Free AI for All From Mistral (under Apache 2.0)

📰 OpenAI lets ChatGPT loose on the Internet again

ChatGPT, created by OpenAI, now has a 'Browse' feature allowing it to access up-to-date information from the web using Bing, a significant leap from its previous knowledge cutoff in September 2021.

• OpenAI has rolled out the browsing functionality for its premium users.

• Users can activate the Browse feature in the Beta settings, listed under GPT-4.

• The AI chatbot will soon have the ability to listen, talk, and view uploaded images.

• OpenAI is enhancing ChatGPT's interface for more intuitive conversations and image discussions.

Why it Matters: Having ChatGPT access real-time information elevates its utility, allowing users to fetch and learn from the latest data rather than being limited to September 2021 insights. However, as the AI delves deeper into the web, concerns arise about data usage, privacy, and the potential implications of such vast access.

📰 Mistral AI makes its first large language model free for everyone

Mistral AI, a French startup, has launched its first free-to-use large language model called Mistral 7B, which they say outperforms other similar-sized models.

• Mistral AI's 7 billion model outperformed larger models like Llama 2 13B and Llama 1 34B in various benchmarks.

• The model offers capabilities similar to other models but at a significantly reduced compute cost.

• Unlike Meta’s Llama 2, which has usage restrictions, Mistral 7B is truly open access, thanks to its Apache 2.0 license.

• Despite its open access, Mistral 7B differs from being "open source" since it was developed privately.

Why it Matters: Mistral AI's launch of a high-performing, free-to-use model can revolutionize accessibility and usage of AI models, potentially leveling the playing field for developers and businesses.

📰 From AI Assistant to image restyler: Meta’s new AI features

Meta is unveiling a variety of AI-driven tools and products for its apps, VR headset, and upcoming smart glasses, all powered by generative AI to enhance user experiences.

• Meta's new features include an AI assistant, image editing, and AI personality chatbots mimicking celebrities.

• The core AI technology is based on "Llama 2," a language model that generates text and code, with Llama 3 anticipated for 2024.

• An AI Studio platform will be available for businesses to craft AI chatbots catering to brand values and customer service.

• Generative AI stickers, powered by "Emu," will be introduced in Meta apps, allowing users to create tailored stickers in seconds.

• Image editing tools, "restyle" and "backdrop," will allow users to transform visuals and backgrounds of photos on Instagram.

Why it Matters: Meta's AI advancements are reshaping user interactions on their platforms. With more personalized and dynamic tools, it's aiming to make digital communications more expressive and immersive.

🖌️ Adobe launches Photoshop’s web version with Firefly-powered AI tools: offering enhanced user workflows, tool identification, and collaboration through shareable links, yet missing some desktop features.

🏦 Generative AI in Finance: FinGPT's efficiency in financial sentiment analysis challenges Bloomberg's heavy investments.

Thank you for reading today’s edition!

Feel free to reply to our emails with questions, suggestions, or topics you'd like to see covered, or drop us a message on Twitter or Facebook.

Until tomorrow,
- Tsvetelin (Bits and Neurons)