Hi all,

Real NYC spring is almost here, now that the calendar is inching towards May. I'm ready for six weeks of good weather before it gets too hot.

The agent space moves fast. OpenClaw was the big breakout project in January and now the hype has calmed down while alternatives have emerged from all sides. It was probably optimistic to think that a vibe-coded hacky proof-of-concept was going to have staying power, but one thing that OpenClaw (and Moltbook) has accomplished is prove that the idea works.

More AGT stuff is coming soon - in the meantime, check out the links below!

Cheers,
Ivan

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

Events

• Apr 28 - The Agentic Bank: Closing the Gap Between Insight and Action
• Apr 29 - Zero to Agent: New York
• Apr 29 - Future-Proofing Healthcare: Turning Agentic AI Innovation and Evolving Regulations into Operational Advantage
• Apr 29 - Building Better Agents: The Agent Improvement Loop
• Apr 29 - Get Your Claude on Steroids
• Apr 30 - Multi-Agent Hackathon
• Apr 30 - Agentic by Eigen
• Apr 30 - Get Your Claude on Steroids V2.0
• May 2 - Beer and Multi-Agent System Coding
• May 4 - AI Agent Conference
• May 4 - KiloClaw Happy Hour
• May 4 - AI Agents Happy Hour with Veris AI and Sondera AI
• May 4 - Autonomous AI Techstars Happy Hour
• May 4 - AI Agent Conference: After Hours with LangChain + Baseten
• May 5 - AI Engineers Happy Hour at Spin 🏓 during AI Agent Conference
• May 5 - Agent builders night
• May 6 - Prototype to Production: Agentic AI in Financial Services
• May 7 - OpenClaw and Beyond
• May 14 - Accelerating Agentic Orchestration with Intelligent Code Assistants and Real‑Time Data Integration
• May 27 - Agents and MCP for PostgreSQL

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

News

• Silicon Valley’s AI agent hiccups: Wasted tokens and ‘chaotic’ systems
• Human scientists trounce the best AI agents on complex tasks
• AI Agents are Coming for Your Dating Life
• No humans allowed: scientific AI agents get their own social network
• Volkswagen to equip Chinese cars with AI agents

Agents are everywhere and they're causing problems. Now would be a good time to think about how to deal with them as they proliferate.

Launches

• Poke
• Cloudflare Agents Week
• Introducing Gemini Enterprise Agent Platform
• Claude Managed Agents: get to production 10x faster
• Introducing Salesforce Headless 360. No Browser Required.

There have been some impressive agent launches recently. One trend that stands out is big companies launching managed agent platforms - way too early to tell who's going to win, but the lack of coordination and collaboration here makes me think this will end up like the agent framework wars, with lots of churn and reinvention.

Deals

• NeoCognition ($40M) - agent research lab
• Synera ($40M) - manufacturing agent platform
• Tasklet ($20M) - agent deployment platform
• BAND ($17M) - multi-agent coordination
• Trent AI ($13M) - security agent platform
• Insight Health ($11M) - clinical AI agents
• Octen ($10M) - agent search engine

Cool to see big rounds for companies that aren't only focused on verticals. This recent batch of raises has been less flashy, with more interesting ideas.

Articles

• AI Agents as Users by NN/g
• Cloud native agentic standards by CNCF
• AI’s Next Operating Model by Bain
• Building the foundations for agentic AI at scale by McKinsey
• Trustworthy agents in practice by Anthropic

Agents as trustworthy users that follow standards will be interesting. Right now trust is limited naturally (sandboxes, permissions) and standards are still evolving (if they're considered at all).

Projects

• Mercury - soul-driven agent
• OpenHarness - open agent harness
• open-multi-agent - multi-agent orchestration
• hippo-memory - biological agent memory
• Bindu - agents as microservices
• [Hermes Atlas] - Hermes community map

Everyone wants to be the next OpenClaw, in adoption if not in design. There's also a push to generalize away from closed-source harnesses like Claude Code now that people realized it's not that complicated of a product.

Learning

• Scaling Managed Agents: Decoupling the brain from the hands
• coSTAR: How We Ship AI Agents at Databricks Fast
• A Practical Guide to Memory for Autonomous LLM Agents
• How To Build a General Purpose AI Agent in 131 lines of Python
• Memory Scaling for AI Agents
• RL for Agents Workshop - Deep Dive on Training Agents with RL and Open Source (1:54:08)

Some hard-earned production experience is making its way into reports and tutorials, which is great. It's possible this wave of agent tech is almost figured out and the next wave is on the horizon.

Research

• PaperOrchestra: A Multi-Agent Framework for Automated AI Research Paper Writing
• Gym-Anything: Turn any Software into an Agent Environment
• The Art of Building Verifiers for Computer Use Agents
• AlphaEval: Evaluating Agents in Production
• ReasoningBank: Scaling Agent Self-Evolving with Reasoning Memory

Just these five papers could add up to an entire startup. As usual, lots of good research coming out about agents in practice. It feels like fairly soon, agents will be too sophisticated to fully understand.

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