What\u2019s new in AI tools
Automated weekly roundup of release notes and changelogs across every AI tool we track. Currently indexing 195 updates.
This week \u2014 2026-W18
AI Tools This Week: 42 updates across 7 products
-
Copilot code review: An improved review experience
GitHub Copilot code review received improvements to enhance the overall review experience for developers.
-
Your Prompts and Skills need a system of record
Studio now provides AI prompts and skills with a system of record that is versioned, owned, and traceable.
-
Unify your marketing data with Lakeflow Connect
Databricks introduces Lakeflow Connect to help organizations unify their marketing data across different sources and platforms.
-
Improving Lakebase Postgres compute cache
Databricks announces performance improvements to the compute cache system in Lakebase Postgres for better query performance.
-
Unifying governance across engines and catalogs in the Open Lakehouse
Databricks enhances governance capabilities across different compute engines and data catalogs within the Open Lakehouse architecture.
-
Control which AI models your agents can use
Workspace owners can now control which AI models are available in Notion Agent and Custom Agents, with separate controls and default model settings for compliance and cost management.
-
Beyond embedding: How to secure AI/BI Dashboards for every viewer
Databricks provides enhanced security features for embedding AI/BI Dashboards in customer-facing applications with viewer-specific access controls.
-
Your Prompts and Skills need a system of record
Studio now provides a system of record for AI prompts and skills with versioning, ownership, and traceability features. This gives developers better control over their AI application components.
-
Gemini 3.8 Flash is now available in GitHub Copilot
Google's Gemini 3.8 Flash model has been added as an available option in GitHub Copilot.
-
A Three-week Task Now Takes 20 Minutes: Kansas City Firm Cuts Demand Prep 99% and Case Time in Half Using AI
Nash Franciskato law firm reduced demand preparation time from three weeks to 20 minutes using EvenUp's AI platform. The firm reports they can now handle 100 additional cases per year with the same team size.
-
Gemini 3.8 Flash is now available in GitHub Copilot
Google's Gemini 3.8 Flash model is now available as an option in GitHub Copilot.
-
Announcing the Databricks Big Book of AgentOps
Databricks released a comprehensive guide on AgentOps, the operating discipline for building, deploying and managing AI agents. The resource provides best practices for agent operations.
-
Content exclusions generally available in Copilot app and CLI
Content exclusion features are now generally available in the Copilot app and command-line interface.
-
Claude Fable 5.1 is generally available in GitHub Copilot
Anthropic's Claude Fable 5.1 model is now generally available for use in GitHub Copilot.
-
How we eliminated $1 million a year of wasted AI agent spend in one hour
Databricks shared how their engineering team identified and eliminated significant waste in AI agent spending through optimization techniques. The case study demonstrates cost management strategies for AI workloads.
-
Claude Fable 5.1 is generally available in GitHub Copilot
Anthropic's Claude Fable 5.1 model is now generally available in GitHub Copilot.
-
GroqCloud: Expanding to Meet Demand
Groq announces expansion of GroqCloud platform capabilities to meet growing customer demand. The expansion includes infrastructure improvements and capacity increases.
-
Groq Becomes an NVIDIA Cloud Partner
Groq announces partnership with NVIDIA as an official cloud partner. This partnership expands Groq's integration with NVIDIA's AI infrastructure ecosystem.
-
Leanstral 1.5: Proof Abundance for All
Mistral released Leanstral 1.5, an updated version of their formal reasoning model. The model provides enhanced mathematical proof capabilities and formal verification features.
-
Prompts and Skills Management in Studio
Mistral Studio now provides a system of record for AI prompts and skills with versioning, ownership, and traceability features. This gives users better control over their AI prompt management workflow.
-
Ask your agent to suggest edits
Agents can now propose changes instead of making them directly, allowing users to review and approve edits line by line for better control over content modifications.
-
Ask your agent to suggest edits
Agents can now propose changes instead of making them directly, allowing users to approve edits line by line. Great for grammar passes and precise content review.
-
Fast, fault-tolerant PyTorch training on AI Runtime
Databricks improves PyTorch training efficiency on AI Runtime by focusing on 'goodput' metrics at scale. The enhancement provides faster and more reliable model training capabilities.
-
Beyond answers: New Genie One features to turn insights into action
Databricks introduces new features for Genie One that extend beyond providing answers to enable actionable insights. The update enhances the AI tool's capability to transform data queries into executable actions.
