_{Welcome back for the Latest AI Drops!}

Today’s Drops:

• Google releases AI model Gemini

• What’s New Across Meta’s AI Experiences

• McDonald’s to use Google AI

• Hottest AI Startup “Pimento”

• Trending on X “Adobe Firefly’s Transitioning Prompt”

• Trending GitHub Projects

• Latest AI Tools

Read Time: 4.5 minutes

Google (DeepMind) releases AI model Gemini

Google has introduced its advanced Gemini AI, which claims to outperform OpenAI's GPT-4 and human experts in various knowledge and problem-solving tasks across 57 subjects, including math, physics, history, law, medicine, and ethics. Gemini is a multimodal AI capable of understanding images, video, audio, text, and code. It surpasses human experts with a 90.0% score on the MMLU test, while GPT-4 scores 86.4%. This multimodality allows it to understand and retain the nuances of various media types.

Gemini has the potential to revolutionize various fields, including science, by generating code and interpreting scientific studies. It's proficient in multiple programming languages and can even create websites that adapt to users' needs in real time. Google plans to integrate Gemini into its devices, starting with the next Pixel phones, and expand its capabilities into touch and tactile feedback.

Additionally, Gemini has been used in AlphaCode 2, where it creates code snippets and ranks them for correctness, beating 87% of human participants in a coding competition. While this technology may not be immediately available to the public due to its computing power requirements, it hints at the potential for future advancements.

Google intends to release Gemini in three model sizes: Gemini Nano for mobile devices, Gemini Pro, and Gemini Ultra, which outperforms GPT-4 on various benchmark tests. Gemini Pro is already accessible for free through the Google Bard service and will continue to gain new capabilities. Google plans to integrate Gemini into many of its products in the future.

Gemini is being introduced as an upgrade to Google's chatbot, Bard, initially in over 170 countries, excluding the UK and EU as regulatory clearance is sought.. The more advanced "Bard Advanced" version will arrive next year. Gemini will initially operate in English but aims to diversify into other languages.Gemini is not only powerful but also efficient, running faster and more cost-effectively on Google's Tensor Processing Units (TPUs) compared to previous models.

Gemini’s launch video

Google Gemini vs OpenAI GPT-4, side by side comparison:

Source

What’s New Across Meta’s AI Experiences

Meta is gearing up to close the year with exciting updates and innovations across its AI-powered experiences on Facebook, Instagram, Messenger, and WhatsApp. Here's a concise overview of what's new:

• Meta AI Evolution: Meta AI, the virtual assistant, is becoming even more helpful with improved responses, greater accuracy in search results, and an expanded range of capabilities. Users can access Meta AI by starting a new message, typing "@MetaAI" in a group chat, or using voice commands with Ray-Ban Meta smart glasses.

• Meta AI Integration: Beyond chats, Meta AI is now behind the scenes, enhancing product experiences on Facebook and Instagram. It generates post comments, chat topic suggestions, search results, and more. It powers a new standalone creative experience called "imagine with Meta AI."

• Reimagine in Group Chats: Meta AI introduces "reimagine" in Messenger and Instagram group chats, enabling collaborative image creation. Users can generate an initial image, and friends can modify it with text prompts, sparking creative exchanges.

• Reels in Meta AI: Reels, a feature for discovering visual content, is coming to Meta AI chats. Users can request recommendations and share Reels to make decisions, such as planning a trip.

• Enhanced Facebook Experience: With Meta AI integrated into Facebook, users can create birthday greetings, edit posts, and receive assistance in various tasks, including setting up groups and finding products. It also helps convert images from landscape to portrait for easier sharing to Stories.

• AI for Creators: Creators can benefit from AI tools like suggested replies in DMs, making audience engagement faster and more efficient.

• Imagine with Meta AI: The text-to-image generation feature called "imagine" is expanding outside of chats and is available online for creative hobbyists.

• AI Improvements: Meta is improving its other AIs, adding more search capabilities and experimenting with long-term memory to enable continuing conversations with select AIs.

• Transparency and Safety: Meta is committed to transparency and safety, introducing invisible watermarking to AI-generated images to distinguish them from human-generated content. They are also investing in red teaming to enhance AI safety.

Meta aims to enhance AI-driven experiences across its platforms, striving to deliver more personalized, immersive, and interactive applications for users.

Source

McDonald’s to use Google AI

McDonald's is partnering with Google to introduce generative AI technology in its operations, starting in 2024. This initiative will involve hardware and software upgrades in thousands of stores, along with improvements to ordering kiosks and the mobile app. The goal is to leverage generative AI to optimize operations, potentially resulting in hotter and fresher food for customers.The specific applications of AI are not detailed, but it will focus on reducing business disruptions and enhancing the overall customer experience.

Source

Hottest AI Startup!

Pimento

Pimento, a French startup, has secured $3.2 million in funding to develop its generative AI tool for creative teams. It helps teams with ideation and moodboarding by generating images, text, and colors based on project briefs. Pimento's personalized approach allows users to save and iterate on its suggestions, distinguishing it from off-the-shelf AI models. The funding will support future feature enhancements as the company aims to assist creative teams more effectively.

So, how do you use the tool exactly?

• You first start by typing some instructions of what you want to achieve with your project, a text brief. You then add a handful of images that will serve as the basis of your project.

• After that, Pimento uses your instructions with AI models to help you create images, text and colours. There are three buttons on the screen that you can use whenever you want to generate images, text or colours.

• If some of Pimento’s propositions seem attractive, you can save them for later. When you’re done, you can generate a link and share a board with all the images, colours and text you saved.

Source

Trending on X-Adobe Firefly’s Transitioning" Prompt

TIP OF THE DAY: Add "transitioning" into your prompt and follow the formula:

(your original prompt) (shade) (color) transitioning to (shade) (color)

Example prompt: a magical waves transitioning from red to the orange

Try it here, it’s free.

Welcome to the innovative world of GitHub projects!

ChatGLM-6B is a language model optimized for Chinese QA and dialogue with 6.2 billion parameters, offering efficient deployment and customization options for various applications.

ModelScope-Agent a general and customizable agent framework for real-world applications, based on open-source LLMs as controllers, offering comprehensive customization, diverse APIs, and practical utility for various tasks

Evals is a framework for evaluating LLMs (large language models) or systems built using LLMs as components. It also includes an open-source registry of challenging evals.

DB-GPT is an open-source framework that simplifies the development of database-related applications using LLMs, offering various technical capabilities and enabling customized application creation with minimal coding in the Data 3.0 era.

Canopy is an open-source RAG framework for building chat applications with Pinecone, streamlining the process with features like query optimization and context retrieval. It offers deployment options and a CLI chat tool for evaluation.

Latest AI Tools

Markprompt helps companies automate customer support, scale without increasing headcount, and deliver exceptional user experiences.

Kommunicate helps easily deploy generative AI bots on any platform.

Superpowered AI is an end-to-end knowledge retrieval solution that makes it easy to build production-ready LLM applications with access to external knowledge.

Read AI's Large meeting models (LMMs) applied to meetings generates a personalized podcast that highlights the past 24 hours and prepares you for your upcoming meetings, perfect for your commute into the office.

Gondolin is an AI-driven app that enhances focus by blocking unrelated web content based on your specified task.