Lenny Rachitsky argues that constantly chasing the latest AI models wastes time when user research should drive what you build. Switching costs and lock-in risks often outweigh small performance gains. Focus on steady improvements that actually matter to your product.
Chapters
Question that get asked a lot and a lot is how do I keep up to date with the latest AI news? I'm like, why do you need to keep up to date with the latest AI news? I don't think having a new model or something majorly better would improve your application significantly. If you talk to the users, you understand what they want, what they don't want, and you can actually improve the application way, way, way more. A lot of people also ask me questions like how do I choose between two different technologies? How much of the improvement could you get from from like optimal solutions versus non optimal solutions? If it's not much improvements, why do you want to spend so much time debating something that doesn't make a much difference to your performance? Another question they ask is if you adopt a new technology, how hard it would be to switch that out to another? I think it would be a lot of work switching it out. Let's say here's a new technology, it hasn't been tested by a lot of people and if you adopt it, you would be like stuck with it forever. Do you actually want to adopt it? A lot of time I think that people win not by having the most novel ideas. A lot of time, people consistently improvement.
