✳flâneur — a map of the web's best reading

Pete Koomen

koomen.dev · 4,508 words · saved by 12 readers

I noticed something interesting the other day: I enjoy using AI to build software more than I enjoy using most AI applications--software built with AI. When I use AI to build software I feel like I can create almost anything I can imagine very quickly. AI feels like a power tool. It's a lot of fun. Many AI apps don't feel like that. Their AI features feel tacked-on and useless, even counter-productive. I am beginning to suspect that these apps are the "horseless carriages" of the AI era. They're bad because they mimic old ways of building software that unnecessarily constrain the AI models they're built with. To illustrate what I mean by that, I'll start with an example of a badly designed AI app. A little while ago, the Gmail team released a new feature giving users the ability to generate email drafts from scratch using Google's flagship AI model, Gemini. This is what it looks like: Here I've added a prompt to the interface requesting a draft for an email to my boss. Let's see what G

AI Horseless Carriages | koomen.dev AI Horseless Carriages I noticed something interesting the other day: I enjoy using AI to build software more than I enjoy using most AI applications--software built with AI. When I use AI to build software I feel like I can create almost anything I can imagine very quickly. AI feels like a power tool. It's a lot of fun. Many AI apps don't feel like that. Their AI features feel tacked-on and useless, even counter-productive. I am beginning to suspect that these apps are the "horseless carriages" of the AI era. They're bad because they mimic old ways of build

Explore this link on the map →

saved by

related reading