"Founder Mode," DOJ alleges Russian podcast op, Kamala flips proposals, Tech loses Section 230?

Chamath and Sacks argue that Paul Graham’s ‘founder mode vs manager mode’ essay is more slogan than substance. The real differentiator is an individual’s ability to ruthlessly go to first principles, break things to the studs when needed, and balance getting into the weeds with scaling through others. This archetype exists in both founders and non-founder CEOs (e.g., Satya Nadella, Shantanu Narayen, Tim Cook, Nikesh Arora). Founders should beware using ‘founder mode’ as an excuse for not leveling up.

Friedberg emphasizes that the best leaders synthesize diverse inputs, ignore cookie-cutter MBA playbooks, and design unique operating models suited to their business. Chamath adds that winners constantly ‘re-underwrite’ reality—updating their model of the world and throwing away what no longer works. Practically, this means founders should read widely, seek out contrarian operators (e.g., Eric Schmidt), and be willing to change their toolkit post–product-market fit.

Jason and Chamath stress that the skills needed to find product-market fit (risk-taking, curiosity, relentless iteration, naivety plus fearlessness) are often different from those needed to scale (systems thinking, focus, organizational design, governance). Very few people excel at both, which is why ~95% of startups fail and why some companies appropriately bring in seasoned operators or partners (e.g., Eric Schmidt at Google, Sheryl Sandberg at Facebook). Founders should be honest about where they sit on this spectrum.

The Bolt/Ryan Breslow situation—fantastical valuations, aggressive pay-to-play round, and disputed investor claims—is raised as a cautionary tale. Sacks notes how the meme-ification of ‘founder mode’ now allows almost any decision (including doubling burn irresponsibly or promotional misrepresentations) to be spun as bold founder behavior. Investors and founders alike should distinguish between smart risk and stupidity: the only real test is whether decisions increase the odds of winning without crossing ethical or legal lines.

In debating the TikTok blackout challenge case, Chamath argues that recommendation systems are effectively editorial because they are designed to amplify whatever is ‘working’—capturing and spreading lightning-in-a-bottle trends, even dangerous ones. Sacks counters that algorithms mostly optimize for user preference and engagement, not a specific editorial viewpoint, and that stripping Section 230 protections would trigger extreme over-censorship. The nuanced insight: modern recommender systems amplify momentum (positive or negative) and can be gamed, so policymakers must update rules without collapsing online speech.