databricks launches google while openai expands its platform

Key Takeaways

• Databricks launched AI Gateway to simplify AI agent access to external tools like GitHub and Atlassian, enhancing security and control via Unity Catalog.
• Databricks introduced Agent Bricks for secure deployment of AI agents using company data, supporting multiple AI models and offering unified governance.
• Google released the Gemini AI app as a native application for macOS 15+, accessible via Option + Space, allowing screen context sharing and quick content generation.
• Google also launched new desktop applications for Windows and Mac, providing quicker access to search and AI capabilities outside a web browser.
• LinkedIn is testing an AI training jobs marketplace, offering users up to $150 per hour for tasks like chatbot training.
• Illinois bill SB3444 proposes to shield AI companies from liability for major harms if unintentional and safety protocols are followed, supported by OpenAI but concerning Anthropic.
• ARK Invest highlights intense competition for AI computing power and "enterprise lock-in" as key challenges for AI adoption in businesses.
• BlackLine's data engineer utilizes Snowflake and other tools to build AI products, such as the Verity Prepare assistant for account reconciliations.
• Ledger plans a 2026 security framework for AI agents, including hardware-anchored security features like Agent Identity and Proof of Human, to ensure human control and verify actions.
• AI tools can identify skin conditions from photos more accurately than internet searches but struggle to provide crucial medical advice, emphasizing the need for professional consultation.

Databricks AI Gateway simplifies tool access for AI agents

Databricks has launched its AI Gateway to make it easier for AI agents to connect to outside tools like GitHub and Atlassian. This new feature helps manage authentication and improves security. It connects with Unity Catalog, allowing administrators to control access to these tools like they do with data. AI agents can now act on behalf of users, ensuring they only access information the user is permitted to see. This simplifies the process of connecting AI to various services and improves overall security and management.

Databricks AI Gateway improves control over AI agent actions

Databricks is enhancing its AI Gateway to offer better control, visibility, and safety for AI agent workflows. The updated gateway helps manage how AI models use tools and APIs, providing unified governance for these interactions. It allows for setting specific permissions, ensuring agents act with user-level access and preventing unauthorized data use. The system also includes customizable guardrails for safety, like detecting personal information and preventing harmful content. Additionally, it offers detailed logging for cost management, debugging, and security audits.

Databricks launches Agent Bricks for secure AI agent deployment

Databricks has introduced Agent Bricks, a new platform for businesses to securely deploy AI agents that use company data. This platform helps manage how agents connect to data, use different AI models, and follow security rules. Agent Bricks aims to make AI agents reliable for important business tasks. Key features include support for multiple AI models, unified governance through Unity Catalog and AI Gateway, and improved accuracy by using company data for context. New capabilities include custom agent apps, workflow orchestration, and better document analysis.

Google Gemini app now available on Mac computers

The Gemini AI app is now available as a native application for macOS version 15 and newer. Users can download it for free to get AI assistance directly on their desktop. A keyboard shortcut, Option + Space, allows users to quickly access Gemini without switching applications. The app enables users to share their screen for context, helping Gemini understand and summarize information from local files or active windows. It also allows for quick image and video generation, aiming to integrate AI help seamlessly into users' workflows.

Google releases new desktop apps for Windows and Mac

Google has launched new desktop applications for Windows and macOS. The Google app for Windows, available since September in beta, now officially supports web and PC searches with features like AI Overviews and screen context sharing. For Mac users, Google has released a native Gemini app, offering AI assistance directly on the desktop. This app can be accessed via a shortcut and can use screen context for more relevant responses. Both apps aim to provide quicker access to Google's search and AI capabilities outside of a web browser.

LinkedIn tests AI jobs paying up to $150 per hour

LinkedIn is testing a new jobs marketplace focused on AI training. Through this platform, users could earn up to $150 per hour for tasks like training chatbots. The marketplace is currently rolling out early job listings and using AI for screening to match people with paid training projects. This initiative aims to connect individuals with opportunities in the growing field of artificial intelligence.

