Scale not bounded by the local machine

What it does

How large a fleet a user runs is no longer a question about the laptop's fans, RAM or battery; running 10,000 subagents at once becomes a question of budget rather than hardware.

Source

Outside beads: the README's Design section (README.md:20-22 and README.md:28-29), two points merged into one idea in #35.

File and line references in this issue are to main as of 75a7a92.

Against the three criteria

• Direction (README): This is the README's direction (README.md:20-22, README.md:28-29).
• Today's terminal users: Parallelism today is bounded by local CPU, RAM and tmux panes.
• Evidence: Nothing of it is built (README.md:31-34).

Depends on

Blocked by #63: scale rests on the cloud execution environment of that idea.

Verdict

• Verdict: do
• Size: XL — rests on a new execution environment and needs a scheduler for many workers.
• Exception to the probe rule: no.
• Date: 2026-09-19
• Verdict ticket: #48

Theme: #51

Start with the README Design section at README.md:20-22 and README.md:28-29, then review blocked issue #63 and the discussion referenced by #48. The issue is done when a cloud execution environment and scheduler support the proposed fleet scale beyond local CPU, RAM, battery and tmux limits.