@VadimStrizheus Nearly 90% of my mac mini is controlled by Hermes and it has only improved the last months but it requires config it is not plug and play as claude or codex.
@LearnWithSubhan I don’t get it- if all these researchers are so afraid of ai, then why do they apply for these jobs? It must be quite expensive for companies to rehire when people leave because they are afraid.
This device gives my Hermes agent the voice and interface I need, but yes, probably too early for most people still depending on phone.
Now I can simply give the Rabbit R1 a command, and Hermes handles the multi agent orchestration behind it. It can activate different agents through the terminal, coordinate their work, and I can monitor everything later through Orca if I need to inspect what the agents did or check the logs.
You definitely need to know what you’re doing to set this up properly, but once the Rabbit is connected to my coding and agent workflow, I barely need to use my phone anymore.
@anshnanda All that capital and infrastructure, and the result is basically a copy of OpenClaw, a project that’s already outdated? Where’s the actual innovation?
When you are running multi agent frameworks where token generation is quite significant,
Jev in multi-agent teams some concrete examples:
- "Which builder takes this task?" → picks one, with probabilities
- "Is this shell command safe to run?" → gates destructive commands before execution
- "Did the build actually pass?" → checks results before they reach a human
- "Is this output good enough to ship?" → second opinion before publishing
- "Is this answer confident enough to act on?" → low confidence goes to a human
Cheap, fast, knows when to ask for help.