Ex-OpenAI engineer pulls the curtain back on a chaotic hot mess https://t.co/X3wPvEE4Ir
Former OpenAI engineer says he received about 10 emails during his year there https://t.co/OYV3Tdc8yW
Inside OpenAI’s and Perplexity’s shady schemes https://t.co/HO9TJghiva https://t.co/cmVeIuxASO
Calvin French-Owen, a software engineer who worked at OpenAI from May 2024 to June 2025, published a blog post on 15 July describing the company’s internal culture, engineering practices and rapid expansion. He writes that OpenAI’s workforce tripled from about 1,000 to 3,000 in a year, straining reporting lines and tooling. Day-to-day communication is conducted almost entirely on Slack—he says he received roughly ten emails in his twelve months at the firm—and teams often pursue overlapping projects because of a strong bias toward bottom-up experimentation. Despite the organisational sprawl, French-Owen says the company can still ship products quickly. A core group of eight engineers, four researchers and several designers and go-to-market staff built the Codex coding agent in seven weeks, drawing on ChatGPT’s user base for immediate uptake after launch. The post also highlights an internal focus on practical safety issues such as hate speech and bio-weapon guidance, a culture of secrecy amid intense external scrutiny, and management’s close attention to social-media sentiment. The first-hand account offers a rare public window into operations at one of the industry’s most closely watched artificial-intelligence labs.
