Katy Shi, a researcher working on Codex behavior at OpenAI, says that while some have described his default personality as “dry bread,” many have appreciated his less sycophantic style. “A lot of engineering is about being able to take critical feedback without interpreting it as malicious,” Shi says.
Several gigantic companies have also started using the Code. “The fact that ChatGPT is synonymous with artificial intelligence gives us a huge advantage in the B2B market,” says Fidji Simo, CEO of OpenAI Applications. “Companies want to use technologies that their employees already know.” Simo said OpenAI’s strategy for selling Codex relies heavily on packaging it with ChatGPT and other OpenAI products.
Cisco president and chief product officer Jeetu Patel says he has told employees not to worry about the cost of using the Code because they will need to be comfortable with the tool. When employees ask if they will “lose their jobs because they use these tools,” Patel replies, “we have to tell our employees no, but I guarantee you will lose your job if you don’t use them because you won’t be important to yourself. So you’ll be out.”
Today there is panic around agents coding artificial intelligence has spread far beyond Silicon Valley. The Wall Street Journal credits Claude Code with causing it $1 Trillion Tech Stock Selloff last month as investors feared the software would soon become completely obsolete. A few weeks later, IBM shares suffered their worst day in 25 years after Anthropic announced that Claude Code could be used to modernize legacy systems that run COBOL, a common language on IBM machines. OpenAI has worked tirelessly to bring its AI coding agent into the public conversation, spending millions of dollars to advertise Codex in the Super Bowl instead of ChatGPT.
At Mission Bay Temple, no one needs to be talked into the Code. Many OpenAI engineers I talked to said they rarely type code anymore. They just spend all day talking to Codex. Sometimes they meet and do it in the church.
At headquarters, I took part in a Codex hackathon – about 100 engineers gathered in a gigantic room. Everyone had four hours to build their best demo using Codex. OpenAI’s senior leader stood at the front of the room, looking up from the laptop in his hands and saying the team names into the microphone. Team representatives nervously walked up to the podium and gave brief speeches about their AI projects with trembling voices. The winners received Patagonia backpacks.
Many projects were created in Codex and designed to assist engineers better utilize Codex. One group built a tool that summarizes Slack messages into weekly reports. Another group created an AI-generated, Wikipedia-style guide to OpenAI’s internal services. Many of these demonstrations previously took days or weeks to get going, but now they can be done in one afternoon.
As I was leaving the house, I met Kevin Weil, a former Instagram executive who now heads OpenAI for Science, the company’s fresh unit dedicated to building AI products for researchers. He told me that Codex was working on a few projects for him overnight and would check them out in the morning. This became regular practice for Weil and hundreds of other employees. One of OpenAI’s goals for 2026 is to create an automated intern that will research (what else?) artificial intelligence.