Former OpenAI Engineer Shares Insights on Company Culture and Rapid Growth

Three weeks ago, Calvin French-Owen, an engineer who contributed to a key OpenAI product, left the company.

He recently shared a compelling blog post detailing his year at OpenAI, including the intense effort to develop Codex, a coding agent rivaling tools like Cursor and Anthropic’s Claude Code.

French-Owen clarified that his departure wasn’t due to internal conflicts but stemmed from his desire to return to startup life. He previously co-founded Segment, a customer data platform acquired by Twilio in 2020 for $3.2 billion.

His insights into OpenAI’s culture confirm some expectations while challenging certain assumptions about the organization.

Rapid expansion: OpenAI’s workforce surged from 1,000 to 3,000 employees during his tenure, he noted.

This growth reflects the company’s status as the fastest-growing consumer product in history, with ChatGPT surpassing 500 million active users by March and continuing to climb.

Scaling challenges: “Rapid growth disrupts everything: communication, reporting structures, product development, team management, and hiring processes,” French-Owen observed.

LIVE NOW! TechCrunch All Stage

Innovate smarter. Grow faster. Network deeper. Connect with leaders from Precursor Ventures, NEA, Index Ventures, Underscore VC, and more for a day of strategies, workshops, and valuable connections.

Save $450 on your TechCrunch All Stage pass

Innovate smarter. Grow faster. Network deeper. Connect with leaders from Precursor Ventures, NEA, Index Ventures, Underscore VC, and more for a day of strategies, workshops, and valuable connections.

Despite its size, OpenAI retains a startup-like environment where employees can act on ideas with minimal bureaucracy. However, this leads to redundant efforts, with French-Owen noting “at least six libraries for tasks like queue management or agent loops.”

Coding expertise varies widely, from ex-Google engineers building scalable systems to recent PhDs less experienced in production code. Combined with Python’s flexibility, this results in a central code repository that’s “somewhat chaotic,” he explained.

Frequent breakages and slow processes are common, though senior engineering leaders are actively addressing these issues, he added.

Startup energy: OpenAI operates like a nimble startup, heavily reliant on Slack and reminiscent of Meta’s early, fast-paced Facebook era, with many former Meta employees on staff.

French-Owen recounted how his team—comprising roughly eight engineers, four researchers, two designers, two go-to-market staff, and a product manager—built and launched Codex in just seven weeks, working tirelessly.

The launch was remarkable. “Simply activating Codex in a sidebar drove massive user adoption, showcasing ChatGPT’s immense reach,” he wrote.

Culture of secrecy: Intense public scrutiny has fostered a secretive environment at OpenAI to prevent leaks. Yet, the company closely monitors X, reacting to viral posts, with French-Owen quoting a friend’s quip: “This place runs on social media buzz.”

Addressing misconceptions: French-Owen suggested that the biggest misunderstanding about OpenAI is its perceived lack of focus on safety. While some former employees have criticized its safety practices, he emphasized practical safety efforts targeting issues like hate speech, abuse, political manipulation, bio-weapons, self-harm, and prompt injection.

OpenAI doesn’t dismiss long-term risks, he noted, with researchers studying potential impacts as millions use its models for critical tasks like medical advice and therapy.

With governments and competitors closely observing, and OpenAI watching its rivals in return, “the stakes feel incredibly high,” he concluded.