The All-In Podcast
Anthropic’s $30B Ramp, Mythos Doomsday, OpenClaw Ankled, Iran War Ceasefire, Israel's Influence
Episode Summary
AI-generated · Apr 2026AI-generated summary — may contain inaccuracies. Not a substitute for the full episode or professional advice.
This All-In Podcast episode features guest Brad Gerstner joining Chamath Palihapitiya, Jason Calacanis, and David Sacks for a wide-ranging discussion covering the latest in AI, venture capital, and geopolitics. The central theme revolves around the rapid, often startling, advancements in artificial intelligence and their profound implications for cybersecurity, market dynamics, and international relations. The panel delves into Anthropic's controversial decision to withhold its Mythos model due to its advanced hacking capabilities, examining whether this is a genuine safety measure or a calculated marketing tactic.
Key discussions include the specifics of Anthropic's Mythos model, which reportedly found thousands of software vulnerabilities, including decades-old exploits missed by automated tools. Gerstner, an investor in Anthropic, defends the company's decision as a responsible act of self-regulation, highlighting their Project Glass Wing initiative to harden internet infrastructure. Sacks, however, points out Anthropic's historical pattern of using "scare tactics" in product marketing, though he concedes this particular cyber threat appears more legitimate. Chamath dismisses much of it as "theater," comparing it to OpenAI's GPT-2 rollout, and expresses skepticism about how much can truly be accomplished in Anthropic's proposed 100-day sandboxing period.
The conversation shifts to Anthropic's unprecedented revenue ramp, which has reached a $30 billion annual run rate, demonstrating a massive and unexpected surge in AI adoption, particularly among enterprise customers. The panel debates the profitability of these AI companies, with Gerstner suggesting that their burn rates might be lower than perceived due to compute constraints and exploding gross margins. They also discuss the intense competition in the AI agent space, with a specific focus on Anthropic's alleged "ankling" of the open-source OpenClaw project, potentially to establish market dominance in agent technology.
Finally, the episode addresses the evolving geopolitical landscape, specifically the Iran-Israel conflict and a recently brokered two-week ceasefire. The hosts discuss former President Trump's role in the ceasefire negotiations and controversial tweets, as well as the perceived influence of Israeli Prime Minister Netanyahu on U.S. foreign policy. Sacks emphasizes the importance of de-escalation, while Palihapitiya notes that markets are pricing in a quick resolution, and Israel needs to secure an off-ramp to preserve its long-standing alliance with the U.S.
Listeners will walk away with a nuanced understanding of the delicate balance between AI innovation and safety, the explosive growth and competitive pressures within the frontier AI model market, and the complex geopolitical forces shaping global conflicts, viewed through the lens of market reactions and political strategy.
👤 Who Should Listen
- AI founders and investors interested in the competitive landscape of frontier models and AI agents.
- Cybersecurity professionals and IT leaders concerned about AI-driven threats and the future of internet infrastructure security.
- Enterprise executives and decision-makers evaluating the adoption and financial impact of large language models.
- Anyone following geopolitical events, particularly the Iran-Israel conflict and its implications for U.S. foreign policy.
- Podcast listeners interested in the intersection of technology, market dynamics, and international relations.
- People interested in the debate between proprietary and open-source AI models and their respective market strategies.
🔑 Key Takeaways
- 1.Anthropic's Mythos model autonomously identified thousands of software vulnerabilities, including 20-year-old exploits in major operating systems and web browsers, leading the company to temporarily withhold its public release for safety.
- 2.Anthropic launched Project Glass Wing, a coalition with Apple, Microsoft, Google, Amazon, and JP Morgan, aiming to use advanced AI to find and fix software vulnerabilities within 100 days before widespread exploitation.
- 3.David Sacks notes that while Anthropic has a pattern of using "scare tactics" to market new products, the cyber hacking threat posed by Mythos is likely legitimate and requires serious attention from companies to patch systems.
- 4.Chamath Palihapitiya views Anthropic's self-regulation as largely "theater," suggesting that sophisticated hackers can already perform similar exploits and that a 100-day period is insufficient to patch global internet vulnerabilities.
- 5.Anthropic's revenue run rate has experienced unprecedented growth, topping $30 billion annually by April, with over a thousand enterprises reportedly paying more than $1 million annually, indicating strong utility and market demand for frontier models.
