🤖Amazon's ChatGPT Competitor

OpenAI postpones launch of advanced ChatGPT voice mode, acquires team behind Multi

Key Points 

• OpenAI now plans to start testing the voice mode next month, with broader access for all Plus customers by fall.

• The company indicated that the exact timelines depend on meeting their high standards for safety and reliability.

• OpenAI acquired Multi, a startup specializing in screen-sharing and collaboration technologies for Mac users.

• The acquisition raises speculation about OpenAI integrating Multi's technology into AI models like GPT-4o, potentially enabling them to execute tasks based on voice or text commands. 

🎙️News - The advanced voice mode for ChatGPT, which led to a dispute with Scarlett Johansson, was a key part of OpenAI’s Spring Update event in May. The company initially planned to roll out this feature to paying users within weeks but now states it needs more time.

OpenAI now plans to release an alpha version to a small group of ChatGPT Plus subscribers in a month, with broader access for all Plus customers by fall. If you're wondering why, OpenAI says it requires additional time to improve the feature's ability to "detect and refuse certain content." The company noted that the exact timelines will thus depend on it  meeting their high standards for safety and reliability.

In related news, the desktop app for macOS is now available to all ChatGPT users. And as for the new video and screen-sharing capabilities, OpenAI indicated that it will keep users informed about the timeline.

💰What's more? Earlier this week, OpenAI acquired Multi, a startup specializing in screen-sharing and collaboration technologies for Mac users. While specific details about the deal's terms have not been disclosed publicly, Multi's co-founder and CEO Alexander Embiricos announced on his X account that he (and likely the entire Multi team have) joined OpenAI's "ChatGPT desktop team." This team is responsible for developing the ChatGPT desktop app for Mac, which was unveiled in May 2024.

Looking ahead, there's a possibility that OpenAI could integrate Multi's tech to enable AI models like GPT-4o to perform tasks on users' computers based on voice or text commands—for example, handling tasks like creating and sending documents.

Amazon is working on an AI chatbot to rival OpenAI’s ChatGPT

Key Points 

• Amazon is developing an AI chatbot named "Metis" to rival OpenAI's ChatGPT.

• Amazon plans to use the RAG method, enabling Metis to reference external information for up-to-date answers.

☕News - Amazon is developing an AI chatbot named "Metis" to compete with OpenAI's ChatGPT. According to Business Insider, the chatbot will be available via a web browser and will use Amazon's internal AI model, Olympus, which is said to be more powerful than their publicly available model, Titan. 

🍨The Scoop - 

• According to an internal document, Metis, Amazon's AI chatbot, can generate conversational text and image responses, provide source links, suggest follow-up questions, and create images.

• Amazon plans to use the retrieval-augmented generation (RAG) method for Metis, allowing it to reference external information beyond its original training data and provide more up-to-date answers. 

• Additionally, Metis will function as an AI agent, performing tasks like turning on lights and booking flights.

• However, Amazon is already late to the AI chatbot race, and with Metis not launching until September, competitors will have more time to improve their products.