Anthropic's $30B Ramp, Mythos Doomsday, OpenClaw Ankled, Iran War Ceasefire, Israel's Influence

with Chamath Palihapitiya, David Sacks, and Brad Gerstenfeld

11 Apr 2026 8 min read 1h 42m

TL;DR

Anthropic is taking a calculated risk by withholding its Mythos AI model to coordinate a 100-day global cybersecurity hardening effort with 40 major companies, finding critical vulnerabilities before release—though skeptics question whether this is genuine safety concern or marketing theater. Meanwhile, Anthropic is aggressively competing against the open-source OpenClaw project by cutting off its API access and launching its own agent product, raising potential antitrust concerns around bundling and price discrimination.

Key Moments

Brad Gerstenfeld

“They the company could have just released Mythos, broken a lot of core things on the internet. 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.”

Brad defends Anthropic's decision to withhold the Mythos model and establish Project Glass Wing

▶ 6:20

David Sacks

“I think Anthropic has proven that it's very good at two things. One is product releases. The second is scaring people. And we've seen a pattern in their previous releases of at the same time they roll out a new model or new model card, something like that. They also roll out some study showing really the worst possible implication of where the technology could lead.”

Sacks raises doubts about whether Anthropic's vulnerability warnings are genuine or marketing tactics

▶ 9:51

Chamath Palihapitiya

“I think it's mostly theater. In February of 2019, when Daario was still at OpenAI, they did the same thing with GPT2. That was a 1.5 billion parameter model, which sounds like a total fart in the wind in 2026. But at that time, this 1.5 billion parameter model was supposed to be the end of days.”

Chamath compares Mythos to previous AI safety claims that never materialized

▶ 15:00

Brad Gerstenfeld

“Very simply, when you buy a subscription to these services, they have blended your usage across many users. So there's, you know, nine out of 10 users use less than the tokens they're paying for and the top 10% use much more. When OpenClaw became a phenomenon, the number one open source project in history on GitHub with all of this usage, people went crazy.”

Brad explains Anthropic's decision to force OpenClaw users to switch from subscription to metered API pricing

▶ 25:40

Jason Calacanis

“I believe open source is going to win the day on the large language models and take 90% of the token usage and I think the entire frontier model space could be undercut by open source and I think they realize that SLMs the the smaller language models that are verticalized now that will run on you know, desktops and laptops and is even starting to run on the top ones. That is their biggest competitive threat.”

Jason argues that open-source models represent the real competitive threat to frontier AI companies

▶ 30:55

About the show

›

All-In is a weekly podcast featuring four prominent tech and business investors discussing the latest developments in AI, startups, and geopolitics. This episode focuses on Anthropic's decision to withhold its powerful Mythos model, the competitive threat to OpenClaw, and major geopolitical shifts including potential Iran-Israel negotiations.

Takeaways

Mythos safety theater or genuine security risk Anthropic withheld its Mythos model claiming it can autonomously chain vulnerabilities across OS, browsers, and infrastructure. The company established Project Glass Wing with 40 major companies for coordinated patching, but skeptics note Anthropic has used fear-based marketing before (GPT-2, blackmail study). The key question: can a 100-day window actually patch decades-old dormant vulnerabilities, or is this primarily a go-to-market tactic?

Anthropic's OpenClaw competitive play may be antitrust risky Anthropic forced OpenClaw developers off the $200/month subscription plan to expensive metered API pricing (increasing costs 10-20x), then released a competing Claude agent product days later. This pricing discrimination and bundling strategy could trigger antitrust scrutiny if Anthropic's own agent product receives preferential pricing while third-party products face market rates.

Open-source smaller models emerging as real threat Rather than OpenClaw itself, the industry's biggest long-term risk is smaller language models (SLMs) running locally on desktops and laptops. Industry players from Alibaba to Elon's xAI are racing to launch agent competitors, suggesting the real competitive battleground has shifted from frontier models to distributed, verticalized open-source alternatives that can't be controlled or monetized by centralized providers.