- 6.The panel discusses whether Anthropic deliberately "ankled" the open-source OpenClaw project by cutting off its flat-rate API access just before launching Anthropic's own agent technology, raising concerns about anti-competitive bundling.
- 7.A two-week ceasefire has been initiated in the Iran-Israel conflict, with VP JD Vance and Jared Kushner heading to Islamabad for peace talks, signaling a potential de-escalation despite recent Israeli bombing in Lebanon.
- 8.The panel suggests that American public opinion perceives Israeli Prime Minister Netanyahu as having too much influence on U.S. foreign policy, potentially endangering the long-standing alliance if an off-ramp is not quickly found.
💡 Key Concepts Explained
Fap Maxing (or 'Chad Maxing')
A philosophy promoted by Eli Long, advocating for detachment, letting go of rumination, and actively living life. Chamath Palihapitiya champions this concept as a healthy way to achieve goals and reduce stress, suggesting that overthinking often hinders progress.
Project Glass Wing
An AI-driven cybersecurity coalition established by Anthropic, involving over 40 major companies like Apple, Microsoft, Google, and Amazon. Its goal is to use advanced AI models like Mythos to find, fix, and harden software vulnerabilities within a 100-day period before they can be exploited by malicious actors.
AGI Models (Artificial General Intelligence models)
Refers to a new generation of AI models, exemplified by Anthropic's Mythos and OpenAI's Spud, that demonstrate massive step-function improvements in intelligence. Brad Gerstner posits that these models are "too smart to be released immediately" without sandboxing and defensive alliances due to their advanced capabilities.
Escalation Trap
A concept in international relations describing how conflicts tend to escalate, often through defined historical stages, making them difficult to pull back from. David Sacks uses this framework to credit President Trump's ceasefire negotiation as a positive step against this tendency.
Jevons Paradox (applied to AI)
While not explicitly named, Brad Gerstner describes how the unit cost of intelligence is plummeting due to improved AI capabilities, potentially leading to increased overall consumption of intelligence. This implies that as AI becomes more efficient, demand and usage might increase disproportionately.
⚡ Actionable Takeaways
- →IT departments and CISOs should take the next few months to gain advanced access to AI models like Mythos (or its competitors) to detect and patch dormant software vulnerabilities in their codebases.
- →Companies with large codebases should prepare for an "AIdriven cyber arms race" by proactively using AI tools for both offense (vulnerability detection) and defense (patching).
- →Monitor the market for emerging open-source AI projects, especially those leveraging crypto incentives for distributed training and development, as they could be highly disruptive to frontier models.
- →Investigate the potential benefits of AI-powered "fap maxing" (detachment from rumination, living life, and trying things) by exploring resources from figures like Elisha Long.
- →For businesses considering adopting large language models, be cautious about entrusting your entire business knowledge to proprietary models from companies like Anthropic or OpenAI due to potential data control and competitive risks.
- →Pay attention to the X (Twitter) platform's auto-translate feature, as it enables unprecedented cross-cultural, real-time engagement and information exchange, providing insights not typically covered by traditional media.
⏱ Timeline Breakdown
💬 Notable Quotes
“Brad Gerstner: "In often times in Silicon Valley, we say move fast and break things. In this case, it means just releasing the model to move further ahead of your competition. But here the company realized it would wreak havoc."”
“Chamath Palihapitiya: "I think it's mostly theater. ... If you actually think that Mythos is capable of doing what it says it can do, two things are true. One is a very sophisticated hacker can probably do those things right now with Opus. And two, if these exploits are this easy to find, whether you use Opus or whether you use Mythos, the reality is you'd have to shut down the internet for about 5 years to patch them all."”
“Brad Gerstner: "This is not zero sum. The TAM of intelligence is dramatically larger than any TAM we've ever seen in our investing careers over the last two decades."”
“David Sacks: "Silicon Valley believes that if you build it, they will come and is willing to finance that build out. And that's basically what's been happening."”
More from this guest
Brad Gerstner
Listen to Full Episode
📬 Get weekly summaries like this one
No spam. Unsubscribe anytime. By subscribing you agree to our Privacy Policy.
Continue Exploring
