770,000 Agents, 0 Humans: Inside the First AI Social Network
Episode
21 min
Read time
2 min
Topics
Productivity, Remote Work, Startups
AI-Generated Summary
Key Takeaways
- ✓Agent versus chatbot distinction: OpenClaw operates as an autonomous agent that completes tasks end-to-end without human intervention, contrasting with ChatGPT or Claude which require constant prompting. Users delegate outcomes like sending emails or managing calendars rather than describing step-by-step processes. This shift from conversational AI to delegative AI represents the fundamental transformation happening in 2026 across enterprise and personal productivity tools.
- ✓Viral growth mechanics: Endorsements from Andrej Karpathy and David Sacks in mid-January 2026 accelerated OpenClaw from a few thousand users to 100,000 GitHub stars by January 30. The project required three name changes due to trademark issues with Anthropic. This trajectory demonstrates how influential voices in AI can compress years of organic growth into weeks for developer tools and open source projects.
- ✓Real implementation examples: Eric Hsu uses OpenClaw to log into ClickFlow, produce content, flag SEO issues, and book meetings autonomously. A Belgian web agency deployed an agent that received a client request at midnight, updated the website menu independently, and notified the client of completion without human involvement. Non-technical users report running design companies and brewing beer operations through OpenClaw with zero coding knowledge.
- ✓Security deployment options: Avoid local installation due to vulnerabilities in the skills marketplace that exposed API keys and authentication credentials. Use hosted services like mycloud.ai for quick setup, xcloud.host for five-minute Telegram deployment, or enterprise solutions like Runlayer and Clari. DigitalOcean offers one-click deployment for developers. Connect an Anthropic Claude or OpenAI API key and start with single use cases before scaling.
- ✓Agent-to-agent knowledge networks: MoltBook reached 770,000 agents within two weeks, creating communities where agents share technical capabilities, surface bugs, and exchange automation strategies. Whether truly autonomous or pattern-matching from training data, this represents AI systems teaching each other skills through distributed knowledge sharing. This self-recursive improvement loop mirrors how OpenAI built Codex 5.3 using AI, signaling the emergence of an agentic internet infrastructure.
What It Covers
OpenClaw transformed from a weekend project in November 2025 to GitHub's fastest-growing open source project with 186,000 stars. This AI agent framework enables autonomous task execution through messaging platforms, spawning MoltBook, a social network with 770,000 AI agents. The movement reveals security challenges and previews an emerging agentic web architecture.
Key Questions Answered
- •Agent versus chatbot distinction: OpenClaw operates as an autonomous agent that completes tasks end-to-end without human intervention, contrasting with ChatGPT or Claude which require constant prompting. Users delegate outcomes like sending emails or managing calendars rather than describing step-by-step processes. This shift from conversational AI to delegative AI represents the fundamental transformation happening in 2026 across enterprise and personal productivity tools.
- •Viral growth mechanics: Endorsements from Andrej Karpathy and David Sacks in mid-January 2026 accelerated OpenClaw from a few thousand users to 100,000 GitHub stars by January 30. The project required three name changes due to trademark issues with Anthropic. This trajectory demonstrates how influential voices in AI can compress years of organic growth into weeks for developer tools and open source projects.
- •Real implementation examples: Eric Hsu uses OpenClaw to log into ClickFlow, produce content, flag SEO issues, and book meetings autonomously. A Belgian web agency deployed an agent that received a client request at midnight, updated the website menu independently, and notified the client of completion without human involvement. Non-technical users report running design companies and brewing beer operations through OpenClaw with zero coding knowledge.
- •Security deployment options: Avoid local installation due to vulnerabilities in the skills marketplace that exposed API keys and authentication credentials. Use hosted services like mycloud.ai for quick setup, xcloud.host for five-minute Telegram deployment, or enterprise solutions like Runlayer and Clari. DigitalOcean offers one-click deployment for developers. Connect an Anthropic Claude or OpenAI API key and start with single use cases before scaling.
- •Agent-to-agent knowledge networks: MoltBook reached 770,000 agents within two weeks, creating communities where agents share technical capabilities, surface bugs, and exchange automation strategies. Whether truly autonomous or pattern-matching from training data, this represents AI systems teaching each other skills through distributed knowledge sharing. This self-recursive improvement loop mirrors how OpenAI built Codex 5.3 using AI, signaling the emergence of an agentic internet infrastructure.
Notable Moment
An Austrian developer described his desired functionality to AI, which built a working OpenClaw prototype in one hour without him writing code. Three months later, that weekend project has spawned an ecosystem including enterprise spinoffs, global meetups called Claw Crew, hardware shortages from Mac Mini purchases, and companies hitting one million ARR within hours of launch.
