Imagine you had to navigate websites by typing.
You'd go crazy. So why do we treat LLMs like that?
I believe our ways of interacting with LLMs are in their infancy and there is immense potential in rethinking them.
Instead of focusing on automation and prompting, we should build use case specific interfaces to streamline our interactions with LLMs.
On this site I'm sharing some thoughts and ideas around my exploration of this topic.
Arne Riethmüller
arne.riethmueller@gmail.com
M.Sc. Business Informatics based in Berlin currently working as an Automation Engineer. I love rethinking how we interact with AI systems (and rethinking systems in general).
I would love to hear your thoughts and ideas!
My A+ Thesis on Improving Coding Agents
Before Claude/Cursor plan mode was released, I had an Idea for a slightly different version of these very useful features. By surfacing ambiguity in coding prompts to the user and by implementing UI elements that let users quickly communicate their intent, we can effectively reduce misaligned outputs in coding agents.