With the coming changes in June, ACP will charge towards the same budget as claude -p and the Claude Code SDK (since it uses the SDK), so ACP no longer solves this. It's (I think) why Zed added "Terminal Threads" [1] to their agent workflow

1: https://zed.dev/blog/terminal-threads

The ACP budget change is so bizarre to me. If i was more adventurous with my subscription i'd be interested to see if you could intercept UI/input from CC TUI and render that in a native GUI without it being a TUI. That would be "interactive Claude Code" but you'd get a programmatic interface.

But that would be banned almost instantly i'm sure lol.

If you're enjoying erlang, I highly recommend giving elixir a spin. Maybe take a look at a Phoenix tutorial. Such a joy of a lang to work in. And NIFs in erlang and elixir are so nice re: ffi. OTP is just a great platform.

You can quite literally just turn off notifications on macOS.

May I ask how?

I'm aware you can disable notifications for a single app at a time. Is there a way to disable them entirely?

You can press ESC-ESC in Claude code to open the conversation history view and go back to a previous point in the conversation. It doesn't restore code, which is still a miss, but it is built in and useful!

I actually don’t want it to restore code when I jump back. I jump back just to clear context and delay compacting.

Excited to follow this! Have been wanting a good open source on-call solution for a while

My experience of this feature in the betas over the last few months (for Notifications) has been excellent. I used to have so many notifs I would just ignore them all, now I can quickly glance and see which groups of notifs I want to actually read. In most cases, the summary contains all the info I'd want.

If you glance at the notification to decide if something needs attention now or later, then read the mail(s) in full - that's something different entirely and not a problem I'd think.

I like this -- I wonder if an articulating frame like the Leo Rover [1] would improve the all-terrain capabilities?

[1] https://www.leorover.tech/

Quite possibly. The simplicity of this design is what allowed us to complete it in the time frame we were given. Q-Octo has similar dimensions but drives ~4x faster, has 3x the battery capacity and likely provides more torque to the wheels. Improvements to the design are already in the works!

I'm still a daily-driver of Flycut and love it.

Just letting you know it's available right now, just specify `gpt-4o` -- for text streaming anyway. I'd hazard a guess that the audio endpoints are open now, just not documented (like most of the last launches)...

Yeah, seems to be a drop-in replacement for the existing inference APIs. But I haven't found any docs yet for streaming audio/video input.