@thdxr Who's your VM provider? I need to try this dedicated work VM thing. I'm a bit sceptical about working via ssh though. Amongst other things, I'll also have to tunnel running services to my local machine or maybe add to my tailnet... urgh
Yes please. I'm generally more interested in what the agent is doing/trying to achieve than what file it's reading. I actually added instruction in my append-system file to have the model always tell me what it's trying to achieve. I'm always in the editor anyways and can read the file myself if I need to
@TechnoAmhara@vimtor@manaflowai I have the same question. Tmux gives me all this. My guess is most of the people super impressed by cmux were not terminal users prior to AI. Either that or I'm missing something
@_justelias@mattpocockuk I was referring to skipping the in-house quality check. I'd like to believe most teams have a QA step in their process, and my worry is that feature flagging everything just makes it easier to skip that step. No strong opinions btw, just thinking out loud.