Illinois bill shields AI companies from liability for major harms

A new bill in Illinois, SB3444, proposes to shield AI companies from liability for significant harms they cause, such as mass casualties or major property damage. If passed, companies would be exempt if they can prove the harm was unintentional and they followed safety protocols. Compliance with the European Union's AI Act would also grant exemption. OpenAI supports the bill, while Anthropic has concerns, highlighting a debate over AI safety and innovation.

AI can help identify rashes but may miss crucial medical advice

New research suggests AI tools can help people identify skin conditions from photos, often more accurately than internet searches. However, these tools struggle to provide guidance on whether to seek medical treatment. Dermatologists note that AI lacks a patient's full medical history and symptom progression, which are vital for diagnosis. While AI can be a useful starting point for questions, doctors warn against relying on it solely and emphasize the importance of seeking professional medical care.

ARK Invest discusses AI compute race and enterprise adoption

ARK Invest experts discussed the intense competition for AI computing power and the challenges businesses face in adopting new AI technologies. They highlighted 'enterprise lock-in,' where companies are hesitant to switch from existing systems. The panel emphasized that access to powerful computing resources is crucial for developing advanced AI models. They also debated whether focusing on compute power or product development is more critical for AI success, noting that integration and clear value are key for enterprise adoption.

BlackLine data engineer builds AI products using Snowflake

Dylan Kaufman, a staff data engineer at BlackLine, works on developing AI products like the Verity Prepare assistant for account reconciliations. He uses tools such as Visual Studio Code and Claude Code daily for tasks like generating code for infrastructure and testing. Kaufman also serves as BlackLine's Snowflake expert, focusing on data model and pipeline design. He collaborates with his team remotely to improve and expand the company's finance and accounting software, emphasizing the satisfaction of getting new features working.

Ledger announces 2026 AI security plan

Ledger is launching a security framework for AI agents throughout 2026 to address risks associated with their access to sensitive resources. The plan includes features like Agent Identity, Ledger CLIs, and Skills in Q2, followed by Agent Intents and Policies in Q3, and Proof of Human in Q4. Ledger aims to provide hardware-anchored security for AI agents, ensuring humans remain in control and verifying actions. This initiative seeks to protect against threats like prompt injection and autonomous execution by integrating Ledger's hardware security with AI agent operations.

AI drives hardware demand for content creators

Smartphones, while popular for content creation, have physical limitations that are driving creators to invest in dedicated hardware like microphones, lenses, and cameras. A report shows that creators are increasingly spending over $1,000 on gear upgrades. The rise of AI is speeding up content production, making capture quality more important and increasing demand for better equipment. This trend is fueling a hardware market expected to grow significantly worldwide, especially in regions like India and the USA.

Google previews Android and AI sessions for I/O 2026

Google has previewed upcoming sessions for its I/O 2026 conference, focusing on advancements in Android and AI. Sessions will cover Google's AI stack, including new model capabilities and infrastructure for building AI apps. Updates on Android development will highlight Jetpack Compose, performance improvements, and new features for various devices, including agentic automation. Other previews include what's new in Chrome, Google Play, Firebase, the Gemma open model family, Android development tools, Web UI, and the Flutter framework.

Sources

• Databricks AI Gateway Tames Tool Access
• Databricks Tames Agentic AI
• Databricks Launches Agent Bricks Platform
• The Gemini app is now on Mac
• Google releases new apps for Windows and MacOS
• LinkedIn AI training jobs could pay up to Rs 14,000 an hour: Who can apply, how it works
• Should AI Companies Be Held Liable for the 'Critical Harms' They Cause? A New Bill Says No
• AI can help identify your rash, but doctors warn it may miss what matters most
• ARK Invest: AI's Compute Race and Enterprise Lock-In
• BlackLine Data Engineering Team: Building AI Products and Using Snowflake
• Ledger's 2026 AI Security Roadmap
• 'Smartphones have physical limitations': Report explains why AI is kickstarting a billion-dollar hardware arms race for millions of creators worldwide
• Google previews I/O 2026 sessions: What’s new in Android, AI, more