In partnership with

Today: enterprise AI agents took center stage at the Databricks summit, xAI pushed Grok into Microsoft Office, and Anthropic's export standoff drew a sharp response from security researchers. Here are five AI developments worth knowing from the past day.

A WORD FROM OUR SPONSOR

Fast browsing. Faster thinking.

Your browser gets you to a page. Norton Neo gets you to the answer. The first safe AI-native browser built by Norton moves with you from idea to action without slowing you down. Magic Box understands your intent before you finish typing. AI that works inside your flow, not beside it. No prompting. No copy-pasting. No switching apps.

Built-in AI, instantly and for free. Privacy handled by Norton. Built-in VPN and ad blocking protect you by default. No configuration. No extra apps. Nothing to think about.

Fast. Safe. Intelligent. That's Neo.

Product & Platform

xAI brings Grok to Databricks and Microsoft Office. At the Databricks Data & AI Summit on June 18, xAI's Grok models became natively available inside Databricks' Agent Bricks platform, letting enterprise agents reason over data held in a customer's lakehouse without routing it through outside systems. xAI separately released free Grok add-ins for Microsoft Word, PowerPoint, and Excel that draft and revise documents and pull in web and connected-app context. The steps extend Grok's reach across the major cloud and productivity platforms it had already joined, including Azure, Amazon Bedrock, and Oracle Cloud. (Basenor)

The PyTorch Foundation launches a certification program. On June 18, the PyTorch Foundation and Linux Foundation Education introduced the PyTorch Certified Associate (PTCA), an entry-level credential that tests foundational skills in the widely used open-source deep-learning framework, alongside a new instructor-led associate training course. The organizations positioned it as a standardized way for employers to assess and hire AI engineering talent. It is the first formal certification tied directly to PyTorch. (Linux Foundation Education)

Industry & Infrastructure

Databricks pushes enterprise AI agents at its annual summit. At its Data & AI Summit in San Francisco (June 15–18), Databricks recast Agent Bricks as a unified platform for building, running, and governing AI agents, and said customers including AstraZeneca and Block have shipped agents on it. The company made Automatic Identity Management for Microsoft Entra ID generally available on AWS and Google Cloud, with Okta support in public preview, aimed at simplifying access control in regulated industries. Its free tier also gained five products, including serverless GPUs and the Lakebase database. (Databricks)

Anthropic's export standoff continues as security experts push back. A week after the U.S. Commerce Department's June 12 directive forced Anthropic to suspend worldwide access to its Fable 5 and Mythos 5 models, more than 150 cybersecurity executives and researchers signed an open letter urging the government to reverse the restrictions, arguing the models' security capabilities are not unique and that constraining defenders is counterproductive. Anthropic said on June 18 it expects the models to return “within days” and opened a Seoul office, while President Trump, speaking at the G7, called talks with the company “going fine.” No official restoration date had been set as of June 18; this follows our earlier coverage of the initial suspension. (Tech Times) (Cybersecurity Dive)

AI & Society

Pew finds Americans wary of AI's pace. In its “Americans and AI 2026” report released June 17, the Pew Research Center said roughly two-thirds of U.S. adults think AI is advancing too quickly, with views skewing more negative among younger adults. Adults under 50 were far more likely to use AI chatbots than older groups; ChatGPT was the most-used tool, followed by Gemini at about a quarter of adults, then Copilot and Meta AI. The survey of U.S. adults was conducted February 17–23, 2026. (Pew Research Center)

A WORD FROM OUR SPONSOR

Postgres Didn't Fail You. Your Architecture Did.

Adding a second database was supposed to fix things. Now you manage sync, drift, and pipelines on top of queries that are still slow.

TimescaleDB extends Postgres instead. Hypertables, 95% compression, continuous aggregates. One database. No pipeline.

That's today's edition of AI Current. Reply with what you're seeing in AI, and forward this to a colleague who would find it useful. See you tomorrow.

A WORD FROM OUR SPONSOR

Stop wiring up backends for your agents. Ghost is ephemeral postgres, forkable, built for this. Try it at ghost.build.

Keep Reading