-
Ask your agent to suggest edits
Agents can now propose changes instead of making them directly, allowing users to review and approve edits line by line.
-
GitHub Copilot in Visual Studio — August update
GitHub Copilot receives an August update for Visual Studio with new features and improvements.
-
How EvenUp Uses AI Agents to Build Ground Truth Data With More Productive and Accurate Labeling
EvenUp published technical details on how they use AI agents to improve data labeling productivity and accuracy for building ground truth datasets.
-
How EvenUp Uses AI Agents to Build Ground Truth Data With More Productive and Accurate Labeling
EvenUp published technical details on their use of AI agents for data labeling to improve ground truth data quality. The approach increases productivity and accuracy in their machine learning data preparation processes.
-
How EvenUp Uses AI Agents to Build Ground Truth Data With More Productive and Accurate Labeling
EvenUp published technical details on their approach to using AI agents for data labeling to improve ground truth data quality. The work was supervised by VP of AI Haixun Wang and focuses on enhancing labeling productivity and accuracy.
-
How EvenUp Uses AI Agents to Build Ground Truth Data With More Productive and Accurate Labeling
EvenUp published technical details on their approach to using AI agents for data labeling to create ground truth datasets. The work was supervised by VP of AI Haixun Wang and focuses on improving labeling productivity and accuracy.
-
Copilot code review: Resolution reasons and expanded capabilities
Copilot code review now includes resolution reasons and expanded review capabilities for better code analysis.
-
10 AI Lessons from Driving 200+ Million Fully Autonomous Miles
Waymo shared ten fundamental insights about building AI systems based on their experience driving over 200 million autonomous miles.
-
GitHub Copilot app Customize tab is generally available
The Customize tab in the GitHub Copilot app has reached general availability for all users.
-
April Drop 2026
Enables slide and interactive HTML page generation, plus embedded MCP Apps in Glean Assistant for cross-tool actions.
-
June Drop 2026
Adds meeting notes functionality to turn meetings into searchable context and interactive content for team collaboration.
-
Setting the Global Standard for Privacy in Autonomous Mobility
Waymo outlined their privacy approach for riders, partners, and road users as they expand to serve more customers globally.
-
Setting the Global Standard for Privacy in Autonomous Mobility
Waymo outlined new privacy standards and protections for autonomous vehicle data collection and user information.
-
Leanstral 1.5: Proof Abundance for All
Mistral released Leanstral 1.5, an updated version of their mathematical proof verification model.
-
A look under our trunk: what's in our compute
Waymo reveals details about their custom compute hardware and silicon designed for real-time autonomous driving. The system is co-designed with sensors and algorithms for deterministic, low-latency performance.
-
Busting SQL Migration Myths: How New SQL Features Make Lift-and-Shift to Lakehouse Easier
Databricks introduced new SQL features designed to simplify migration from traditional data warehouses to the lakehouse architecture.
-
Inbound Private Link now supports account-level Genie One, the account console, and custom URLs
Databricks expanded Inbound Private Link support to include account-level Genie One, the account console, and custom URLs for enhanced enterprise security.
-
Your Developer Portal, now in the sidebar
Adds a new Developer section to the sidebar for managing Workers, connections, and personal access tokens, plus a developer bar for easy API object access.
-
Your Developer Portal, now in the sidebar
New Developer section in workspace sidebar for managing Workers, connections, and tokens. Includes developer bar for easy API ID copying.
-
Your Developer Portal, now in the sidebar
New Developer section in workspace sidebar for managing Workers, connections, and tokens, plus a developer bar for copying API IDs.
-
Your Developer Portal, now in the sidebar
New Developer section in workspace sidebar for managing Workers, connections, and tokens, plus developer bar for copying API IDs.
-
Your Developer Portal, now in the sidebar
Added a new Developer section to the sidebar for managing Workers, connections, and personal access tokens. Includes a developer bar for easily copying API IDs and objects.
-
Enterprise managed settings in GitHub Copilot for JetBrains
Enterprise administrators can now manage GitHub Copilot settings centrally for JetBrains IDEs.
-
How Databricks Feature Store serves features with sub-second freshness
Databricks enhanced its Feature Store to deliver machine learning features with sub-second latency for real-time applications like fraud detection.
-
Bringing more control over your connectors
Mistral enhances connector functionality with improved control and management capabilities for enterprise integrations.
-
Leanstral 1.5: Proof Abundance for All
Mistral releases Leanstral 1.5, an updated version of their mathematical proof generation model.