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

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 https://www.anthropic.com 0:00 Unix vs.

Mer att läsa

DeepSeek open sources Super-Agent as a research prototype, a proactive coding agent it says can monitor group chats for coding issues and suggest solutions

Techmeme · för 1 tim sedan

Insurance software maker Guidewire agrees to acquire Hyperexponential, which makes AI-powered insurance underwriting and pricing software, for $1.7B in cash

Techmeme · för 1 tim sedan

One word for AI agents in 2027

Nate Herk · för 5 tim sedan

Tests: chatbots rely on three obscure AI-built voter sites for US midterm answers; Copilot cited the error-prone Ballot Brief in 75% of tests and ChatGPT in 22%

Techmeme · för 6 tim sedan