In partnership with
Hello from The AI Night,
Today in AI:
Anthropic Launches Claude Managed Agents for Production
Perplexity Launches Billion Dollar Build Challenge for AI Startups
Google Launches Gemini Notebooks to Turn AI Chats Into Knowledge Bases
Here's the deal: Anthropic released Claude Managed Agents in public beta on the Claude Platform. It's a set of composable APIs that handle sandboxed execution, state management, authentication, and tool orchestration so teams can skip months of infrastructure work and deploy agents directly.
The Breakdown:
Includes secure sandboxing, long running sessions (hours of autonomous operation), scoped permissions, identity management and end to end tracing out of the box.
Multi agent coordination (agents spawning and directing other agents) is in research preview with gated access.
A self evaluation loop lets agents iterate against defined success criteria. In internal testing on structured file generation, this improved task success by up to 10 points over standard prompting.
Pricing: standard Claude Platform token rates plus $0.08 per session/hour for active runtime.
Early adopters include Notion, Rakuten, Asana, Sentry and Vibecode, reporting deployment timelines dropping from months to days or weeks.
The bigger picture: This positions Anthropic as a full-stack agent platform, not just a model provider. Teams building coding, finance, legal or productivity agents can now offload orchestration and governance entirely, which lowers the barrier to shipping agentic products at scale.
Here's the deal: Perplexity announced "The Billion Dollar Build," an 8-week challenge offering up to $1M in seed investment and $1M in Computer Credits to founders who build a company using Perplexity Computer as their primary AI tool.
The Breakdown:
Registration opens April 14; submissions due June 2, with a live reveal on June 9.
Up to 3 winners split the seed pool via Perplexity Fund, though investment is not guaranteed and terms are set case by case.
Entrants must hold an active Max or Pro subscription before April 13 and be US legal residents (18+).
Winners must incorporate as a Delaware C-Corp to qualify for funding.
Perplexity claims Computer saved over $1.6M internally and performed the equivalent of 3.25 years of work in its first 4 weeks.
Submissions require a product video, application, and traction data showing real users or revenue.
Other AI tools are allowed, but Computer must be the primary system.
The bigger picture: This positions Perplexity Computer as a full-stack company-building platform, not just a search tool. If winning entries show real traction, it could validate agentic AI as a serious alternative to traditional team scaling.
Here's the deal: Google is rolling out Notebooks inside the Gemini app, giving users a dedicated space to organize chats, files, and custom instructions around specific topics. Notebooks sync bidirectionally with NotebookLM, so any source added in one app automatically appears in the other.
The Breakdown:
Notebooks let users group conversations, custom instructions, and uploaded files (documents, PDFs) around a single topic or project.
Gemini draws on those curated sources alongside its own tools and web search to generate responses.
Notebooks sync bidirectionally with NotebookLM, so sources added in either app appear in both.
This unlocks NotebookLM features like Video Overviews and Infographics from within a Gemini workflow.
Rolling out this week on web for Google AI Ultra, Pro, and Plus subscribers.
Mobile, additional European countries, and free-tier access are coming in the following weeks.
The bigger picture: This positions Gemini as more than a chat interface. Persistent, source-grounded workspaces let users run longer projects without losing context between sessions. The NotebookLM sync is the real lever: it bridges conversational AI with structured research tools, giving paying subscribers a reason to stay inside Google's ecosystem.
88% resolved. 22% stayed loyal. What went wrong?
That's the AI paradox hiding in your CX stack. Tickets close. Customers leave. And most teams don't see it coming because they're measuring the wrong things.
Efficiency metrics look great on paper. Handle time down. Containment rate up. But customer loyalty? That's a different story — and it's one your current dashboards probably aren't telling you.
Gladly's 2026 Customer Expectations Report surveyed thousands of real consumers to find out exactly where AI-powered service breaks trust, and what separates the platforms that drive retention from the ones that quietly erode it.
If you're architecting the CX stack, this is the data you need to build it right. Not just fast. Not just cheap. Built to last.
What else you need to know:
Intel and Google announced a multiyear collaboration to co-develop future Xeon processors and custom ASIC-based infrastructure processing units (IPUs) designed to improve performance and efficiency across Google's AI cloud infrastructure.
Cursor's Bugbot now resolves 78% of flagged bugs in PRs, up from 52% at launch, and introduced self-improving "learned rules" generated from developer feedback across 110,000+ repos.
Google is expanding its AI-powered Finance platform to over 100 countries, adding AI-driven market research, advanced charting, live earnings call access, and expanded real-time data.
OpenAI celebrated Codex reaching 3 million weekly users by resetting usage limits, promising to repeat the reset at every million-user milestone up to 10 million.
Google's Gemini can now turn questions into interactive visualizations mid-chat, complete with adjustable variables and rotatable 3D models, making it a hands-on learning tool.
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.
3 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
Reply to this email — I read every response