hey Boris, with multi-agent work, e.g. Agent Teams, following other agents is not possible because they work so fast. (my tmux panes are just matrix)

do you think something like a /speed config can be introduced to adjust agent working speed and let people adjust?

How is this different, or how does it complement Agent Teams? When should I use which?

> "you are out of tokens" notices

there's your answer. You're likely a heavy user. It's just that in this particular case Auto is shuffling models & reasoning efforts just right so you stay afloat.

There are a lot caveats when people say they're running out credits in hours: lots of people runnings clawS (multiple) 24/7; lots of people using Agent Teams or similar; lots of people just max reasoning for everything; etc.

If you get your work done in the 20 plan, thank the credit Gods and sleep easy. Price hikes are probably coming, or just reasons to draw more credits out from you in one form or another (e.g. new OAI device draws from you 20 sub to use)

My main driver has been Ghostty but I've been looking at Warp for a while. Warp seems like a full on IDE (~ADE) though, as opposed to a minimalistic terminal. Can anyone add some thoughts? Are these 2 very different?

tangential: I've seen Mitchel tweet that people in SF have ran up to him showing him how they fully riced their Ghostty setup. How many people here have done this and how easy/manageable is it? e.g. just forking the repo and implementing whatever Warp feature I like?

Warp founder here. We actually are chatting with Mitchell about integrating Ghostty so it's the terminal grid renderer within Warp.

Warp failed to launch. Perhaps too much AI pushed onto the users in the early days that failed to show its charm.

Ghostty remains incredible stable and usable and fast compared to competition.

Warp is equally stable, almost as fast, while being more usable than Ghostty.

What makes you say that Ghostty is less usable than another?

Well, for starters, I get obscure errors if I get something wrong editing the Ghostty config file.

I think there's a reason people are very interested in libghostty. It's a fully-featured library to build off of, but Ghostty proper still lacks a lot of polish.

libghostty makes it pretty easy to do. I spent about two weeks setting something up until it was advanced enough to daily drive. I wanted to have a modal workflow similar to vim or tmux copy mode, but without having the overhead of using tmux... that's probably a lot more complicated than "I want Ghostty but with $X tweak". You can poke around in the repo to get a feel for what's involved if you want: https://github.com/milch/mistty

check out yaw terminal for a terminal first experience that also treats ai cli as a first class citizen. and if you're on windows is very dialed into git bash.

apologies, just trying to get the word out. and why is it crap? it seems exactly relevant to what i'm replying to.

well maybe meaningfully participate in the thread before pitching your product

I'm 6 days late but if anyone reads this: what does it mean we should "upgrade now to quantum-resistant cryptography" ? If I'm using RSA am I supposed to switch to something else?

I was reading through CC MCPs docs ... MCP Notifications kind of did this right? The server/client could update each other automatically. So there was already this channel-like communication.

This is like that but instead of the server/client sending messages it's you.

people who use pi: is this stuff easy? do I just clone the repo and give SKILL.md arguments to implement features & customize ??

2 days late to this post so not sure anyone's here, but what fascinates me is finding out how the different segments (or other such insects) are connected. Like, at each of those junctions it seems like a bottle-neck scenario where everything has to be condensed into some "wire" form to then connect to the next segment.

Very useful breakdown; thank you!

so are people pairing OpenRouter with open agent harnesses like Pi or OpenCode?