Cursor's support bot invented a rule and its cofounder had to say so

The conversation

In April 2025, users of the code editor Cursor found they were being logged out when they used the app on more than one machine. Some who emailed Cursor's support address for an explanation received a reply stating this was expected behaviour under a usage-limit policy. According to reporting by The Register's Thomas Claburn, published 18 April 2025, the reply had come not from a person but from an AI support bot, and the stated policy did not exist. Cursor cofounder Michael Truell responded directly in a Hacker News discussion of the incident, and separately on Reddit, to apologise and correct the record.

What the documents show

The Register quotes Truell's public post: 'Hey! We have no such policy. You're of course free to use Cursor on multiple machines. Unfortunately, this is an incorrect response from a front-line AI support bot.' On Hacker News, Truell posted directly, in a comment tagged by the site as coming from Cursor's cofounder, that 'this user's complaint was the result of a race condition that appears on very slow internet connections,' which created excess sessions that crowded out real ones, and that 'we've rolled out a fix.' Truell also stated the affected user had been 'completely refunded.'

The system boundary

Truell's own Hacker News comment names the specific remediation Cursor adopted for the support channel itself, distinct from the underlying session bug: 'Any AI responses used for email support are now clearly labeled as such. We use AI-assisted responses as the first filter for email support.' That statement draws a boundary the original support reply had blurred, between an automated first-line reply and a human-confirmed company policy, by committing to label which is which going forward.

Where it fails

The Register reports this was not Cursor's first publicly acknowledged case of an AI system producing an incorrect account of its own product; the outlet notes a separate February 2025 episode in which Grok, a different company's chatbot, briefly suppressed unflattering mentions of public figures after what that company called an unauthorised change, a comparison The Register draws to illustrate the pattern rather than to link the two companies' systems.

• When a support channel's first-line responder is an AI system, is that disclosed to the person receiving the answer?
• What technical bug, if any, sat underneath the incorrect policy claim, and was it fixed independently of the support-labelling change?
• What remedy did the specific affected user receive, and was it offered before or after the incident became public?

Cursor's cofounder chose to correct the record in the same public venues where users had raised the complaint, which is what makes his own words, rather than a formal press statement, the primary record of the company's response here.