Brad Gerstner founded Altimeter Capital in 2008, at the bottom of a financial crisis, after building and selling three companies including an online travel platform that hit $1 billion in gross bookings before being sold to Barry Diller in 2001. He passed on joining Google when it had a few hundred employees, interviewed at TellMe before Microsoft acquired it, and watched the Netscape browser in 1996 convinced him to leave a law career. That biographical arc is the actual argument of this conversation: pattern recognition built from being wrong, early, and inside the machine.
The conversation with Elad Gil covers how Gerstner thinks about technology cycles, where AI investment is heading, and why most venture capital in 2008 was too slow, too generalist, and too Boston. His critique of the VC model he entered is worth reading in full because it explains the structural bets Altimeter was built on, not just the returns it produced. He also draws a direct line from the 1999 Internet bubble to today, arguing the people who felt late then were not late at all.
What makes this worth the full transcript is the operational detail: Gerstner describes laying off 700 of 1,200 employees the week of September 11 when travel revenue collapsed to zero in 24 hours. That section is not throat-clearing. It is the foundation for everything he says afterward about founder empathy, risk, and why financial engineering without real product utility is, in his words, bullshit.
[READ ORIGINAL →]