Hi all,

Here is this week's bonus edition. The weather in NYC is great, New York Tech Week is right around the corner, and a lot keeps happening in the agent space as usual.

Hope you enjoy your weekend!

Cheers,
Ivan

Invite a friend to join AGT NYC at agtnyc.com!

Events

• Jun 3 - AI VIPs: Enterprise Edition #NYTechWeek - hosted by Topher Price (from our Dec '25 event!)
• Jun 2 - Inside the Agent Engine: A Deep Dive with Traversal and LangChain
• Jun 4 - AI Agents: How To Keep Pace With Rapid Change
• Jun 23 - Daytona AI Builders w/ Oracle & Datadog
• Jun 24 - Agentic Alpha: Engineering High-Velocity Autonomy in a Millisecond Economy

You can find more events on the AGT NYC Luma calendar.

Also, a NYTW-specific list is coming next.

News

• Apple is working to incorporate AI agents on the App Store
• Companies Have a New AI Problem: Too Many Agents
• NSA joins the ASD’s ACSC and Others to Release Guidance on Agentic Artificial Intelligence Systems
• Sam Altman backs “micropayment” model for AI agents to compensate publishers
• Fujitsu develops self-evolving multi-AI agent technology that learns and adapts to business operations

Once the big slow players like Apple start making moves around agents, it's probably a sign the tech has started to mature. I think agents only get more sophisticated from here.

Launches

• The Gemini app becomes more agentic
• Notion launches AI agent hub
• Agents can now create Cloudflare accounts
• Everything shipped at LangChain Interrupt 2026
• Announcing Cofounder 2: Run an entire company with agents

Just these five launches cover consumer, professional, enterprise, and developer markets. Agents are everywhere, changing everything.

Deals

• Sierra ($950M) - customer experience agents
• Exaforce ($125M) - security operations agents
• Standard Intelligence ($75M) - computer use agents
• XBOW ($35M) - offensive security agents
• NanoCo ($12M) - secure OpenClaw alternative
• Jit acquired by Torq
• Symmetry Systems acquired by Zscaler

I see a few trends above: increased acquisitions, emphasis on security, and continued funding towards general-purpose agents. Sierra in particular has broadened its focus recently - it might position itself as "the enterprise agent company" in the near term.

Articles

• Computer use is 45x More Expensive Than Structured APIs by Reflex
• I don’t think we are close to “AI scientists” in Understanding AI
• Securing the agentic enterprise: Opportunities for cybersecurity providers by McKinsey
• 4 in 10 AI agents headed for demotion or the rubbish bin in The Register
• Most AI Agents Fail in Production Because They’re Built Backwards in Towards Data Science

While the sentiment in the pieces above leans sour, I think that's just the outcome of an overhyped first wave of agent products. As the industry sobers up and prioritizes security, efficiency, and quality, we'll likely see fewer articles about disappointing results.

Projects

• Agent Experience Interface - agent design principles
• Webwright - simple browser agent
• Flue - agent harness framework
• Cline SDK - agent application framework
• Adam - embeddable agent library

This is a fun roundup: a database project releases an agent library, another coding agent company moves their core engine to a framework, a web framework company builds a harness framework, and Microsoft releases something simple for once.

Learning

• How to evaluate AI agents
• Why Agentic AI Fails: Infinite Loops, Planning Errors, and More [12:44]
• Multi-Agent Collaboration in Production: Lessons from 500,000+ Agent Deployments
• Agents for Everything Else — swyx [14:09]
• Agent Evaluation: A Detailed Guide

Encouraging to see so much high quality material out there covering real world experience with agents. The knowledge keeps advancing and the products keep getting better.

Research

• Anthropic - Project Deal
• Your Agents Are Aging Too: Agent Lifespan Engineering for Deployed Systems
• AutoScientists: Self-Organizing Agent Teams for Long-Running Scientific Experimentation
• SkillOpt: Executive Strategy for Self-Evolving Agent Skills
• LiveBrowseComp: Are Search Agents Searching, or Just Verifying What They Already Know?

Though the headlines imply that agents are both self-improving and eternal, the reality is that even the virtual robots are faulty and imperfect, like their creators. It may be a long time before they're as good as we want them to be.

Comments, suggestions? Reply to this email, let me know what you think!