• Hyrise AI
  • Posts
  • 🤖 Two Open Source Rivals to OpenAI's 'Multimodal' GPT-4V

🤖 Two Open Source Rivals to OpenAI's 'Multimodal' GPT-4V

Square Unveils AI-PLUS: Powered Tools for Generating Websites and Restaurant Menus

Author

Hyrise AI
October 19, 2023

Welcome, AI Enthusiasts.

OpenAI's GPT-4V is an advanced AI model that combines text and image comprehension, but it has concerns such as privacy issues and biases..

And

Square, the financial services platform, has launched 10 new generative AI capabilities to enhance customer content creation, onboarding, and setup, aiming to revitalize its platform amid declining revenues from other services and growing competition.

In today’s issue:

  • 🤖 Two Open Source Rivals to OpenAI's 'Multimodal' GPT-4V

  • 🟪 Square Unveils AI-Powered Tools for Generating Websites and Restaurant Menus

  • 🛠️ 3 New AI tools

  • 💻 Custom code in ChatGPT

  • 🤖 3 Quick AI updates

Read time: 3 minutes.

LATEST HIGHLIGHTS

Image source: Leonardo

To recap: OpenAI's GPT-4V is a powerful multimodal AI model with privacy and bias issues, while open source projects have introduced LLaVA-1.5 and Fuyu-8B as alternatives, each with its strengths and weaknesses, sparking concerns about misuse and bias in the AI community.

The details:

  •  OpenAI's GPT-4V is a potent multimodal AI model, capable of comprehending both text and images, but it raises concerns about privacy and bias.

  • Open source projects like LLaVA-1.5 and Fuyu-8B provide alternative models, with LLaVA-1.5 excelling in object detection and meme interpretation, while Fuyu-8B offers the ability to understand unstructured data.

  • The development of these models triggers concerns about their potential misuse and biases, as the AI community leans toward open sourcing models with fewer restrictions for fostering innovation.

Here is the key takeaway: OpenAI's GPT-4V offers powerful multimodal capabilities, it faces privacy and bias issues. Open source alternatives like LLaVA-1.5 and Fuyu-8B aim to address these concerns, but the AI community is grappling with the balance between open sourcing models for innovation and potential risks.

SQUARE
🟪 Square Unveils AI-Powered Tools for Generating Websites and Restaurant Menus

Image source: Leonardo

In Summary: Square has introduced 10 generative AI capabilities to improve customer content creation and reinvigorate its platform amid financial challenges and growing competition. One feature, the Menu Generator, enables quick menu creation for restaurants, although concerns about AI-generated menu accuracy are acknowledged.

Key points:

  •  Square has unveiled 10 new generative AI capabilities to enhance customer content creation, part of its strategy to rejuvenate the platform amid financial challenges and increased competition.

  •  One of these capabilities is the Menu Generator, designed to help restaurants create menus quickly, though concerns about the accuracy of AI-generated menus have been raised.

  •  The AI features also cover email and website copy generation, with the human review process incorporated, and item description generation for seller catalogs.

  •  These AI advancements are part of Square's efforts to leverage technology and stay competitive in the financial services sector, addressing its declining revenues from other services and positioning itself as a technology partner for seamless AI applications.

Our thoughts: As Square ventures into this technology-driven territory, the company appears determined to remain at the forefront of innovation in the financial services sector. How these AI tools perform and adapt in real-world scenarios will undoubtedly be of great interest to industry observers and stakeholders alike. Square's bold move underscores the transformative potential of AI in reshaping the financial landscape.

TRENDING TECHS

🎤 IStory- The first tool for digital content creators that improves audience experience and engagement by integrating rich media with speech to text, text to speech, and a natural language understanding engine, all on a no code platform.

🔮 Sibyl- The world's first spiritual AI for Neophytes, Adepts, and Holistic Practitioners trained on volumes of rare metaphysical knowledge and experience. Freemium Plan, AI-to-AI Conversations, Chat Visualization, Speech Enabled, Multilingual support in 99 Languages.

🪄 Cascadeur- Cascadeur is a software for 3D keyframe animation of humanoid or other characters. AI-assisted tools help you to make key poses really fast, instantly see the physical result and adjust secondary motion. All while retaining complete control at any point.

AI DOJO

Custom ChatGPT
📱 Image Understanding:

    1. To begin, tap the photo button to capture or select an image. If you're using the iOS or Android app, tap the plus button before proceeding.

    2. You also have the option to discuss multiple images or utilize the drawing tool to provide guidance to your AI assistant.

    3. Image understanding is powered by advanced models, specifically multimodal GPT-3.5 and GPT-4. These models combine language reasoning with the analysis of a wide range of images, including photographs, screenshots, and documents containing both text and images.

QUICK BYTES

Almost a decade after the "Twitch Plays Pokémon" phenomenon, where over a million players collectively controlled a pixelated avatar in Pokémon Red, software engineer Peter Whidden has trained an AI using reinforcement learning to play the same classic game. This AI has spent more than 50,000 hours playing Pokémon and its journey is shared in a viral YouTube video. The AI uses a Pavlovian reinforcement model, earning points for leveling up Pokémon, exploring new areas, and winning battles. While the AI's incentives sometimes lead to quirky behaviors, its journey has captured the interest of many, demonstrating the appeal of blending AI and gaming in a creative way.

Amazon has revealed that it will commence testing Agility's bipedal robot, Digit, in its facilities, primarily focusing on tote recycling, with the intent of finding a potential fit for its processes. While this move is part of Amazon's exploration of innovative technologies following its $1 billion "Industrial Innovation" fund, the company's history suggests that actual deployment remains uncertain, despite its extensive robot system usage in warehouses.

Superfy, a mobile app, aims to connect users with real people through AI, facilitating live chats for answers to personal queries, encouraging more subjective and meaningful conversations than AI chatbots. The app uses proprietary AI technology called Matchpoint AI to match users with relevant people based on various factors, including expertise and personal relevance. Superfy boasts 550,000 registered users, primarily Gen Z users in the U.S., Canada, and the U.K., who prefer connecting with real people for their queries. The app has evolved to create a personal chat-like interface, resulting in 85% of users actively engaging, with average daily usage at 31 minutes. Superfy seeks to expand its reach beyond Gen Z in the future.

SPONSOR US

🦾 Get your product in front of AI enthusiasts

THAT’S A WRAP

If you have anything interesting to share, please reach out to us by sending us a DM on Twitter: @HyriseAI