Episode Transcript
I'm gonna go behind the scenes on the biggest AI movement on the web, a weekend project that became the fastest growing open source project in GitHub's history, a social network with with 770,000 users, but the catch is those users are agents and humans aren't allowed to post, a security nightmare that companies are scrambling to contain and buried under all of this chaos is a genuine preview of how an agentic first web might look. This is the story of Open Claw and mold book, one you cannot afford to miss. So stay with me for the the next twelve to fifteen minutes because I'm gonna give you the complete story, how it was built, how it created its own social network, the security nightmare that companies are having at the moment to try to wrangle this in, and how you can get started with your own open claw in a very safe and quick manner. All of that and more on this episode of Marketing Against the Grain. Before we get into today's show, here's a quick word from HubSpot. Cutting your sales cycle in half sounds sounds pretty incredible, but that's exactly what Sandler Training did with HubSpot. They used Breeze, HubSpot's AI tools, to tailor every customer interaction without losing their personal touch, and the results were pretty incredible. Click through rates jumped 25%, qualified leads quadrupled, and people spent three times longer on their landing pages. Go to hubspot.com to see how Breeze can help your business grow. So in this episode, we are gonna talk about the lobster that has eaten the Internet, and that lobster is called Open Claw. Now let's start at the start. Where did this all originate? Well, in November 2025, an Austrian developer named Peter Steinberger, he had a simple idea. He wanted to take an AI model, and he wanted to integrate it into Telegram so he could message his model and have it actually do some things. He didn't want a chatbot. He wanted a true agent that could do work for him. Now, he previously built and sold a company called PSPDFKit, what a great brand name, for a lot of money. It was developer tools. He sold it globally. He's known for shipping software. But for this little kinda hack that he did, he didn't want to code himself. He described something that he wanted, and AI built it. It actually built a prototype in in a single hour, and that prototype was an agent called Claude, c l a w d. It was a pun on Entropic's Claude, and he open sourced it. And for two months, it grew pretty quietly. It's been out since November 2025. A few thousand developers tried it. It had its own Discord community. And then in mid January twenty twenty six, so very recently, something actually changed and it took off like a rocket ship. And what happened was Andrej Karpathy, one of the most respected AI …
Get the full transcript (3,696 words) + summary by email — free
One-time email with the complete transcript and AI summary of this episode. No account needed.
One email, no spam. We’ll also show you what SignalCast does.
You just read a 3-minute summary of a 18-minute episode.
Get Marketing Against the Grain summarized like this every Monday — plus up to 2 more podcasts, free.
Pick Your Podcasts — FreeKeep Reading
Books, tools, and gear mentioned in this episode
SignalCast may earn commission on purchases via these links.
Tools
“Use hosted services like mycloud.ai for quick setup, xcloud.host for five-minute Telegram deployment, or enterprise solutions like Runlayer and Clari.”
“Eric Hsu uses OpenClaw to log into ClickFlow, produce content, flag SEO issues, and book meetings autonomously.”
“SPONSORS [{'name': 'HubSpot', 'url': 'hubspot.com'}]”
“Use hosted services like mycloud.ai for quick setup, xcloud.host for five-minute Telegram deployment, or enterprise solutions like Runlayer and Clari.”
“OpenClaw transformed from a weekend project in November 2025 to GitHub's fastest-growing open source project with 186,000 stars. This AI agent framework enables autonomous task execution through messaging platforms, spawning MoltBook, a social network with 770,000 AI agents.”
“OpenClaw transformed from a weekend project in November 2025 to GitHub's fastest-growing open source project with 186,000 stars. This AI agent framework enables autonomous task execution through messaging platforms.”
“Use hosted services like mycloud.ai for quick setup, xcloud.host for five-minute Telegram deployment, or enterprise solutions like Runlayer and Clari.”
“Use hosted services like mycloud.ai for quick setup, xcloud.host for five-minute Telegram deployment, or enterprise solutions like Runlayer and Clari.”
More from Marketing Against the Grain
We summarize every new episode. Want them in your inbox?
The death of the CMO
Turn Every Comment Into a Lead - Automate DMs with Claude
Using Codex to build a Research Dashboard/System
Claude vs. Human Marketer: Which Is Actually Better?
How to Stop AI from Killing Your Critical Thinking w/ David Epstein
Similar Episodes
Related episodes from other podcasts
The AI Breakdown
Feb 16
OpenClaw Goes to OpenAI
Latent Space
Jun 2
GitHub's plan for Agents — Kyle Daigle, GitHub
This Week in Startups
Mar 5
Is Anthropic Making the Biggest Mistake in AI History | E2258
This Week in Startups
Mar 3
How the OpenClaw foundation bullet-proofed its future (w/Dave Morin) | E2257
This Week in Startups
Feb 17
Will OpenAI Tank OpenClaw? | E2251
Explore Related Topics
This podcast is featured in Best Marketing Podcasts (2026) — ranked and reviewed with AI summaries.
Read this week's Startups & Product Podcast Insights — cross-podcast analysis updated weekly.
You're clearly into Marketing Against the Grain.
Every Monday, we deliver AI summaries of the latest episodes from Marketing Against the Grain and 192+ other podcasts. Free for one show.
Start My Monday DigestNo credit card · Unsubscribe anytime