David Weisburd 🚀

Everyone thinks Renaissance Technologies won because of a secret algorithm, guarded like the Coca-Cola recipe.⁣
⁣
Scott Malpass doesn't buy it.⁣
⁣
The real edge was the machine that built the algorithm: recruiting PhDs, over and over, for decades.⁣
⁣
Ken Griffin will tell you the same about Citadel. Not the execution. Not the models. The culture and the recruiting engine behind them.⁣
⁣
Algorithms decay. Execution gets copied. The day a firm stops recruiting at that level is the day the half-life starts.

China started building space weapons in 2004. The US shelved its own in the early nineties. Most defense investors are pricing the next decade as if that gap doesn’t exist.

@jollyrogersta, CEO of @The_TrueAnomaly, puts the head start plainly:
“China, which is our principal adversary for space warfare, has been at this problem now for about twenty years.”

The US, he says, “effectively mothballed the research and development of all of its defensive and offensive space capabilities” after the fall of the USSR.

Most investors still treat space as a commercial market with defense on the side. Dual use made sense while space was a peaceful domain.

Once an adversary can shoot back, the performance envelope changes. The product that wins in combat may have no commercial use at all. Aviation did the same thing between the wars.

His read on how that gap closes:
“Democracies are very good at responding to a crisis. Not necessarily preventing the crisis from existing.”
That is the investable part. A twenty-year deficit is a demand signal, not just a threat.

Capital moves slowest right before the crisis and fastest right after. The edge is in the gap between.

Thank you @thedatavc for the kind introduction!

This episode is presented by @juniper_square.

Full episode below 👇
youtu.be/0K9cRDIT2ug

Will Robbins (@whrobbins), on why some software businesses never get commoditized:⁣
⁣
"You would think Stripe would be competed to zero margin if you can just replace it with a single API call."⁣
⁣
They don't, because sophisticated companies run multiple payment portals as fallbacks, and raw execution quality plus centralization of data compounds.⁣
⁣
Same logic applies to NetSuite, Oracle, Palantir. Most hated, least usable products in the world. Still some of the best businesses of all time.⁣
⁣
His actual contrarian take: capital has overcorrected from software into deep tech, nuclear, defense, physical.⁣
⁣
"These are fundamentally extremely dilutive. Fundamentally very hard to do. Fundamentally unforgiving with the iteration."⁣
⁣
Software forgives bad bets. Hardware doesn't.

Two 17 year olds got into Stanford, then decided Zoom college was a waste of time.⁣
⁣
Will Robbins (@whrobbins) backed them instead. They stayed in India, skipped school, and started organizing grocery deliveries for neighbors over WhatsApp.⁣
⁣
The insight: India's retail market was an ocean of mom and pop shops, too fragmented and custom to ever get software built for them. So instead of selling to them, the founders decided to replace them.⁣
⁣
They built dark stores covering a mile radius, each fed by one giant mother warehouse per city. Drivers in, out, order packed and delivered in 10-15 minutes, over and over, dozens of times a day.⁣
⁣
Five years later: Zepto does 2 million orders a day, about $4B in annualized sales, and the CEO just turned 24.⁣
⁣
Better selection, better prices, better speed, all from economies of scale mom and pop could never touch.

Will Robbins (@whrobbins) on what winning in retail actually looks like:⁣
⁣
One team redesigned their entire packaging flow to save a single rupee per order. About a penny.⁣
⁣
That's the level of operational discipline the category demands before any of the fun stuff happens.⁣
⁣
Because once you have scale, the business changes shape entirely. Instacart didn't start advertising until 2019. Now ads are a third of their revenue and almost all of their margin.⁣
⁣
Zepto saw this early and built it in from day one: ads, private label, verticalized farmer relationships. Not bolted on later, designed in from the start.⁣
⁣
The operational grind isn't separate from the cash-flow machine. It's the price of admission to build one.