# OpenAI — How Ultrafast cut incident investigation from hours to minutes at OpenAI

- Company: OpenAI (openai.com)
- Announced: 2026-10-26T07:00:00+00:00
- Category: not stated
- Coverage: not counted
- Announcement: no
- Group: routine
- Source: https://openai.com/index/openai-cut-incident-investigation
- Record: https://forck.live/items/27941-how-ultrafast-cut-incident-investigation-from-hours-to-minutes-at-openai
- Subject: GPT / ChatGPT / API
- Models affected: Codex

OpenAI engineers have made Ultrafast, a high-speed mode in Codex, their default for incident response, cutting security investigations from an hour or two to 10 to 15 minutes. The company published the internal accounts as part of a series on how its own teams use its technology and APIs. Aaron Friel, a member of the technical staff, says Codex lets him search many possible causes at once instead of one haystack at a time. AJ Trbojevic, a datacenter security architect, says pulling structured and unstructured data together and normalizing it now takes 10 to 15 minutes and is at times near real time. Jeff Niu says most of his work now returns an answer in under 30 seconds, the threshold past which he says he would check Slack or ask a coworker instead. Engineers also describe using Codex for large refactors, morning planning summaries and browser form-filling, with Adam Perry saying he can now stay focused rather than split his attention.

## Evidence

Verbatim from https://openai.com/index/openai-cut-incident-investigation:

> AJ Trbojevic, Member of the Technical Staff and a Datacenter Security Architect, says investigations into security issues used to take an hour or two to pull structured and unstructured data together, normalize it, and provide context to make a decision. “Now that happens in 10 to 15 minutes,” he says. “And at times it’s near real time.”

---

Record: https://forck.live/items/27941-how-ultrafast-cut-incident-investigation-from-hours-to-minutes-at-openai
Catalogue: https://forck.live/llms.txt
Current issue: https://forck.live/feed.md