Why One AI Agent Isn’t Enough (And How to Run 1,000) - CJ Avilla & Harrison Stall, Anthropic @aiDotEngineer
Why One AI Agent Isn’t Enough (And How to Run 1,000) - CJ Avilla & Harrison Stall, Anthropic  @aiDotEngineer
Uploaded October 2026 | Updated November 2026, 4 days ago
Ken Thompson built Unix in a month. IBM put 1,000 engineers on OS/360. And in 1998, Linus Torvalds hit his limit. Now we're hitting that same wall managing AI agents.

Anthropic's CJ Avilla and Harrison Stall introduce Flows in Claude Managed Agents. You hand Claude a goal, and it plans and runs its own multi-agent topology of up to 1,000 agents through the API. In this hands-on workshop they demo a stock watchlist scanner, show how to spend more compute on the phases that matter, walk through the Bug Hunter and Explainer Studio templates, and explain why single agents fail through laziness, self-preference bias and goal drift. They finish with a deep dive on running agents safely at scale: budgets, interrupts, auto mode, credential vaults and network lockdown.

Speakers:
CJ Avilla, Member of Technical Staff, Anthropic
X: https://x.com/cjav_dev

Harrison Stall, Member of Technical Staff, Anthropic
X: https://x.com/HarrisonStall

anthropic.com

0:00 Unix vs. OS/360: the scaling problem
6:00 Meet CJ & Harrison
7:00 Introducing Flows in Claude Managed Agents
9:05 Why single agents hit a wall
14:00 How dynamic workflows work
17:25 Demo: stock watchlist scanner
23:20 From console to production: tuning each phase
32:15 Q&A, round 1
1:04:45 Bug Hunter: multi-agent PR review
1:07:15 Explainer Studio: flows that make videos
1:12:15 Laziness, self-bias & goal drift
1:15:20 Agent safety at scale
1:20:05 Declare, contain, detect
1:21:15 Separating the brain from the hands
1:27:50 Credential vaults & locking down tools
1:31:45 Wrap-up + Q&A, round 2

Live from the Sheraton New York Times Square Hotel in New York City, AI Engineer New York 2026 is an AI conference for engineers and technical leaders building production AI across banking, hedge funds, fintech, insurance, and asset management.

Event: AI Engineer NYC
Dates: October 12–14, 2026
Venue: Sheraton New York Times Square Hotel

Learn more about the event:
https://ai.engineer/nyc/2026#

Subscribe for more AI Engineer talks, livestreams, and event coverage:
youtube.com/@aiDotEngineer
Why One AI Agent Isn’t Enough (And How to Run 1,000) - CJ Avilla & Harrison Stall, AnthropicAI Writes the Code. So Whats Left for Engineers? - Laurie Voss, Arize AIThe Agents Were Fine. Everything Around Them Broke. - Joey Wang, HarveyYour Code Has Bugs. Lean4 Has Proofs: Formal Verification for Engineers — Varun Pant, AWSBuying GPUs Should Work Like Buying Stocks — Carmen Li, Compute ExchangeAIE Singapore Day 2 ft. Google DeepMind, OpenClaw, Adaption, Arize, Cloudflare, Robot Company & moreThe Next Trillion Internet Users Still Don’t Have an Identity — Adi Singh, AgentMailWe Trust AI Agents That Break Every Day - Shy Ruparel, Temporal1,368 AI Failures and What They Teach About Evals — Vishal Srivastava, New York Life & Tanmay SahVideo Has No Memory. Heres How We Built One. — James Le, TwelveLabsHow to Get Frontier Coding Results at 70% Lower Cost - Nouha Dziri, CohereHow to Build a Software Factory That Improves Itself - Zach Lloyd, Warp
AI Engineer |

Why One AI Agent Isn’t Enough (And How to Run 1,000) - CJ Avilla & Harrison Stall, Anthropic

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER