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

Episode Summary

Executive Summary: Patrick O'Shaughnessy interviews Palantir CTO Sham Sankar about heretics, talent, culture, AI, and U.S. reindustrialization. The conversation argues that breakthroughs come from builders who defy bureaucracy, that Palantir’s edge is its ontology-driven operating system and forward-deployed model, and that America must rebuild industrial capacity to preserve deterrence and prosperity.

Main Topics: Heretics as founders (Priority: 10/5): Sankar frames military and industrial innovators as founder-like heretics who beat bureaucracy to create lasting change. Personal worldview and American exceptionalism (Priority: 9/5): His family history, immigrant roots, and 80s Orlando optimism shaped a positive-sum, pro-building worldview. Unlocking talent at Palantir (Priority: 9/5): Palantir recruits, stretches, and protects uneven talent by focusing on superpowers, kryptonite, and honesty. Forward-deployed engineering (Priority: 10/5): Palantir sends technical people to the field to solve real operator problems and build product from outcomes. What Palantir is (Priority: 10/5): He describes Palantir as an enterprise operating system that turns data into decisions through ontology and kinetics. U.S. reindustrialization and deterrence (Priority: 10/5): Sankar argues America must rebuild manufacturing and defense capacity to restore learning, resilience, and deterrence. AI, culture, and self-disruption (Priority: 9/5): He sees AI as a chance to amplify the American worker and says Palantir must institutionalize rebellion to stay ahead.

Key Arguments: Heretics drive real change; bureaucracy rarely does. Great builders need obsession, resilience, and willingness to look foolish. Talent is uneven; excellence comes from superpowers, not fixing every weakness. Forward-deployed work beats sales-driven software by proving outcomes in the field. Palantir's ontology maps reality to decisions, making businesses programmable. America's strength is exceptionalism plus plasticity: the ability to learn and pivot. U.S. deterrence depends on rebuilding industrial capacity, not just reforming process. AI should make American workers dramatically more productive, supporting reindustrialization.

Data Points: Expense reviews automated: 85% - Ramp ad read claims AI automates the majority of expense reviews. Expense review accuracy: 99% - Ramp ad read claims its automation achieves near-perfect accuracy. Company savings: 5% - Ramp ad read says its tools save companies five percent. Enterprise adoption capabilities: SSO, SCIM, RBAC, and audit logs - WorkOS ad lists core enterprise features top AI teams need on day zero. Market cap threshold: more than 100 billion euro - Sankar contrasts U.S. and Europe’s track record of creating giant companies. Palantir company size: 4500 people - He notes the company is small enough that every change still feels like a mini insurgency. Three-year crucible: roughly around the three year mark - He says many employees decide whether the company is feature or bug around then. Customer delivery speed: inside of eight weeks - He says Palantir once delivered value in eight weeks and now can do it in about a week. Customer delivery speed now: a week - He says enterprise autonomy and value delivery have accelerated sharply over time. Defense industrial base share: 6% - He says only this share of major weapon system spending went to defense specialists in 1990. Defense industrial base share today: 86% - He says this share of major weapon system spending now goes to defense specialists. Primes in 1993: 51 - He references the pre-consolidation number of major defense primes. Primes today: 5 - He cites the current number of prime defense contractors. Nuclear submarine safety: no deaths due to nuclear incidents - He contrasts U.S. submarine safety with Russian submariners recovering from radiation exposure. World War II production window: Lend-Lease period before Pearl Harbor - He argues the U.S. used pre-war time to build industrial capacity before direct combat. Patrol/recuperation cycle: six months at sea and then six months in Sochi - He uses Russian submariners as an example of poor radiation shielding. Airframe count: 41 airframes - He cites Kelly Johnson’s lifetime output as evidence of founder-level engineering. U-2 timeline: 13 months - He says Kelly Johnson built the U-2 in 13 months. Modernization estimate: two extra capex investment - His pharma policy idea would extend patents to recover two extra capex investments. Generic drugs dependency: 80% - He says 80% of U.S. generic drugs come from China in one way or another. AI compounding window: from eight weeks to a week - He uses this as evidence that Palantir's delivery speed is compounding dramatically.

Pivotal Quotes: "The only shit that ever worked, the things that helped us win all the wars were the things that the heretics actually did." — Sham Sankar: On why he admires military and industrial innovators who defied bureaucracy. "You can either be somebody or you can do something, but you can't really have both." — Sham Sankar: On the tradeoff between status and meaningful execution. "The person is the program." — Sham Sankar: On the primacy of talent and founders over organizational process.

Implications: The unresolved test is whether American institutions can reward builders faster than bureaucracy can dilute them; listeners should look for places where ownership, urgency, and reality-based feedback loops exist.