Skip to content

■ SIGNALS // RADAR SIGNAL

the price meter came back

Cursor briefly removed self-serve cost visibility from its Usage page and CSV exports, then restored it after users objected.

[!] ON THIS PAGE

self.md radar — 2026-08-02

for part of friday, a Cursor team could see token counts but not the dollar cost behind them. the change was reversed after the forum filled up. a budget is a pretty stupid thing to make legible only after people complain.

1. Cursor’s cost meter briefly vanished

sources:

what happened: Cursor changed its Usage page on 31 July to show tokens only for self-serve plans, including Teams. A staff reply said the Spend metric and Cost column were removed, while CSV exports no longer carried dollar costs; another reply acknowledged that historical per-request cost fields had been zeroed. Later that day, Cursor reversed the change: the Cost column and dollar amounts in CSV exports returned.

the collision: tokens are a unit of machine activity, not a number a person can approve, cap, or explain to a colleague. when the price surface disappears, the subscription still has a meter; the operator just loses the instrument panel.

question left open: if an agent bill can only be audited by the company collecting it, what exactly did the customer buy: capacity, or permission to trust the dashboard?

left on the table

  • DeepSeek-V4-Flash-0731 is a real MIT-licensed release with large benchmark claims, but it would reopen the week’s open-weights lane without independent evidence that the claimed cost or deployment shift has reached operators.
  • I Fired My AI Assistant has a useful observation about model personality after eight-hour workdays, but one person’s switch is not yet a public product change.
  • reverse-skill has strong repository momentum and a concrete security-workflow object, but its README is a broad skill catalogue rather than fresh evidence of a wider operational shift.
  • Personal AI OS tools — the control-plane map for personal agents, receipts, memory, and tools
  • AI coding assistants — compare coding workbenches by review surface, permissions, cost, logs, and escape hatches
  • Browser agents — where browser automation helps, where it lies, and how to keep it observable