Inside OpenAI: What a Former Engineer Really Experienced

Share
Inside OpenAI: What a Former Engineer Really Experienced

Calvin French‑Owen, who joined OpenAI in mid‑2024 and departed three weeks ago after roughly a year, has shared his candid reflections about the fast-paced, chaotic, and exhilarating environment inside one of the world’s leading AI companies.

Rocket‑Ship Growth and Startup Chaos

During his time, OpenAI expanded from about 1,000 to 3,000 employees—tripling its size in just 12 months. With that scale came all the growing pains: internal communication gaps, duplicated efforts when multiple teams built similar tools, and evolving reporting structures. Behind the scenes, the workplace still “felt like a startup,” with low bureaucracy and high autonomy—but that led to a backend code repository becoming a dumping ground of mixed-quality scripts .

Building Codex in a Frenzy

French‑Owen was part of a small, focused team—eight engineers, four researchers, two designers, two go‑to‑market staff, and a product manager—that launched Codex, OpenAI’s new AI coding assistant, in just seven weeks. The sprint was intense, bordering on sleepless, but symbolic of the company’s "launch-first" mindset. Even a sidebar launch drew immediate user adoption, underscoring the power of the ChatGPT ecosystem .

Secrecy and “Twitter Vibes”

Despite its size, OpenAI operates in high secrecy due to constant public scrutiny. Everything unfolds within Slack and internal tools, with buzz on X (née Twitter) closely monitored. A viral post can set off immediate reactions internally—prompting quick pivots or damage control.

Safety: More Than Just Doomsday Scenarios

Contrary to the outside perception that OpenAI downplays risk, the company has devoted teams focused on real-world safety concerns: hate speech, bias, misinformation, self-harm prompts, prompt injection, and dual-use threats such as bio-weapon creation. These function alongside long-term AGI considerations, reflecting a more practical and proactive safety approach than many assume.

Key Takeaways

  • OpenAI’s explosive growth brought energy and opportunity, but also friction: duplicated work, unstable code, and organizational chaos.
  • The "move-fast-and-break-things" culture fuels rapid innovation, exemplified by Codex’s swift rollout, though it comes with extreme pressure.
  • Despite the startup vibe, OpenAI is acutely aware of its public visibility, keeping tight internal controls and monitoring external sentiment.
  • Safety remains a strategic focus, grounded in tangible risks—not just hypotheticals—debunking the myth that OpenAI is indifferent to responsible AI use.

Read more