Grok Bot vs OpenClaw and HermesThe ladderA staff in one thread
Station 07 · pass at 75%
A staff in one thread
General staff, researcher, and Max Recomp sit in one chat and argue a widget into existence using a week of health memory. The same thread lives on the phone. That is the difference from Discord agents going feral.
1 · Video
Clip
6:48 – 13:01
Three bots, one widget
6:48 – 13:01 · Staff + researcher + recomp negotiate an iOS glance widget from memory.
13:18 – 14:00
The same chat on the phone
13:18 – 14:00 · He taps the thread on iOS and it just continues.
2 · Document
Thesis
Multi-bot is not 'more tokens.' It is specialists with permission to correct each other against a shared memory, on a surface you already check.
- Summon multiple bots into one conversation. They keep their jobs.
- Recomp bot is not technical but knows Ray's targets, so it dictates glance-able widget fields.
- Researcher names platform constraints (iOS widget = glance, no camera).
- General staff holds the machines and the source of truth.
- Phone parity is the feature people underestimate — you are not tied to the desk.
3 · Assignment
Field work
Design a three-bot argument
Add two specialists beside your first bot. Give each a lane and a veto. If they all have the same job, you built a committee.
4 · Quiz
Station quiz
Pass at 3/4. You can retake.
1.Which three bots share the widget chat?
2.What does the recomp bot contribute that the others cannot?
3.Why does he say this beats his OpenClaw Discord setup?
4.Do you have to stay in the thread while they work?
Grok Bot vs OpenClaw and Hermes