This Week in Generative AI – 15th to 21st September

Microsoft’s Copilot AI coming to Windows on Sept. 26

Microsoft is expanding the release of its AI assistant, Copilot, across Windows 11, Microsoft 365, Edge, and Bing.
The centralized AI assistance will be accessible through an app or a right-click and will continue to add capabilities and connections to Microsoft’s most-used applications. The release will coincide with an operating system update on September 26, with additional features and integrations to follow.
Bing will also receive new features, including personalized chat history and improved search options. Microsoft 365 Copilot will be available to enterprise customers on November 1, 2023, alongside a new Microsoft 365 Chat assistant.

Amazon has introduced a new version of its Alexa voice assistant powered by its own large language model (LLM).
The updated Alexa offers improved conversational interactions, contextual understanding, and the ability to handle multiple requests. It integrates with over 200 smart home APIs, allowing for proactive management of the smart home environment.
The new Alexa is being rolled out slowly in the US and may have charges associated with enhanced features in the future.

Uber Eats is launching a chatbot powered by Google’s PaLM 2 LLM to enhance the food delivery experience.
The AI assistant will provide personalized recommendations and streamline the ordering process, aiming to engage customers and increase loyalty.
This move follows similar AI integrations by competitors DoorDash and Instacart.

Kinetica has launched a native Large Language Model (LLM) for language-to-SQL on enterprise data. The native LLM is secure and tailored to specific industries, providing reliable and accurate SQL generation. It prioritizes SQL generation optimization for consistent and accurate results, offering peace of mind for businesses.
The US Air Force is already using Kinetica’s embedded LLM to detect threats in airspace.
Kinetica plans to integrate with other LLM platforms in the future. The native LLM is available immediately to customers in a containerized, secure environment.

Oracle Cloud has launched a bare metal cloud infrastructure that utilizes Nvidia’s H100 Tensor Core GPUs for generative AI and LLM workloads.
A more affordable option with Nvidia’s L40S GPU will be available in early 2024.
The move aims to capitalize on the growing demand for AI technologies and generate profit for Oracle.