State of the Claw — Peter Steinberger

137.4K views · Apr 17, 2026 · 44:12 min · Watch on YouTube ↗

Takeaway

Open Claw's hyper-growth has made it the largest target for AI-generated security reports, forcing the project to rebuild around hardening and a healthier maintainer bus factor.

Summary

Peter Steinberger, creator of the open-source Open Claw project (an open Claude/agent CLI), reports it is the fastest-growing GitHub project in history at five months old with ~30k commits, ~2k contributors, and 30k PRs.
Project security is dominated by AI-generated reports: 1,142 advisories (16.6/day, 99 critical), roughly 2x Linux kernel and 2x curl rates, with most 'critical' reports turning out to be AI-slop.
Steinberger now juggles two roles — OpenAI engineer and Open Claw Foundation lead — with contributors from Nvidia, Microsoft, Red Hat, Tencent and Bytedance, where Chinese users are the largest cohort.
Argues AI tools will break all existing software by surfacing weird multi-chain exploits, and points to vendors like Nvidia Nemo Claw as a sandbox/plug-in security layer for Open Claw.

open-sourceagentssecurity

Original description

Peter Steinberger gives the 5 month update on OpenClaw, the fastest growing open source project in history, and what it's like as a maintainer, from security to community. Keynote followed by audience Q&A moderated by @swyx.

Speaker info:
- https://x.com/steipete
- https://www.linkedin.com/in/steipete/
- https://openclaw.ai/


Timestamps
0:00 Project Growth and Statistics
2:23 Management Challenges and the OpenClaw Foundation
3:47 Addressing Security Advisories and Vulnerabilities
10:33 Misinformation and Media Fearmongering
14:50 The Burden of Open Source Maintenance
16:12 OpenAI Involvement and Future Independence
18:57 Audience Q&A Begins
19:53 OpenClaw's Relationship with OpenAI
22:28 The Importance of Open and Local Models
24:57 Coding Workflow and Agent Interactions
28:28 Defining 'Taste' in AI Development
30:31 Developing Personality for AI Agents
33:22 Future Vision: Ubiquitous Agents and Smart Homes
35:58 Addressing Prompt Injection Risks
38:33 Future Vision: Implementing 'Dreaming' and Modularity
40:24 Life as a Maintainer and Future Skills