations, bike lanes, parking, and more.
- Magic Editor and Magic Compose: AI-powered photo editing tool that can make complex edits in specific parts of photos and rewrite texts in different styles.
- Bard: Now available in English in over 180 countries, with support for Japanese and Korean. Can generate images and partner with Adobe for art generation.
- Google Workspace: AI enhancements in Sheets for automatic table generation, and in Slides and Meet for image creation and custom backgrounds.
- Pixel 7a: New Pixel phone with improved features and lower price compared to Pixel 7, including enhanced camera capabilities.
1/Today at #GoogleIO we shared the bold and responsible approach we're taking to use AI to reimagine our products, including Search with a new generative search experience coming first to Labs. You can sign up starting today. https://t.co/rjQscvjWzQhttps://t.co/jmzhruXX2C pic.twitter.com/mgFpxpS8dc
— Sundar Pichai (@sundarpichai) May 10, 2023
Sundar Pichai also announced more such innovations that will pave the way for AI in years to come.
Sources: TechCrunch, The Verge
💬 Supercharged chats: Claude hits 75K words
Anthropic's chatbot Claude now boasts an expanded context window of 75,000 words, a major improvement over existing models.
The AI startup tested Claude's capability by processing the entirety of "The Great Gatsby" and accurately identifying a modified sentence in just 22 seconds.
With 100K context windows, you can: Digest, summarize, and explain dense documents like financial statements or business reports.
— Anthropic (@AnthropicAI) May 11, 2023
This advancement signifies increased processing capacity for AI language models, enabling faster reading, summarizing, and analyzing of long documents, while also facilitating longer and more coherent conversations.
⚡Meta unleashes ImageBind: a multisensory experience
Meta has unveiled ImageBind, an open-source AI model that integrates various data streams, including text, audio, visual data, temperature, and movement readings.
This research project hints at a future of immersive generative AI systems that can create multisensory experiences.
Meta's model is the first to combine six types of data into a single embedding space, paving the way for AI systems to cross-reference diverse sensory inputs.
6. Meta’s open-sourced multisensory model
— Rowan Cheung (@rowancheung) May 12, 2023
Meta is back (again!) with yet another exciting open-source project.
Introducing ImageBind, a new AI research model that understands and combines text, audio, visual, movement, thermal, AND depth data. pic.twitter.com/tNlKB0aQvX
The open-source approach by Meta contrasts with the increasing secrecy of rivals like OpenAI and Google in the AI research landscape.
Sources: The Verge, Interesting Engineering
🎥 Script to screen: stable animation SDK debuts
On 11th May, Stability AI released the Stable Animation SDK, enabling developers and artists to leverage Stable Diffusion models for text-to-animation generation.
With the Stable Animation SDK, we are giving users a new way to unleash their creativity and imagination using generative AI. We are excited to see what amazing things people will create with this tool.
- Emad Mostaque, Stability AI CEO
The SDK offers three input options: text only, text with an initial image, and text with an input video.
5. Stability AI just released Stable Animation
— Barsee 🐶 (@heyBarsee) May 15, 2023
A powerful text-to-animation tool.
You can instantly turn:
- Text to animation
- Image to animation
- Video + Text to animation pic.twitter.com/d9AiLssG5x
Unlike other models, accessing this tool requires advanced technical skills for installation and UI operation.
Costs are credit-based, ranging from 0.058 to 0.174 credits per operation based on video dimensions and 3D render mode.
Sources: ZDNET, Voicebot.AI
Also read:
AI Roundup #9: Last week in AI