I’ve been consistently impressed with how fast new, AI-native companies are growing. Cursor is a great example:
Cursor isn’t alone here. I’ve noticed a few things consistent between them that’ll inform new products I build going forward.
3 keys to successful products
- AI-native: Predicated on different assumptions and reimagining a better way to do things. Aim for 10X better on some dimension for differentiation.
- Scratch your own itch: A product you use daily or understand deeply.
- Built-in distribution: Thanks to #2 above, you and early adopters talk about how you use it. AI is constantly improving, so there’s also always something new to show, driving an awareness loop.
Solving distribution (the hard part)
The distribution angle resonates since I was listening to Nikita Bier on Lenny’s podcast last night and he talked about how he targets teens because 1) they’re still meeting new people and hanging out with each other, which means 2) they’re more likely to share frequently. This slows as people age.
But you know where it doesn’t stop? Within interest/hobby areas like AI on X/Twitter, who are sharing new things all the time regardless of age. I’m sure the same is true within other hobby groups like running, chess, math, indie hackers, programming languages, open source projects, etc.
So if you can tap into that, you can give yourself a leg up in distribution. The best way to do that is to start with your own interests.
A comment on durability
The consistent observation in Nikita’s typical approach is his products are (intentionally?) short-lived and not durable. But scratching your own itch and focusing on things people care about gives it a better chance at durability.
The takeaway
I don’t think any of this is new. But connecting these dots finally resonated for me because I have pretty good product sense and my teams have always been good at building solid products. But distribution has always been the hard part, so starting with that and building it into the product makes sense.
This approach won’t work for everyone, but if you can 1) scratch your own itch in an area with a 2) sufficient number of others passionate about the same thing to share it with, then 3) leverage AI to solve in a better way, you might be onto something 🚀
I’m going to try this approach with my next side project and I’ll share an update about how it goes.
