@TokenGremlin if chat/work/codex merge without smart routing it's just three modes fighting. would you rather pick the mode, or have it infer from context?
@pcshipp Claude all day, OpenAI is having some issues right now. They don't have enough compute, they just cut usage for $200 plan in half, they promised a reset today at 10am PST that never came and have been silent about since.
@SahilPanhotra This opens it up for mass adoption, do you think every joe schmo knows how or wants to buy a vps and set it up for hermes. Even a lot of the heavy AI users here on X dont want to do that.
@TheBlueMeanieUK homelab how-posts are easy; the why is the hard one to write down. what was the actual reason yours exists — control, cost, or just not trusting the cloud?
@anilpervaiz mods that can rewrite prompts and approve tools is a supply-chain nightmare. are you blocking third-party mods entirely, or sandboxing somehow?
@kaushik_holla digging three open-source Open Dots just to steal bits for Orbit is the right kind of homework. which piece are you actually borrowing first?