🧠 Microsoft launches Copilot Cowork

Good Morning! Microsoft just introduced a new AI agent powered by Anthropic. Plus, you’ll learn how to transfer your memory from any AI provider to Claude in seconds.

Plus, in today’s AI newsletter:

• Microsoft announces Copilot Cowork

• Anthropic Builds AI to Review AI-Generated Code

• OpenAI Acquires Promptfoo to Secure its AI Agents

• How to Transfer Your Memory From Any AI Provider to Claude

• 4 new AI tools worth trying

AI TOOLS

🧠 Microsoft announces Copilot Cowork

Microsoft just dropped "Copilot Cowork," a cloud-powered AI agent that doesn't just chat, it actively executes multi-step workflows across your entire Microsoft 365 suite. The biggest surprise? It's heavily powered by Anthropic's Claude, not OpenAI.

• Instead of waiting for prompts, Cowork runs autonomously in the background. You can ask it to prep for a client meeting, and it will analyze your Outlook threads, read your SharePoint files, schedule the prep time, and build a polished PowerPoint deck while you do other things.

• While Anthropic recently launched its own local desktop version ("Claude Cowork"), Microsoft's offering lives in the cloud and taps directly into a company's massive enterprise data graph, fully protected by existing IT security protocols.

• This marks a massive strategic shift. Despite sinking $13 billion into OpenAI, Microsoft is taking a "model-agnostic" approach, plugging rival Anthropic’s tech right into the heart of its flagship enterprise suite to get the best results.

Microsoft choosing Anthropic to power its ultimate enterprise agent proves that the AI war won't just be won by whoever has the smartest model. It will be won by whoever is most deeply embedded in the workflows and security systems that Fortune 500 companies already trust.

VIBE CODING

🛑 Anthropic Builds AI to Review AI-Generated Code

The rise of "vibe coding" has completely changed how developers work, but it's created a massive new problem: humans simply can't review the sheer flood of AI-generated code fast enough. To fix this bottleneck, Anthropic just launched "Code Review" inside Claude Code.

• Instead of nagging engineers about code style, this multi-agent system strictly hunts down critical logic errors before they merge into the codebase.

• It automatically analyzes pull requests, explaining exactly what the issue is, why it’s a problem, and how to fix it with color-coded severity rankings.

• The tool uses multiple AI agents working in parallel to examine the code from different dimensions, with a final agent aggregating and prioritizing the most urgent fixes.

• It’s a premium, token-based feature estimated to cost $15 to $25 per review, targeted squarely at enterprise giants like Uber and Salesforce.

AI tools have solved the blank page problem, but they've just shifted the bottleneck from writing code to reviewing it. By having AI police its own output, Anthropic is attempting to close the loop on fully automated software engineering. It’s also a massive strategic flex, as Anthropic battles the DoD in court over "supply chain risk" designations, they are leaning hard into their booming enterprise business, which just crossed a staggering $2.5 billion in run-rate revenue.

AI PARTNERSHIPS

🛡️ OpenAI Acquires Promptfoo to Secure its AI Agents

OpenAI just announced the acquisition of AI security startup Promptfoo. As independent AI agents take on more digital tasks, bad actors are finding fresh opportunities to manipulate these systems, and OpenAI is scrambling to prove its technology is safe for critical enterprise use.

• Promptfoo's testing technology will be integrated directly into OpenAI Frontier, the company's enterprise platform for AI agents.

• The startup, which was valued at $86 million in its last funding round, builds tools that allow companies to automate red-teaming, evaluate workflows for vulnerabilities, and monitor compliance risks.

• Promptfoo's open-source interface and library are already utilized by more than 25% of Fortune 500 companies.

• OpenAI has stated it expects to continue building out Promptfoo’s open-source offerings alongside the new native enterprise integrations.

By buying a dedicated security firm to bake automated red-teaming natively into its platform, OpenAI is signaling that robust, enterprise-grade safety is now the mandatory cost of entry for the next phase of agentic AI deployment.

HOW TO AI

🗂️ How to transfer your memory from any AI provider to Claude

In this tutorial, you will learn how to instantly transfer your saved memories, custom instructions, and personal context from ChatGPT (or any AI provider) directly into Claude using their native import tool, ensuring you never lose your tailored AI setup.

🧰 Who is This For

• AI researchers streamlining model integrations to improve small business efficiency

• Techies and developers migrating their local coding workflows

• Content creators who want to keep their generated brand voice intact

• Anyone tired of manually re-teaching a new AI their daily preferences and workflows

STEP 1: Initiate the Memory Import

Head over to Claude and ensure you are logged into your account. Once you’re in, click your profile picture in the bottom left corner. Navigate to Settings, select the Capabilities tab, and look for the Import memory from other AI providers button. Click Start import, which will reveal a specific prompt designed by Anthropic to extract all your saved context. Copy this prompt to your clipboard.

STEP 2: Extract Your Data from ChatGPT ( Or any AI Provider )

Head over to your ChatGPT account. Before pasting the prompt, click your profile picture, go to Personalization, and verify that Reference saved memories, browser memories, and chat history are all toggled on.

Next, switch your model from Auto to a Thinking model. These reasoning models perform a much deeper dive into your account history than standard models. Then Paste the prompt you copied from Claude and hit enter to generate a comprehensive code block of your data.

STEP 3: Import and Clean Up in Claude

Now copy the generated code block from ChatGPT and return to the open import window in Claude, paste the entire block into the text field, and click Add to memory. Claude will automatically format this raw data into cleanly categorized, easy-to-read memories.

You can review the list and click the pen icon to manually correct any minor mistakes, but otherwise, your base profile is fully migrated and ready to go!

STEP 4: Transfer Project-Specific Context

If you organize your ChatGPT workflows into specific Projects (with their own files and instructions), you will need to move those containers separately. Open your ChatGPT Project, run the exact same extraction prompt using a Thinking model, and copy the output. Then, go to Claude Projects, create a new project, and paste that extracted data directly into the project's instructions and memory fields.

Within 12 hours, Claude will fully sync this context to that specific workspace.

Anthropic sues to block the DOD from designating it a supply chain risk, says the designation is unlawful and violates its free speech and due process rights.

Jay Graber is stepping down as Bluesky CEO, saying a “seasoned operator focused on scaling and execution” is needed; VC Toni Schneider is named interim CEO.

A study finds LLMs from Anthropic, Google, OpenAI, and xAI can facilitate academic fraud, specifically helping non-researchers submit fabricated papers to arXiv.

Apple delayed the release of its smart home display, planned for this month, until later this year to let the company finish work on the new Siri.

🤝 Copilot Cowork: Microsoft's agentic AI assistant powered by Claude

🔒 Codex Security: OpenAI Codex security agent that scans code and fixes bugs

⚙️ Autoresearch: Andrej Karpathy’s tool for AI-driven LLM training

🎥 Replit Animation: Turn text prompts into animated videos

THAT’S IT FOR TODAY

Thanks for making it to the end! I put my heart into every email I send, I hope you are enjoying it. Let me know your thoughts so I can make the next one even better!

See you tomorrow :)

- Dr. Alvaro Cintas