How Clay runs 350 million GTM agents a month | Interrupt 26
Jeff Barg, Head of AI at Clay, breaks down what it actually takes to run go-to-market agents at production scale — not just one agent, but 350 million a month across an entire addressable market. He covers the four hard problems Clay solved: infrastructure reliability, throughput under spiky workloads, cost (including a 70% reduction via caching), and agent quality. He also introduces Audiences, Clay's new product for giving agents the context they need to recommend plays autonomously. Chapters: 0:00 What Clay does and why GTM is an agent problem 0:55 350 million agents a month: Clay's scale 1:10 Why no creative advantage lasts forever 1:44 How to actually win: the fastest to iterate wins 2:00 Go-to-market alpha: the three levels 3:16 Why most teams stay stuck at level one 3:47 The loop Clay's best customers run 4:18 Why this looks like an engineering challenge 4:47 Four challenges at production scale 5:25 Challenge 1: infrastructure and durable workflow execution 6:21 Challenge 2: rate limits and the TCP/IP approach to throughput 7:30 Challenge 3: cost and caching strategies 8:36 Challenge 4: quality, context, and evals 9:39 What's next: Audiences and agent memory 11:12 Recap Extra resources: • Everything we shipped at Interrupt: https://www.langchain.com/blog/interr... • Meet LangSmith Engine: https://www.langchain.com/blog/introd... • About LangChain: https://www.langchain.com/

The Future of AI Agents with Andrew Ng | Interrupt 26

How Model Context Protocol (MCP) actually works

Stop Prompting Claude. Use Karpathy's Method Instead.

MIT Just Revealed the AI Bubble's Fatal Flaw

What Nobody Tells You About Being a Quant

The Agent Development Lifecycle: Build, Test, Deploy, Monitor | Interrupt 26

How I deleted 95% of my agent skills and got better results — Nick Nisi, WorkOS

How Lyft Builds Evals That Actually Matter in Production | Interrupt 26

Pick The RIGHT CI/CD Strategy for Your Team (Most Engineers Get This Wrong)

The best AI agents are simpler than you think

How Instagram Scaled Postgres to 2 Billion Users

Andrej Karpathy's Wiki Idea Was Just Shipped by Pinecone

Is ontology replacing the semantic model?

Are we really doing this again

I Made Opus 4.8 and Fable 5 Build the Same App (RAW RESULTS)

Anthropic's Boris Cherny: Why Coding Is Solved, and What Comes Next

Billionaire's WARNING: I'm SELLING. The Crash Is Already Here!

LangGraph + Pipecat: Voice Agents, Interruption Handling, and Full LangSmith Tracing

Full Walkthrough: Workflow for AI Coding — Matt Pocock

