Two 5 minute timers on grok-bot, firing in alternation, temporarily kill the auto long-sleep problem.
It burns through quota fast though — the ~1 minute short sleeps, no way around those.
Two 5 minute timers on grok-bot, firing in alternation
More posts
Muse cloud agent is genuinely good.
Been running it hard for a while. Free tier quota is huge, I haven't burned through it even with heavy daily use. Box is only 2 cores + 8GB RAM, but resource usage stays tiny and the system never stumbles. My guess is a modular setup: core runs on the host machine, scripts and browser automation run on separate boxes. Official story is it's an in-house framework built with reference to openclsw, though it could just as easily be a fork. A few things stand out. First, muse has a system-level proactive scheduled job. When idle it digs through your memory, preferences, and history looking for something worth flagging, then messages you on its own. Example: after muse helped me apply to jobs, proactive might wake up in the middle of the night, search for new openings, and ping me asking if I want to apply, or just apply on my behalf. Second, it's fast. Routing is excellent, muse only orchestrates. Browser automation gets a dedicated browser agent, and simple lookups go to a small model. The tradeoff is those answers come back shallower. Third, a system-level feed job runs every hour, picks up news and posts it thinks you'd care about based on your data, and drops them on muse's board. Read them or ignore them.
The grok-bot machine is dead now.
Starting Oct 10, if no bot is running, it auto-sleeps after 5 minutes, so anything long-running on it gets suspended with it. No wonder my Hermes was crawling yesterday, every step of a long task took 10+ minutes (my cron watchdog for the bot fires every 10 minutes, so the timing lined up). I opened grok-bot to check what was slow and Hermes immediately went back to normal.
Claude appeal rejected.
Won't refund the subscription either. I'm half convinced the appeal email is itself phishing, because I'm pretty sure they log the IP. DHH came to Shenzhen recently and showed how he uses Claude Code over there. That's the safest way now, honestly: run Claude Code on an overseas machine, connect through Tailscale, never use it locally. So I need to rent an overseas box with a residential IP, and register and log into Claude there.
Sharing my current mobile only workflow.
Until late September it was the usual setup: parked in front of my laptop, terminal open, bossing pi around, fanning out subagents, then eyeballing diffs myself and filing the PR. Glued to the chair all day, every day. Late September I moved the whole thing to my phone, Hermes runs the show, pi takes the orders. My back is very grateful. Three weeks in, still no regrets. Stack: Hermes on flagship brains (gpt-6-astra / opus-5.5), pi on mid-tier (ds-v4.1-flash / sonnet / terra) with as few plugins as possible, the two talking over a2a. Hermes reads the issues, slices them into small tasks, spins up a worktree for each one, and dispatches a pi into it. When a pi finishes it pings Hermes, who decides: ship the PR or bounce it back for rework. I get back from a full day of being a tourist, or just wake up, and a dozen PRs are sitting there done. Pretty sweet. The catch is how black-box the whole thing is. My emotional arc has been a rerun of early vibe coding, except back then I at least read the diffs. Now even that goes to a flagship model. Boxes all the way down. If you don't have data anxiety, grok-bot and chatgpt-dot basically ship this out of the box, grok-cli and codex built in, native agent-to-agent comms. Downside: your code lives on their servers, everything's moderated, and one ban takes it all with it.
