In partnership with
Hello from The AI Night,
Today in AI:
Anthropic Launches Claude Security for Enterprise Customers
Gemini Can Now Generate and Download Files Directly
Cursor Launches SDK for Coding Agent Deployment
Here's the deal: Anthropic is rolling out Claude Security, a code vulnerability scanner powered by Opus 4.7, to all Claude Enterprise customers. The tool scans repositories, identifies security flaws, and generates targeted patches directly within Claude.
The Breakdown:
Claude Security reasons through code like a security researcher, tracing data flows across files rather than matching known patterns.
Each finding goes through a multi-stage validation pipeline with confidence ratings to reduce false positives.
Teams can schedule recurring scans, scope to specific directories or branches, and export results to Slack, Jira, or CSV.
No API integration or custom agent build required.
Technology partners including CrowdStrike, Microsoft Security, Palo Alto Networks, SentinelOne, and Wiz are embedding Opus 4.7 into their platforms.
Services partners like Deloitte, Accenture, BCG, Infosys, and PwC are deploying Claude-integrated security solutions.
Access for Claude Team and Max customers is coming soon.
The bigger picture: Anthropic is positioning Claude as a frontline defense tool, not just a coding assistant. With AI compressing the gap between vulnerability discovery and exploitation, giving defenders access to the same frontier capabilities changes the security equation for enterprises that lack dedicated red teams.
Here's the deal: Google rolled out native file generation inside the Gemini app. Users can now prompt Gemini to create complete, downloadable files across multiple formats without leaving the conversation.
The Breakdown:
Supported formats include Google Docs, Sheets, Slides, PDF, .docx, .xlsx, .csv, LaTeX, Plain Text, RTF and Markdown.
Files can be downloaded directly to a device or exported to Google Drive.
The feature is live now for all Gemini app users globally.
The workflow eliminates the copy-paste-reformat loop. Users go from prompt to finished file in one step.
Use cases highlighted: budget proposals exported to Excel, brainstorm notes structured into bulleted drafts, long collaboration threads condensed into a single-page PDF or Word doc.
The bigger picture: ChatGPT and Claude generate files through workarounds. Google built it natively into the product that already owns your documents, spreadsheets, and slides. No competitor can copy that without building an entire office suite first. Every other AI assistant exports files. Google exports files into the ecosystem where your team already works.
Here's the deal: Cursor released the Cursor SDK in public beta, letting developers build programmatic coding agents using the same runtime, sandboxing, and models that power Cursor's desktop app, CLI, and web app. Install it with npm install @cursor/sdk and deploy agents in a few lines of TypeScript.
The Breakdown:
Agents run locally, on self-hosted workers, or on Cursor's cloud VMs with full sandboxing and durable sessions.
Cloud agents survive laptop sleep or network drops, can auto-create PRs, push branches, and attach screenshots.
SDK includes Cursor's full harness: codebase indexing, semantic search, MCP server connections, skills, hooks, and subagents.
Supports every model available in Cursor, including Composer 2, a specialized coding model built for cost-efficient agent tasks.
Teams are already using it in CI/CD pipelines to auto-fix failures, summarize changes, and build internal agent platforms.
Billed on standard token-based consumption pricing.
The bigger picture: Cursor spent two years solving sandboxing, indexing, and state management for its own product. Now it is selling that infrastructure to everyone else. That is the same playbook AWS ran with Amazon's internal tools. The dev tools companies building agent deployment platforms just discovered their biggest competitor is also their biggest customer's IDE.
The era of manual marketing ends this May!
Manual marketing had a good run.
But the teams winning right now aren't briefing, approving, and repeating. They're directing AI agents that execute the whole strategy for them.
The Agentic Marketing Summit (May 4–8) is a free, five-day event that shows you exactly how it works in practice. Not theory. Not a PDF checklist. Step-by-step insight to help you become an expert in AI marketing agents.
Hosted by 3x Inc 5000 founder Manick Bhan alongside the sharpest minds in the marketing world today.
The era of doing it yourself is over!
What else you need to know:
OpenAI launched Advanced Account Security, an opt-in setting requiring passkeys or security keys, disabling password and SMS recovery, and automatically excluding conversations from model training for high-risk users.
Manus launched Cloud Computer, a persistent cloud VM that lets non-technical users deploy 24/7 bots, scheduled scrapers, databases, and open-source tools using plain English prompts instead of code.
Microsoft posted $82.9 billion in Q3 revenue, up 18% year-over-year, with its AI business hitting a $37 billion annual run rate, growing 123% from last year.
Anthropic's claude-api skill, originally launched in Claude Code, is now bundled into CodeRabbit, JetBrains, Junie, Resolve AI, and Warp to help developers write production-ready Claude API code across more tools.
Mistral released Medium 3.5, a dense 128B model with 256k context and configurable reasoning effort, scoring 77.6% on SWE-Bench Verified while running on as few as four GPUs.
That’s it for today’s edition of The AI Night.
Our goal is to cut through the noise, surface what actually changed, and explain why it matters.
2 ways to support us:
Forward this to your AI-curious friend → https://www.theainight.com
Sponsor The AI Night and reach 500+ AI builders daily → passionfroot.me/theainight