Is Anthropic Making the Biggest Mistake in AI History? | E2258
This episode discusses the explosive growth of OpenClaw, an open-source AI project that has become the most-starred project in GitHub history, and explores its implications for the AI industry's shift from proprietary models to open-source alternatives. The hosts debate Anthropic's controversial stance against military AI use and autonomous weapons, discuss the company's remarkable $20 billion revenue run rate, and showcase emerging applications and tools built around the OpenClaw ecosystem. The episode features insights from crypto entrepreneur Erik Voorhees, venture investor Logan Allin, and product demos highlighting how OpenClaw is democratizing access to AI agents for non-technical users.
Key takeaways
- • OpenClaw has surpassed React to become the most-starred open-source project in history, signaling that open-source AI models are meaningfully closing the gap with proprietary LLMs on intelligence, speed, and price.
- • Anthropic's decision to refuse Pentagon contracts over autonomous weapons and mass surveillance concerns was principled but politically risky, as the military retaliated by discouraging other companies from working with the company.
- • Anthropic is on track to generate nearly $20 billion in annual revenue, more than doubling from late 2025, with Claude Code alone accounting for $2.5 billion and enterprise relationships representing over 80% of revenue versus less than 5% for OpenAI.
- • Privacy-first AI platforms like Venice AI are gaining traction by offering uncensored, non-logging alternatives to mainstream chat platforms, with tokenomics that provide marginal-cost-free API access to users holding the platform's token.
- • AI agents visiting websites require new analytics tools to track behavior and attribution; Sighteline is emerging as "Google Analytics for agents" to help businesses capitalize on agentic traffic and commerce.
- • OpenClaw Studio simplifies agent orchestration for non-technical users through intuitive UI, permission controls, and scheduled automation, bridging the gap between developer-focused tools and mainstream consumers seeking AI assistance.
- • The future of AI adoption will be driven by user-friendly interfaces and security-first design rather than raw model capabilities, as executives and non-technical power users represent the largest untapped market for agentic AI.
Recommendations (7)
"I bought this new pen, this Zebra G-750. You know, it's an $8 pen, but it's really quite nice"
This Week in Startups · ▶ 1:15:45
"I do want to give a shout out to NotebookLM. It was one of the first AI applications that I ever got really obsessed with. I still love it. I still use it for research all the time"
Logan Allin · ▶ 1:19:28
Mentioned (11)
More from these creators
The 5-Step Framework for AI Agents That Improve While You Sleep | E2269
The $60 billion resource hiding in space, and the start trying to mine it | E2268
This Bittensor Subnet Could Cut Drug Discovery Costs in HALF | E2267
32 Million Dollar Fraud? Delve Scandal | E2266
The Drone Company Everyone Thought Was Illegal (Now Worth $4B+) | E2265
How Robinhood became a $68B company w/ Vlad Tenev