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:

• ChatGPT Gets Vision & Screensharing Capabilities

• Microsoft Launches New ‘Phi-4’ AI Model

• Anthropic’s Claude 3.5 Haiku Goes Global

• Meta Launched an Open-Source Watermarking Tool

• And more AI news….

Top Developments

OpenAI

ChatGPT Gets Vision & Screensharing Capabilities

The Gist: OpenAI's ChatGPT now supports screensharing and real-time video analysis, enabling users to interact with their surroundings and apps in new ways, rivaling capabilities recently launched by Google’s Gemini 2.

Key Details:

• Video mode allows ChatGPT to see and respond in real-time, identifying objects, assisting with tasks, and remembering introduced individuals.

• Screen-sharing lets users display apps or browsers to ChatGPT for context-aware assistance, expanding its utility for mobile and desktop environments.

• The features are available for Teams, Plus, and Pro users, with Enterprise and Edu support coming in January, though unavailable in parts of Europe.

• A festive “Santa Mode” offers a playful holiday-themed interaction, accessible across all advanced voice mode platforms until early January.

Microsoft

Microsoft Launches New ‘Phi-4’ AI Model

The Gist: Microsoft’s Phi-4, a 14-billion-parameter AI model, surpasses larger rivals like Google’s Gemini Pro 1.5 in mathematical reasoning, demonstrating that smaller, efficient models can deliver top-tier performance without massive computational costs.

Key Details:

• Phi-4 achieves superior results in complex mathematical reasoning, challenging the “bigger is better” philosophy of AI development.

• Its efficiency makes it more cost-effective, reducing energy and computational resource requirements for businesses adopting AI solutions.

• The model excels in specialized applications such as scientific research, engineering, and financial modeling, particularly in rigorous mathematical tasks.

• Released on Azure AI Foundry under a research license, Phi-4 includes safety features, monitoring tools, and content filters to ensure responsible AI use.

Anthropic

Anthropic’s Claude 3.5 Haiku Goes Global

The Gist: Anthropic has launched Claude 3.5 Haiku, its fastest and most advanced AI model, for all users worldwide. The model outperforms competitors like GPT-4o in benchmarks, making it ideal for diverse personal and business applications.

Key Details:

• Claude 3.5 Haiku is available to free and paid users, offering faster performance, better instruction adherence, and optimized tool use compared to its predecessor, Claude 3 Opus.

• The model excels in benchmarks, including Software Engineering (40.6%), HumanEval, and Graduate-Level Google-Proof Q&A, surpassing GPT-4o and GPT-4o Mini.

• Designed for versatility, it supports user-centric applications, specialized sub-agent tasks, and data-driven experiences for enterprise needs.

• Optimized for AWS Trainium2 AI chipset, it enables low-latency inference through Amazon Bedrock, improving efficiency for cloud-based deployments.

Quick Gist

Meta launched an open-source watermarking tool, Meta Video Seal, to detect AI-generated videos (Read More).

Meta appointed Clara Shih from Salesforce to head a new Business AI unit, leveraging Llama models to enhance advertising and customer engagement (Read More).

Meta introduced Meta Motivo, an advanced AI model that boosts the realism of avatars in the Metaverse, advancing the development of immersive virtual experiences (Read More).

Google added a "Summarize this folder" tool to its Gemini AI assistant in Google Drive, allowing users to generate overviews and locate files more efficiently (Read More).

Microsoft enhanced its Copilot AI in Edge on Android, introducing a native experience for summarizing content and delivering context-aware responses on mobile browsers (Read More).

Apple is collaborating with Broadcom to develop its first AI-focused server chip, "Baltra," slated for mass production in 2026 (Read More).

Elon Musk's Grok AI chatbot on X now includes a feature that describes memes and images for premium users (Read More).

Apple updated its iWork suite to version 14.3, introducing AI features like Writing Tools and Image Playground (Read More).

THAT’S IT FOR TODAY!