Thought for the day...
I am a big believer in tools. Like most developers, I'm willing to spend a 8 hours building a tool that solves a 5 minute annoyance. I've been thinking a bit differently about tooling lately though, particularly, who will consume and use the tool interface. For my entire career, tools have been aimed at human consumption, provide an app or a CLI and the user can interact with it and use it. However, in the last 6 months it seems that changed pretty drastically, it seems that the primary consumers for tools will be AI. I think in the same way that the smart phones changed how humans access the internet, I think AI is going to be the primary means that users are going to interact with tools.
This means that when designing and creating new tools, we need to think how AI will consume it and what sort of levers and escape hatches we might want to provide. We also will need to think intentionally about context and how our tools might harm the AI experience by providing too much context and feedback. So, how can we go about thinking AI first when creating new tools or updating existing ones?
