Just In Time Interfaces

I think the next ChatGPT moment we are going to see is just in time interfaces.

ChatGPT got popular, it had a very familiar interface, it was chat and it was like chatting to a very intelligent human. Later came Canvas where if it thought it was writing a longer text it could write it a kind of like text editor, but so far we have not seen anything more advance.

We have seen products like v0 and lovable where you can generate outputs like websites and landing pages, and the LLM is able to navigate and combine a large list of files to make a website.

The next waves of ChatGPT we are going to see LLM build their own interfaces to the user which the user can then fill out and the LLM can read back the information.

Imagine a todo list, you can have a todolist with ChatGPT today, the problem is that you have to ask it to show you the latest state and you have to tell it when to check of an item.

Now imagine that ChatGPT knows you want a todolist, it builds the interface, then you can either talk to it and it generates todo items, but you can also check and add items yourselfs. It is reactive so when you do any actions the LLM knows and can update any related actions it knows to do.

The benefit of a Just In Time Interface is that it can scale with your users, it can start out simple while still being advance on the backend, and slowly it can add fields and workflows when the user becomes a more advance user. Imagine a CRM that starts out with only having one table like customers, name, email and company name. It does not need to be more advance. Slowly as you are growing and you are two people, it adds a Customer Responsible so you both can see who takes care of that user, the interface adapts and onboarding is fast.