Hi all,

Spring is almost over! A lot has been happening in the agent space lately, including a ton of upcoming events. This week I'll send out multiple editions since (as always) there is too much to fit into one newsletter.

Cheers,
Ivan

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

Events

• May 26 - Scaling Enterprise AI Agents: Security, Privacy & Real-World Deployment
• May 27 - ACM Conference on AI and Agentic Systems
• May 27 - Agents and MCP for PostgreSQL
• May 28 - Voice Agent Build Night with ElevenLabs and Tavily
• May 28 - Building Agentic Marketing: Foundations, Real Cases, and Getting Started
• May 28 - Build a Bilingual Multi-Agent Voice AI System with Vapi
• June 1 - ERA/Remarkable Ventures Agentic AI Demo Showcase
• June 2 - Agentic Commerce: Build Day @ NY Tech Week
• June 2 - Agents & APIs Demo Night
• June 2 - Agents of Change
• June 3 - Interrupt Resume NYC
• June 3 - Camp AI: Agents at Work
• June 3 - The Situation Room: Agent Autonomous
• June 4 - Agentic AI Summit
• June 4 - Agents at Scale
• June 4 - SOSV Presents: The Agentic VC
• June 12 - Alpha | NYSE : The Agentic Enterprise Summit
• June 16 - AWS Summit New York Welcome Party: Agents at Scale
• June 17 - AWS Summit NYC w/ Rootly AI, Cloudsmith, Mend.io, Clickhouse and Docker
• June 23 - Executive Dinner: Leveraging AI Agents

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

News

• Citi introduces Arc, an AI agent platform for use across the firm
• The Economist prepares for a two‑track internet: one for humans and one for AI agents
• Anthropic Releases New AI Agents for Financial Services Firms
• NVIDIA and ServiceNow Partner on New Autonomous AI Agents for Enterprises
• We gave an AI a 3 year retail lease in SF and asked it to make a profit

Internal agents with autonomy are an interesting development, an outline of the emerging "engine" of organizations. Early experience has been mixed, but clearly the interest in making internal agents successful is there.

Launches

• Plurai introduces the first vibe-training platform for evals and guardrails
• Infinitus Launches Studio, the First Healthcare-Specific No-Code AI Agent Builder
• Guild.ai Introduces the First Control Plane for AI Agents
• SoundHound OASYS: the Orchestrated Agent System for AI in the Real World
• Fiserv Launches agentOS: The Operating System for Agentic AI in Banking

Sometimes launches are split between brand new no-name companies and the old guard coming out with completely unexpected products. I take it as a sign that tech is moving in the same direction no matter the vertical.

Deals

• Adapt ($10M) - agents for business
• Cloneable ($4M) - agents for infrastructure
• NudgeBee ($3M) - cloud operations agents
• Brev ($3M) - agents for work
• Quantro ($2M) - cyber defense agents

The funding keeps coming and the ideas are much more grounded lately. Huge rounds for ambiguous missions aren't as common as they used to be.

Articles

• Your harness, your memory by LangChain
• Long-running Agents by Addy Osmani
• Why AI Engineers Are Moving Beyond LangChain to Native Agent Architectures in Towards Data Science
• How to Design for Human-agent Interaction in Every
• AI agents may be skilled researchers—but not always honest ones in Science

We are on a glide path to a complete understanding of how the current generation of LLM agents works, what their limitations are, and how to build around those limitations. It's likely that a proper abstraction (or two) for agents will emerge soon that integrates all the lessons learned so far.

Projects

• Coral - SQL for agents
• Motus - agent serving project
• Adala - autonomous data labeling
• OpenSwarm - Claude Code for X
• Mirage - agent file system

It's interesting to see the interfaces useful for agents end up being the standards used in tech for decades, like SQL and file systems. We're really building towards a fully virtualized machine user.

Learning

• Learn Harness Engineering
• AI Agents Vibe Coding Course from Google and Kaggle
• Background Agent Summit 2026
• Build Long-running AI agents that pause, resume, and never lose context with ADK
• Building Physical AI agents with MCP and MQTT on AWS IoT Core

There's still a lot to learn about the latest approaches for building agents. The larger field has started specializing and I wouldn't be surprised if there are niche agent disciplines before long.

Research

• HeavySkill: Heavy Thinking as the Inner Skill in Agentic Harness
• Recursive Multi-Agent Systems
• Synthetic Computers at Scale for Long-Horizon Productivity Simulation
• Autodata: an automatic data scientist to create high-quality data
• MegaFlow: Large-Scale Distributed Orchestration System for the Agentic Era

Research into making agents self-sufficient and self-improving continues, with constant improvements. Agents are also increasingly more reliable in production and have graduated from experiments to dependable systems.

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