Calvin French-Owen, former OpenAI engineer and co-founder of Segment, shared insights about his year at the company after resigning three weeks ago.

During his time there, OpenAI grew rapidly from 1,000 to 3,000 employees.

He discussed the challenges of this growth, including unclear reporting structures, overlapping team efforts, and communication issues.

While employees have freedom to act without bureaucracy, this often led to duplicated work and inefficiencies.

He also pointed out technical challenges like a messy central code repository and varying coding skills among staff. Despite these issues, he said engineering leadership is working on improvements.

French-Owen compared OpenAI’s culture to an early-stage startup, fast-paced and Slack-driven. He shared his experience launching Codex in seven weeks with a small team, which gained popularity through ChatGPT integration.

French-Owen addressed misconceptions about OpenAI’s safety focus, stating the company prioritizes issues like hate speech and abuse prevention, while also considering long-term risks.

He noted OpenAI’s secretive nature and awareness of public perception, especially on social media.

.source-ref{font-size:0.85em;color:#666;display:block;margin-top:1em;}a.ask-tia-citation-link:hover{color:#11628d !important;background:#e9f6f5 !important;border-color:#11628d !important;text-decoration:none !important;}@media only screen and (min-width:768px){a.ask-tia-citation-link{font-size:11px !important;}}

🔗 Source: TechCrunch

🧠 Food for thought

1️⃣ Hypergrowth creates predictable chaos across tech companies

OpenAI’s expansion from 1,000 to 3,000 employees in a year mirrors classic scaling challenges that plague even the most promising tech companies.

Research shows that over 50% of U.S. startups fail within five years, with this figure rising to over 70% after ten years, often due to inability to manage rapid growth effectively 1.

The communication breakdowns, duplicate efforts, and code quality issues French-Owen described are textbook symptoms of hypergrowth. Each doubling of company size typically necessitates complete process redesign 2.

This pattern appears consistently across fast-growing companies. For example, POD Point faced challenges when they doubled staff from 24 to 54 while growing revenue from £4.5 million to £6.5 million in a single year 2.

What makes OpenAI’s case particularly notable is maintaining a startup-like culture (“running entirely on Slack”) despite its size and importance. This demonstrates how organizational structure often lags behind headcount growth during periods of extreme scaling.

2️⃣ The speed-stability paradox intensifies with AI development

French-Owen’s account of building and launching Codex in just seven weeks highlights a fundamental tension in AI development: balancing innovation speed with operational stability.

This “move-fast-and-break-things” approach resembles early Facebook culture, which is unsurprising given French-Owen noted OpenAI is “full of hires from Meta” 3.

The financial stakes of this balance increase dramatically with scale. What might be a $10,000 mistake in a small startup can escalate to multi-million dollar consequences in larger organizations 4.

For AI companies specifically, this tension carries unique complications. Codex was trained on 159 GB of Python code from 54 million GitHub repositories 5, representing massive computational investment that contrasts with the described “back-end monolith” that’s “a bit of a dumping ground.”

OpenAI’s ability to achieve instant product adoption (“I’ve never seen a product get so much immediate uptick just from appearing in a left-hand sidebar”) creates a powerful advantage that may temporarily mask the operational costs of prioritizing speed over stability.

Recent OpenAI developments

……