How Ultrafast Cuts Incident Investigation Time at OpenAI
OpenAI details how the high-speed mode in Codex has transformed internal incident response, reducing investigation times from hours to minutes.
Radwan AbdulhadiEditor and technology writer
Transforming Incident Response at OpenAI
OpenAI has shared internal examples illustrating how frontier AI supports use cases across its teams, highlighting how a high-speed mode called Ultrafast in Codex has become the default choice for incident response. When unexpected issues occur, engineers often lack the time to wait minutes for answers while problems compound. Ultrafast allows them to investigate multiple potential causes simultaneously and retrieve answers without dividing their attention, a workflow detailed in the official [OpenAI News](https://openai.com/index/openai-cut-incident-investigation/) report.
Aaron Friel, a Member of the Technical Staff at OpenAI who frequently runs point on incidents, explains that finding the right needle in a haystack is the chief job during an emergency. By using Ultrafast, he can spin up a Slack channel, point Codex at the problem, and essentially search all potential failure paths at the same time to respond much faster.
Accelerating Security Analysis and Datacenter Operations
The operational impact of the high-speed mode is particularly visible in security investigations. AJ Trbojevic, a Member of the Technical Staff and Datacenter Security Architect, notes that looking into security issues previously required one to two hours to gather structured and unstructured data, normalize it, and contextualize it for decision-making.
According to Trbojevic, those investigations now routinely take only 10 to 15 minutes, and occasionally occur in near real time. Sayan Sisodiya, who works on the core agent team alongside Adam Perry to build code enabling Codex to take actions and use tools, notes that he now uses Codex to dig through Slack messages as well as logs to identify suspicious changes or similar symptoms discussed by colleagues.
Eliminating Context Switching and Maintaining Focus
The speed improvements have also fundamentally changed how engineers interact with AI during their daily tasks. Jeff Niu, another Member of the Technical Staff, mentions that he previously had to switch to other tasks while waiting for AI responses, often losing his train of thought in the process. With Ultrafast keeping conversations under a 30-second threshold, interactions happen in real time.
Adam Perry shares a similar sentiment regarding workflow continuity. Previously, he would launch a query and immediately begin investigating the issue himself because the tool was not fast enough to wait upon. Now, Codex operates quickly enough that he can remain focused on the task at hand without splitting his attention.
Expanding Daily Workflows, Refactoring, and Automation
In addition to emergency investigations, the speed enhancements have altered how technical debt is managed. Jeff Niu reports that he is now much more willing to tackle large code refactors because reworking existing code takes considerably less time. Code is inexpensive to produce, and refactoring tasks that previously carried high costs now happen almost instantly.
Engineers are also leveraging Codex for administrative duties and manual browser tasks. Adam Perry maintains a running thread acting as a chief of staff that summarizes meetings and unreplied messages every morning. Meanwhile, Sayan Sisodiya utilizes the model's computer-use capabilities to handle tedious form-filling and browser clicks that were previously avoided due to latency. For more updates and announcements, readers can [View all](https://openai.com/news/) articles published by the company.
Sources
- OpenAI NewsHow Ultrafast cut incident investigation from hours to minutes at OpenAI
Continue chronologically
Related entity coverage
Related stories
Oracle turns days of work into minutes with AI
Oracle says ChatGPT Work and Codex are helping recruiters, business users and engineers turn specialist work into repeatable workflows.
Qodo Brings Organizational Context and Code Review to Codex
Discover how Qodo's new Codex plugin integrates organizational knowledge, coding standards, and independent review to catch P0 bugs early.
Heimstaden Transforms Real Estate Management With OpenAI
European real estate operator Heimstaden is transforming property management, cutting costs, and empowering employees using OpenAI technology.
OpenAI Prepares New Security Sprint With Cathedral Systems
OpenAI announces its next internal security sprint in partnership with Cathedral Systems to test and strengthen its AI defenses and systems.
