Anthropic's $30B Ramp, Mythos Doomsday, OpenClaw Ankled, Iran War Ceasefire, Israel's Influence
The All In Podcast discusses Anthropic's decision to withhold their powerful Mythos model due to cybersecurity risks, Anthropic's unprecedented $30B revenue ramp, OpenClaw being restricted by Anthropic, and ongoing Middle East conflicts including the Iran war ceasefire.
Summary
The episode begins with discussion of Anthropic withholding their new Mythos model, claiming it's too dangerous for public release as it can autonomously find thousands of vulnerabilities in major operating systems and web browsers, including decades-old exploits. The hosts debate whether this is legitimate safety concern or marketing theater, with Brad Gerstner defending Anthropic's approach of creating a 100-day sandboxing period with 40 major companies to patch vulnerabilities before release.
The conversation shifts to Anthropic's explosive revenue growth, reaching a $30 billion run rate by April 2025, representing the fastest revenue ramp in tech history. This growth is driven by enterprise customers, with over 1,000 companies paying more than $1 million annually. The hosts discuss the competitive dynamics between Anthropic and OpenAI, noting Anthropic's focus on coding capabilities while OpenAI appears to be struggling with internal issues.
A significant portion covers Anthropic's restriction of OpenClaw, an open-source coding agent, by forcing users to switch from subscription plans to more expensive API pricing while simultaneously releasing their own competing agent product. This raises antitrust concerns about bundling and price discrimination.
The episode concludes with discussion of Middle East conflicts, particularly the Iran war and ceasefire negotiations, with debate about Israeli influence on U.S. foreign policy. The hosts also praise X's new auto-translate feature for enabling cross-language communication and promoting their upcoming Liquidity summit.
Key Insights
- Anthropic claims their Mythos model can find thousands of vulnerabilities including 20+ year old exploits missed by security audits for decades
- The company established Project Glasswing with 40 major companies to spend 100 days hardening systems before releasing the dangerous capabilities
- Anthropic achieved a $30 billion revenue run rate by April 2025, representing the fastest revenue growth in technology history
- Over 1,000 enterprises are now paying Anthropic more than $1 million annually, demonstrating massive enterprise adoption
- Anthropic restricted OpenClaw users from using subscription plans, forcing them to expensive API pricing while releasing their own competing agent
- The hosts argue this represents potential antitrust issues around bundling and price discrimination in AI services
- Brad Gerstner invested in Anthropic at $130-150 billion valuation and now expects it could reach $600+ billion, representing a 5-7x return
- The TAM for intelligence is described as dramatically larger than any previous technology market they've seen
- OpenAI is reportedly preparing to release their 'SPUD' model which early users say is on par with Mythos capabilities
- The Middle East ceasefire represents efforts to prevent escalation traps, with JD Vance leading diplomatic negotiations
- Americans increasingly believe Netanyahu has too much influence over U.S. foreign policy decisions
- X's auto-translate feature is enabling real-time cross-language political discussions that were previously impossible
Topics
Full transcript available for MurmurCast members
Sign Up to Access