I've been thinking about the concept of "small software" and how AI will turn more founders into holdco founders builders rather than startup founders. It kinda reminds me of department stores. They died because Amazon could sell everything cheaper and faster. The only retailers that survived were the ones that went deep into niches. The running shoe store that perfectly served marathoners. The coffee shop that obsessed over single-origin beans. Software is about to go through the same transformation. When it cost millions to build products, you had to be Oracle – selling everything to everyone. But AI has changed that equation entirely. Now, the most interesting opportunities look more like specialty stores. Not because they're small, but because they can go deeper than big companies ever could: A $50k/month product that serves accountants at law firms better than QuickBooks. A tiny CRM that understands real estate agents better than SAP. A marketing tool that serves Substack writers better than Mailchimp. The winners won't just build one specialty store. They'll build collections of them. That's where holdcos come in. Instead of raising venture capital to build one massive product, smart founders are bootstrapping portfolios of focused products. Each might be small, but together? They're building a cash-flowing portfolio. Not crazy that we'll see $50M-$1B exits — but with teams of 25 or less. Big, medium and small exits, probably mostly bootstrapped. Lots of dividends along the way which is rare in VC startup land. Think of it like collecting art pieces or real estate properties. Each piece cash flows on its own. But the portfolio compounds into something much bigger. "BuT loTs oF PeOpLe WiLL fAiL." Yeah, you're right. Not everyone will create products and portfolios people love. They won't. It won’t get any easier to build a startup that people love. Competition will probably go up. But costs are going down, and the ability to find customers has never been easier — if you know how to build internet audiences and create ad creative that stops the scroll. Buuuuut I am saying this: How people are building software is changing. I'm saying small code is the new big code.