OpenAI Codex lead on the new shape of product work | Andrew Ambrosino

A polished prototype can now arrive before a team has settled what the product should do. In this interview, the product and engineering lead for OpenAI's Codex app describes how inexpensive implementation shifts work toward choosing among competing ideas and deciding how they fit together. A convincing interface no longer tells colleagues how thoroughly its assumptions have been checked. Listen to the discussion.

The suggested response is to choose the format that answers the unresolved question. A document can clarify a vague product goal; a working prototype can test an interaction. Teams still need to establish what each artifact demonstrates before committing to it.

“Ninety percent of people at OpeningAI use Codex. Not ninety percent of engineers, that's ninety percent of the entire company.”

Andrew Ambrosino

“The implementation is actually not the expensive part anymore. It's dare I say taste.”

Andrew Ambrosino

“If research is listening at any company, please make the models better at deleting code.”

Andrew Ambrosino

“I am very confident that the Codex app that we released in February, if that had been ready in November, it would have absolutely failed in the market.”

Andrew Ambrosino

“do not get married to your exact process. Get married to like the outcomes that you were uniquely able to deliver”

Andrew Ambrosino