@grok@neogoose_btw I did however make the mistake of adding an Apache 2 license to pretty much all my projects, while MIT is probably the better choice. But I can't just change all my projects to MIT?
It seems like everyone in my feed is building either an agent terminal multiplexer, or an agent memory store. I'm definitely not working on a multiplexer.
Have been playing with oxc in the meantime and I like it. Just switched my new project to use that instead of Biome. As for pnpm, still one of the best obviously, but I also really like bun's toolkit as a while, like the testing and a bunch of builtin libraries.
@PeaweeBoomhall Did some limited testing with Qwen-3.5-9-Q4_K_M. It wasn't too bad, but I think context window is limited, so it started derailing fast from my initial request. I could probably run a larger model though. Need to give that a try next
I love the idea of opensource coding agents like Pi and OpenCode, and to an extent OpenClaw, but how do people actually use this with SOTA coding models without burning literally $1000 per month? Especially with multiple agents. I stick to Claude Code with my Max plan ($200) because of that. (And because it's awesome). What am I missing here?
I know, but it's not the same, claude code would still use its internal context. Also for OpenClaw I mainly use gemini-3-flasp-preview for non coding tasks, which works really well. But for dev I'd like to switch to Pi. Seems like oauth from my Copilot and Codex subs should work without breaking TOS, so will try that
@iamdothash Just saw you're the cliamp guy. Have been having a lot of fun with it! And I have to check out Aether, but not for my Neovim color scheme ;)
I'm curious on what models API/subscription you use. The main reason I stick with Claude Code is because I have a Max subscription. As far as I understand it's against the TOS to use these oath tokens with Pi. If I hook up the API, then I'm easily burning 100$ per day. So what am I missing here?