Hi all,

We're almost halfway through the year. New York Tech Week starts today, with a ton of agent-related events - I picked some I thought would be a good fit for AGT. See you around the city!

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 1 - ERA/RV Portfolio Agentic AI Demo Showcase - #NYTechWeek
• Jun 1 - Agentic Voice Hackathon - #NYTechWeek
• Jun 1 - AI Agents in Finance - The Next Wave of Adoption #NYTechWeek
• Jun 1 - AI x Investing: The Age of Agents #NYTechWeek
• Jun 1 - AI Agents for Allocators and Asset Managers - #NYTechWeek
• Jun 2 - Build Agent-first Companies with Hyperagent #NYTechWeek
• Jun 2 - Building & Entrepreneurship in The Agent Era - Cloudflare x Shopify #NYTechWeek
• Jun 2 - Agentic Commerce: Build Day - #NYTechWeek
• Jun 2 - 🗽Agents & APIs Demo Night #NYTechWeek
• Jun 2 - No AI Team, No Problem: The Multi-Agent Playbook For Teams & Enterprises. - #NYTechWeek
• Jun 2 - How We Build with Agents - #NYTechWeek
• Jun 2 - Creator Hackathon: Build the Next AI Agent That Finds Your Next Brand Deal - #NYTechWeek
• Jun 2 - Building Safe Agents for Business - #NYTechWeek
• Jun 2 - Agent Era: How Women Founders Are Building AI That Works for Their Business - #NYTechWeek
• Jun 3 - Camp AI: Agents at Work #NYTechWeek
• Jun 3 - AI Agents: From Prototype to Production - #NYTechWeek
• Jun 3 - Agents in Production: The Memory Problem - #NYTechWeek
• Jun 3 - Voice Agents Panel - #NYTechWeek
• Jun 3 - Agentic Finance Summit
• Jun 4 - Autonomously Improving Agent Swarms with Weights & Biases - #NYTechWeek
• Jun 4 - Agents at Scale - #NYTechWeek
• Jun 4 - Building with AI Agents in Finance Operations - #NYTechWeek
• Jun 4 - Proof of Build - An agent building hackathon by Lyzr #NYTechWeek
• Jun 4 - Build Agents with TheVibeFounder - #NYTechWeek
• Jun 5 - Bare Metal Happy Hour: Autonomous Systems & Agent Fleet Orchestration - #NYTechWeek
• Jun 5 - Building AI Agents for Science - #NYTechWeek
• Jun 6 - Agents Hackathon by Lightning AI & Validia - #NYTechWeek
• Jun 6 - Multimodal Hacks: build the interface for agents - #NYTechWeek
• Jun 6 - We live in a simulation - #NYTechWeek

Find more events on the AGT NYC Luma calendar.

News

• Robinhood now lets your AI agents trade stocks
• Anthropic’s Claude Managed Agents can now “dream,” sort of
• Intercom, now called Fin, launches an AI agent whose only job is managing another AI agent
• Amazon WorkSpaces now lets AI agents operate desktop applications
• BCG trained an AI agent on the worst things salespeople do

I wonder if we'll look back on headlines like these next year and see a mania around agents that was a bit ahead of its time. Not everything is ready yet.

Launches

• The Figma design agent is here
• Introducing Inherent: a lab designed from scratch to build AI agents that discover new knowledge
• Agents that transact: Introducing Amazon Bedrock AgentCore payments, built with Coinbase and Stripe
• SAP launches AI Agent Hub: a vendor-agnostic command center to inventory and govern AI agents, LLMs, and MCP servers
• Upwind launches Agentic Pack: blue, red, and green AI agents for security

The theme here is "laggards catching up." Whether startups or software old guard, everyone is repositioning to capture the agent wave.

Deals

• Viktor ($75M) - professional AI coworker
• Pace ($46M) - insurance operations agents
• Dust ($40M) - agent operating system
• Geordie AI ($30M) - agent governance platform
• General Analysis ($10M) - agent security infrastructure

Funding for agent companies continues undisturbed. It's more mature companies this time around, and the round sizes are less gigantic than the broader AI space.

Articles

• Call your AI agent in Nature
• Why Coase needs Hayek by Rohit Krishnan
• AI in SRE: Where and how Google is deploying agentic AI to improve operations by Google
• Teams of AI agents boost speed of research in Nature
• AI Agents Forecast to Boost Tech Cash Flow as Usage Soars by Goldman Sachs

Usually there's more criticism than approval from those who write about agents, but this time around the reception is more measured and level-headed. Even though we're still figuring out how to build agents, some really are producing useful results.

Projects

• ax - distributed agent runtime
• HALO - agent loop optimization
• browse - agent browser CLI
• forge - tool reliability layer
• Autodidact - self-evolving agent

What's interesting about this group is that they're all innovating on different parts of the agent space. Some are improving the internals, others are rethinking what an agent should be oriented around, and yet others are considering the layers above.

Learning

• What I Learned Building Multi-Agent Systems from Scratch [35:28]
• MemEx: A Programmable Scratchpad for LLM Agents
• Even (very) noisy LLM evaluators are useful for improving AI agents
• From 46% to 90%: Fine-Tuning Tiny LLMs for On-Device Agents — Cormac Brick, Google [21:00]
• Startup School: Agentic AI

It feels like we're continuing to map out what a proper agent architecture looks like, not having reached the full picture yet. There is still a lot to learn.

Research

• Polar: Agentic RL on Any Harness at Scale
• Why You Shouldn’t Treat AI Agents Like Employees
• SocialReasoning-Bench: Measuring whether AI agents act in users’ best interests
• LinuxArena: A Control Setting for AI Agents in Live Production Software Environments
• Multi-User Large Language Model Agents

A variety of topics here, but I do think it's interesting how the relationship between agents and their users is becoming more important. These things are unlike anything we've dealt with in technology so far, so we should adjust our expectations.

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