Maybe a feature for clarity on the context bloat or warning to let users know how many tokens they are burning on unneeded system context on every message sent
Because 12k tokens per message on context with opus goes fast and multiply by multiple sessions thats 120k tokens after just 10 messages
I think many people are unaware of the optimal claude code settings
Did a context bloat audit last night and was very surprised cut context bloat by 75% PER MESSAGE !!
Using voice to text with ai just increased efficiency insanely because when you talk it out, it can be more descriptive and natural language for the AI to actually understand what you even want
instead of having to type every single word out every time you wanna get a thought out
and then correct it when it doesnโt do it correctly the first time
this is way more efficient, and you can speak more words than you can read or write per minute.
It is the most efficient form of communication.
It also helps you sharpen your own speech ๐ค ๐คฏ