Another Massive Week in AI! (Summed Up in 10-Min)

TL;DR

The week’s biggest AI announcements included Google Bard access, Nvidia Foundations, Adobe Firefly beta, Bing Chat image generation, Opera’s built-in AI tools, Microsoft Loop, 10 new Canva features, GitHub Copilot X, Ubisoft’s dialogue generator, Unreal Engine MetaHuman, and ChatGPT plugins. Bard arrived within 48 hours for the presenter but felt underwhelming, while ChatGPT plugins stood out as game-changing. Read on for what each release does and why it matters.

Transcript

the pace at which new AI announcements have been coming out has been absolutely insane lately in this video I'm going to try to quickly give you the tldr of the week and try to break down all of the crazy announcements that have come out of the AI space now most of these I've made individual videos about so I'm not going to try to go too deep into ... Read More

Key Insights

• 🔍 Google's new AI tool, Bard, is underwhelming compared to other tools like chat GPT and Bing chat, as it still has a long way to go in terms of capabilities and user experience.
• 💻 NVIDIA's new foundations allow companies to create their own proprietary large language models using NVIDIA's Cloud computers, removing the need for supercomputers to run AI chat bots or other models.
• 🖼️ Adobe's Firefly beta, which is purely trained on licensed images from Adobe Stock and Creative Commons, brings a user-friendly interface to image generation, although there are still ethical concerns regarding the use of images without consent.
• 🔍 Bing chat's new image generation feature, powered by an improved version of DALL-E, generates impressive images, raising questions about whether this is the next-gen version of DALL-E.
• 🌐 Opera now integrates AI features, such as Chat GPT and Rate Sonic, into its browser, allowing users to summarize, rewrite, and creatively modify text directly within the browser.
• 💼 Microsoft Loop, an AI-powered version of Notion, integrates AI features and connects seamlessly with other Microsoft tools like Excel, Word, and PowerPoint.
• 🎨 Canva adds a range of AI features, including Magic Designs, automatic translation, and Magic Slide Presentations, making design creation more accessible and intuitive for users.
• 💻 GitHub's Copilot X, powered by GPT-4, is an AI tool that helps coders by offering code suggestions, fixes, and a voice interface for a more seamless coding experience.
• 🎮 Ubisoft's new AI tool allows game developers to automate dialogue creation for non-playable characters, reducing the time and effort required for scriptwriting and game development.
• 📱 Unreal Engine showcases the impressive capabilities of MetaHuman, using a basic iPhone to record expressive and realistic facial movements, which can then be applied to different characters, raising the bar for realism in game development.
• 🔌 OpenAI's announcement of plug-ins for Chat GPT unlocks new functionality, allowing integration with existing tools, enhanced internet search capabilities, and the ability to manipulate various file types within the interface, opening up new possibilities for users and third-party developers.

Questions & Answers

Q: What were the biggest AI announcements of the week?

The announcements covered Google Bard, Nvidia Foundations, Adobe Firefly beta, Bing Chat image generation, Opera’s AI features, Microsoft Loop, Canva’s 10 new additions, GitHub Copilot X, Ubisoft’s dialogue generator, Unreal Engine MetaHuman, and ChatGPT plugins. The presenter characterized ChatGPT plugins as a game-changing announcement because they extend ChatGPT through integrations with other tools and platforms.

Q: How did Google Bard compare with ChatGPT and Bing Chat?

The presenter received access to Bard within 48 hours of joining the waitlist but found it underwhelming compared with ChatGPT and Bing Chat. Bard even said that Bing had a better user experience and would probably provide better information when asked about its advantages over Bing.

Q: What does Nvidia Foundations enable companies to do?

Nvidia Foundations lets companies create proprietary large language models using Nvidia’s cloud computers and cloud GPUs. A business could build a chatbot trained on its own information, create its own version of Stable Diffusion, or develop biomedical language models without running its own supercomputer.

Q: What is Adobe Firefly beta, and what ethical concerns does it raise?

Adobe Firefly beta is an AI image-generation model trained on Adobe Stock and Creative Commons images. Although those images are licensed, the presenter notes that some Adobe Stock contributors may not have knowingly agreed to AI training and that some stock images were themselves generated by tools trained on outside images.

Q: How does image generation work inside Bing Chat?

Bing Chat can generate images directly from a written request, such as asking it to draw an astronaut riding a horse on the moon. The feature uses a more advanced version of DALL-E, and its results led people to wonder whether Bing was showing a next-generation version of the model.

Q: What AI features did Canva add?

Canva unveiled 10 new additions, largely consisting of AI features. Examples include Magic Designs, translation, Magic Eraser, Beat Sync, topic-based Magic Slide Presentations, Magic Write across Canva tools, and updated text-to-image generation for use in Canva projects.

Q: How does GitHub Copilot X use GPT-4 to help coders?

GitHub Copilot X is a code helper powered by GPT-4. It offers code suggestions and fixes, helps users solve coding problems, and includes a voice interface for communicating with the assistant.

Q: What do ChatGPT plugins add to ChatGPT?

ChatGPT plugins extend the system by connecting it with existing tools and platforms. The described capabilities include enhanced internet search and manipulating different file types within the interface, creating new possibilities for users and third-party developers.

Summary & Key Takeaways

• Google opened the waitlist for Bard, but it fell short of expectations compared to other AI tools like Chat GPT and Bing chat.

• Nvidia announced Nvidia Foundations, enabling companies to build their own AI models on Nvidia's Cloud GPUs, including language models and biomedical models.

• Adobe introduced Firefly beta, an AI image generation model trained on licensed images, raising ethical concerns.

• Microsoft added image generation capabilities using Dolly in Bing chat, potentially showcasing a new version of Dolly.

• Opera now integrates Chat GPT and Rate Sonic into its browser, making AI features accessible for text summarization and creative writing.

• Microsoft Loop, akin to Notion but AI-powered, connects with Excel, Word, PowerPoint, and other Microsoft tools.

• Canva launched ten new AI features, including magic designs, translation, eraser, beatsync, magic slide presentations, and text-to-image enhancements.

• GitHub unveiled Copilot X, a code helper implementing GPT4, offering voice interface for assistance in coding.

• Ubisoft introduced an AI tool to automatically generate dialogue for non-playable game characters, reducing the time needed for script writing.

• Unreal Engine showcased the new version of metahuman, allowing highly realistic facial animations captured with just an iPhone.

• Chat GPT now supports plugins, enabling extended functionality by integrating with other tools and platforms, a game-changing announcement.