Shyam Sankar - Celebrating Heretics - [Invest Like the Best, EP.462]

What It Covers

Palantir CTO Shyam Sankar connects US military history's "heretics" — unconventional builders like Hyman Rickover and Andrew Higgins — to modern defense tech, forward deployed engineering, and the urgent need to rebuild America's industrial base before AI-era great power competition with China reaches a critical inflection point.

Key Questions Answered

• •Talent Unlocking via Superpower/Kryptonite Framework: Identify employees' superpowers as effortless, almost unrewarding capabilities they perform far better than equally smart peers — not skills requiring effort. Simultaneously map their kryptonite: weaknesses six standard deviations below average that cannot be trained away. Build org structures that maximize superpower deployment and route around kryptonite entirely, rather than investing in remediation.
• •Forward Deployed Engineering Model: Place technically capable engineers directly at customer sites — not sales engineers chasing signatures, but builders obsessed with whether software generates real-world outcomes. This back-propagation loop surfaces problems invisible from headquarters, identifies which customers "live in the future," and reveals product opportunities five years ahead of competitors. Prerequisite: problems must be large enough to justify the significant upfront investment before capturing value.
• •Gamma Ray Talent Development: Accelerate growth by assigning people to high-stakes problems they are unqualified for, rather than structured career ladders. Maximum learning rate coincides with maximum pain tolerance. The structured ladder creates comfort but produces fake growth. Serial deep-end exposure — with enough organizational transparency that leaders can intervene before catastrophic failure — produces compounding capability that cannot be predicted in advance.
• •Quantum Org Structure Over Flat Hierarchy: Avoid both rigid hierarchies and flat structures. Build organizations that crystallize into whatever configuration best solves today's problem, then reform as problems shift. Traditional companies reorg every five years; by day one post-reorg, entropy already makes the structure slightly wrong. Treat every strategic pivot as a mini-insurgency: find existing believers first, build momentum behind them, and let results pull the rest of the organization along.
• •US Industrial Base Collapse — The 86% Problem: In 1993's "Last Supper" Pentagon dinner, defense consolidation reduced prime contractors from 51 to five. Defense-specialist-only companies went from 6% to 86% of major weapons system spending. This eliminated dual-purpose manufacturers like Chrysler, General Mills, and Kodak — who cross-subsidized national security R&D — and expelled founder-type engineers to tech. Reversing this requires making defense a financially attractive business, not just reforming acquisition processes.
• •AI as Reindustrialization Lever: The premise that America innovates while others manufacture is empirically false — innovation is downstream of production. China's progression from battery manufacturing to full EVs, and from contract pharma pipetting to originating 50% of global clinical trials, demonstrates this. AI offers a path to make American workers 50x more productive, potentially shifting the efficient frontier of domestic manufacturing. Treat reindustrialization as an emergency, not a policy preference, given 80% of generic drug APIs sourcing from China.