Welcome to Get The Gist, where every weekday we share an easy-to-read summary of the latest and greatest developments in AI—news, innovations, and trends—all delivered in under 5 minutes! ⏱

In today’s edition:

• Meta Opens AI Models to US Defense and Security Agencies

• OpenAI Hires Former Meta AR Lead to Robotics Team

• Microsoft's Copilot Vision Brings Screen-Aware AI to Your Browser

• ChatGPT Becomes Fastest Growing Subscription Service Globally

• And more AI news….

Top Developments

Meta

Meta Opens AI Models to US Defense and Security Agencies

The Gist: Meta is making its Llama AI models accessible to U.S. defense agencies and contractors, marking a strategic shift to support national security efforts and reinforce the AI model’s presence in the public sector.

Key Details:

• Meta’s Llama models, previously restricted from military use, are now accessible to over a dozen U.S. defense agencies and contractors, including Lockheed Martin and Palantir.

• This decision extends to the intelligence-sharing "Five Eyes" nations—UK, Canada, Australia, and New Zealand—but excludes Israel and other allies.

• Meta’s Nick Clegg described the move as part of Meta's commitment to U.S. security and global democratic values, while Meta emphasizes that these agencies are responsible for ethical AI use.

• CEO Mark Zuckerberg aims to make Llama a dominant AI model, positioning it as a preferred choice in government, with potential applications in data analysis, cybersecurity, and anti-terrorism efforts.

OpenAI

OpenAI Hires Former Meta AR Lead to Robotics Team

The Gist: OpenAI has hired Caitlin Kalinowski, Meta's former head of AR glasses hardware, to lead its revived robotics and consumer hardware initiatives, signaling a renewed focus on integrating AI with physical technology.

Key Details:

• Caitlin Kalinowski, known for overseeing Meta’s AR glasses prototype Orion, will lead OpenAI’s robotics team, aiming to bring AI-driven robotics into real-world applications.

• This hire follows OpenAI’s efforts to rebuild its robotics division, which was previously closed in 2020, with a focus on supporting third-party robotic developers rather than creating in-house hardware.

• OpenAI is also collaborating with former Apple design chief Jony Ive on a new AI device startup, further expanding its reach into consumer hardware.

• Orion, Meta’s high-profile AR glasses project led by Kalinowski, remains unreleased but was showcased at Meta Connect, underscoring her experience in advanced tech hardware.

Microsoft Copilot

Microsoft's Copilot Vision Brings Screen-Aware AI to Your Browser

The Gist: Microsoft’s Copilot Vision is set to launch widely, letting the AI assistant see and respond to whatever is on your screen in real time, enhancing tasks like web browsing, reading, and recipe adjustments.

Key Details:

• Copilot Vision, initially limited to Copilot Labs, will soon be available for all users, accessible through the Microsoft Edge browser with a single click.

• The AI can view and interact with text on-screen, helping with tasks like offering travel details, recipe modifications, and suggestions based on the content you’re viewing.

• Microsoft assures that user privacy is a priority; browsing data from each session will not carry over, and the feature will initially be limited to selected, secure websites.

• As competition intensifies with AI tools like ChatGPT and Claude, Copilot Vision aims to provide unique, screen-aware assistance directly within the browser.

Quick Gist

Microsoft's Visual Studio Code 1.95 introduces Copilot Edits for AI-powered code suggestions across multiple files and enhances support for multiple GitHub accounts (Read More).

xAI, Elon Musk's AI startup, has launched a public API for its Grok language models, offering $25 in monthly credits for developers to encourage adoption amidst competition with other AI platforms (Read More).

Elon Musk encouraged users to submit medical images to his Grok AI chatbot for diagnosis, raising significant privacy concerns and eliciting mixed reviews on its accuracy from healthcare professionals (Read More).

Google Research unveiled InkSight, an AI system that accurately transforms handwritten text from photos into digital text without requiring intermediary devices (Read More).

Claude 3.5 Haiku API launched on Amazon Bedrock and Google Cloud, delivering superior speed and coding capabilities, outpacing leading models like GPT-4o in performance benchmarks (Read More).

Anthropic has increased the pricing for its Claude 3.5 Haiku AI model, now set at $1 per million input tokens and $5 per million output tokens, citing significant improvements in performance (Read More).

OpenAI's CEO Sam Altman announced plans to focus on releasing GPT-o1 instead of GPT-5 this year, emphasizing a more methodical approach to model complexity and capabilities (Read More).

ChatGPT has surged by 296% in growth compared to 2023, becoming the fastest-growing subscription service globally, amid rising multi-subscription trends in various online categories (Read More).

THAT’S IT FOR TODAY!