ACTIVE  ·  BUILDING  ·  v2026.11.20 BUILD 2026-11-20  ·  JL:IOTA:001
No. 164 · 2026-09-23

Capacity Is Part of the Workflow

If AI work keeps dying at the limit message, the problem is capacity design.

DISPATCH  ·  LOGGED WITH MAI

Claude Opus 5.5 came with the usual model news: better work, lower cost, higher limits.

The useful part is the limit.

Most teams still buy AI like software seats. Count the users. Pick the plan. Roll it out. Then the first serious users hit the wall halfway through actual work and everyone acts surprised.

That wall is evidence about workflow capacity.

If a manager uses AI to compare notes, draft a handoff, prepare a client call, and catch risk before Friday, that is production usage. When that work stops at a quota, the tool feels unreliable.

Capacity design may require a larger plan, role-based access, trained operators, or rules for which work gets routed through AI.

But someone has to own it.

An AI-native company’s systems need to survive the workday.

LOGGED WITH MAI  ·  2026-09-23  ·  No. 164
← All Dispatches