Experience working at OpenAI ends for an employee, three weeks ago, due to his desire for a fresh start, the ex-engineer clarified, denying any work-related drama.
In a recent blog post, Calvin French-Owen, a former engineer at OpenAI, shared his experiences during his year-long tenure at the leading AI research company. French-Owen joined OpenAI in May 2024, when the startup had around 1,000 employees, and left three weeks ago, as the company's workforce had tripled to 3,000 employees.
French-Owen's departure was not due to dissatisfaction or internal conflict, but rather a desire for a fresh start and a return to his roots as a startup founder. His tenure coincided with a period of rapid scaling for OpenAI, which presented significant operational challenges such as communication, reporting structures, product shipping, people management, and hiring processes becoming more chaotic as the company grew.
One of French-Owen's most significant contributions at OpenAI was the development and launch of Codex, an AI-powered coding assistant. Working alongside a lean team of about 17 people, he helped build and launch Codex in just seven weeks. The intensity of this sprint was, by his own account, "probably the hardest I’ve worked in nearly a decade," with most nights ending after midnight and early starts, often on weekends.
Codex was quickly integrated into ChatGPT, where it saw immediate and substantial user uptake, demonstrating both the tool’s utility and the scale of ChatGPT’s influence. French-Owen observed that the product’s rapid adoption was unprecedented in his experience, crediting the massive reach of ChatGPT for its success.
French-Owen also provided insight into OpenAI’s internal priorities, addressing concerns about AI safety. He noted that the company focused more on addressing practical, real-world risks such as hate speech, abuse, political bias, bio-weapons, self-harm, and prompt injection, rather than abstract, long-term existential threats. This perspective contrasts with published criticisms from some former employees and safety advocates who felt OpenAI did not prioritize safety enough.
OpenAI operates with a policy of minimal email usage, with most communication happening on the workplace messaging platform Slack. The team responsible for Codex's development consisted of eight engineers, four researchers, two designers, two go-to-market managers, and one project manager. French-Owen was in the top 30% by tenure at OpenAI.
The departure of key personnel from OpenAI has been a topic of interest, with Meta reportedly poaching talent from the company, including ChatGPT co-creator Shengjia Zhao and ChatGPT voice mode co-creator Shuchao Bi, with reported pay packages in the nine figures. Despite this, OpenAI remains one of the world's leading AI organizations, with its contributions shaping the future of artificial intelligence.
- Despite his departure, Calvin French-Owen's tenure at OpenAI was marked by significant achievements, such as the development and launch of the AI-powered coding assistant, Codex.
- French-Owen's departure was motivated by a desire for a fresh start, acting as a startup founder, rather than dissatisfaction or internal conflict at OpenAI.
- The scaling of OpenAI presented considerable challenges during French-Owen's tenure, particularly regarding communication, reporting structures, and hiring processes.
- The rapid adoption of Codex, following its integration into ChatGPT, was unprecedented in French-Owen's experience, attributing its success to ChatGPT's massive reach.
- OpenAI's focus on addressing practical, real-world risks is in contrast with some criticisms from former employees and safety advocates, who felt the company did not prioritize safety enough.
