The Horse Cart Problem
Every week I see another press release: "Company X adds AI to their platform!" And every week I think the same thing — that's a jet engine on a horse cart.
The cart wasn't designed for speed. The chassis can't handle it. The wheels will fall off. But the marketing team gets their headline.
What AI-Native Actually Means
When I say AI-native, I mean something specific: the AI isn't a feature — it's the architecture.
Think about it this way. A traditional research tool has a search box, a database, and maybe some filters. You type, it fetches, you read. Adding AI means throwing a chatbot on top of the same workflow.
An AI-native research tool doesn't have a search box. It has an intent. You tell it what you're trying to understand, and it figures out what data to gather, where to look, how to synthesize, and what to present. The workflow is fundamentally different.
Why This Matters Now
We're at an inflection point. The tools are finally good enough — fast enough, cheap enough, capable enough — to rebuild entire workflows from scratch. Not improve them. Rebuild them.
The companies that understand this will build 10x products. The ones that don't will build 10% improvements and wonder why nobody cares.
My Bet
I'm betting everything on AI-native. Every product I build starts with a blank page and one question: if AI could do this entire job, what would the human's role be?
Usually the answer is: making decisions, not doing busywork.
That's the future I'm building toward.