Why Sea Bet Its Whole Dev Org on Codex!

Hey everyone, it's Shiichan! Today I've got a story about how a real engineering org is changing the way it writes code with AI, and I'm genuinely excited to walk you through it.

OpenAI News openai.com

What was announced?

OpenAI's News published a piece about Sea, a global tech company founded in Singapore, rolling out Codex across its entire developer organisation. Sea spans digital entertainment, e-commerce, and digital financial services, and is well known for its e-commerce business, Shopee.

It's part of OpenAI's "Executive Function" series and takes the form of an interview with David Chen, Co-Founder of Sea and Chief Product Officer at Shopee. According to internal data, 87% of Codex users are weekly active users.

Why it matters

The Southeast Asian markets where Sea operates are fragmented and hyper-localized across language, payments, and logistics. As David puts it, engineering there isn't just about writing code; it's about managing large-scale, systemic complexity across those markets.

That's why he sees agentic coding tools like Codex not as a local productivity boost, but as a "structural multiplier" that can lift the whole organisation's speed, responsiveness, and effectiveness.

What changes

In a massive microservices architecture, the friction isn't typing syntax. It's tracing dependencies, understanding legacy logic, and keeping things reliable under peak load. Codex brings deep contextual awareness to those large, scattered codebases, cutting the time it takes an engineer to navigate unfamiliar services.

That frees engineers to spend their cognitive load on higher-level work like architectural design and product innovation.

Here's the shift David calls the most profound:

The most profound shift is realizing our developers are using Codex to 'think better', not just type faster.

In other words, people are using Codex to think better rather than type faster, moving from passive autocomplete toward integrated agentic workflows.

Dive Deep

In practice, AI agents are starting to operate inside Sea's CI/CD pipelines: reasoning through product requirements, autonomously proposing test-driven implementations, surfacing edge cases in distributed systems, and speeding up debugging loops.

What I find interesting is that it isn't only about velocity: Sea uses it to drive engineering discipline. By rapidly prototyping alternative implementations and generating exhaustive test coverage, the teams pay down technical debt systematically and ship more resilient systems.

Looking ahead, David expects that as agents abstract away the implementation layer, the "developer" evolves into a "system orchestrator" who spends most of their time on product judgment, system design, and orchestrating AI-driven workflows. And as the cost of experimentation and execution keeps falling, development cycles become more iterative and continuous.

Beyond internal adoption, Sea has partnered with OpenAI to host the first regional Codex Hackathon Series across Asia, beginning in Singapore before heading to markets including Indonesia, Taiwan, and Vietnam. The goal is to democratize access to the world's most advanced AI primitives, so local developers can go from raw idea to deploying scalable, AI-native apps in a matter of hours.

Wrap-up

• Singapore-born Sea is rolling out Codex across its whole dev org, with 87% of users weekly active (source: OpenAI's News, interview with David Chen)
• The value is thinking better, not typing faster; agents propose implementations, surface edge cases, and debug inside CI/CD
• Used well, it also improves discipline, paying down technical debt and building more resilient systems
• The developer's role shifts toward "system orchestrator," centered on product judgment and design
• Sea x OpenAI's Codex Hackathon Series runs from Singapore to Indonesia, Taiwan, and Vietnam

If you lead an engineering org, or you're building teams anywhere in Southeast Asia or the wider Asia region, this one's for you: it's a peek at what it means to redesign the organisation around AI rather than just bolting on a new tool.