Keen's Clippings

2026-03-12-wired-openai-codex-race-claude-code

5 min read

Inside OpenAI’s Race to Catch Up to Claude Code — WIRED

Quick summary

This WIRED piece frames OpenAI as late to the modern coding-agent wave despite having pioneered Codex years earlier.

Main points:

  • OpenAI had an early lead with the original Codex and GitHub Copilot, but then shifted focus to GPT-4, ChatGPT, multimodal models, and agents rather than building a dedicated coding product.
  • Anthropic stayed more focused on coding, especially real-world messy repositories, and that helped Claude Code take off earlier.
  • OpenAI later had to scramble internally to merge separate coding efforts into a real Codex product, while also unsuccessfully trying to acquire Cursor and later Windsurf.
  • The article argues the coding-agent market is now one of the few places where enterprises are clearly willing to pay a lot for AI, so this became strategically urgent.
  • OpenAI’s recent catch-up appears to be driven by stronger reasoning models, command-line-native agents, aggressive distribution through ChatGPT/enterprise packaging, and heavy subsidization of usage.
  • The broader implication is darker: management may increasingly expect employees to use coding agents or risk becoming irrelevant.

My take

Pretty plausible overall.

The most believable part is organizational drift: OpenAI invented a lot of the underlying capability, but didn’t productize it fast enough because ChatGPT and broader AGI ambitions pulled attention elsewhere. Anthropic looked more disciplined here.

The article is also basically saying the winner won’t just be the model with the best benchmark. It’ll be whoever combines:

  • strong coding reliability
  • the right interface (CLI / repo-native workflows)
  • enterprise distribution
  • enough subsidy to get teams hooked

One eyebrow-raiser: some of the revenue and usage claims are sourced to unnamed insiders, so treat those numbers as directional, not gospel.

Full clip

Inside OpenAI’s Race to Catch Up to Claude Code | WIRED

Inside OpenAI’s Race to Catch Up to Claude Code

Why is the biggest name in AI late to the AI coding revolution?

Sam Altman sits with his legs pretzeled in an office chair, staring deeply into the ceiling. To be fair, the new OpenAI headquarters—a temple of glass and blond wood in San Francisco’s Mission Bay—seems to invite this kind of contemplation. A kiosk behind reception holds booklets that describe the “Eras of AI” as if they were steps on the path to enlightenment. Posters along the stairs mark AI’s milestone victories, like the time thousands of humans watched on livestream as a machine beat a top-ranked esports team at Dota 2. In the hallways, researchers pass by in sacred merch. One shirt reads “Good research takes time.” Ideally, not too much.

Altman and I are in an enormous conference room. The question I put to him is about the AI coding revolution—and why OpenAI doesn’t seem to be leading it. Millions of software engineers have started delegating their programming tasks to AI, forcing many in Silicon Valley to reckon with the automation of their jobs for the first time. Coding agents have emerged as one of the few areas where enterprises are willing to pay a lot for AI. This moment could, and arguably should, be the next triumphant poster along the stairs for OpenAI. But the name in big print right now belongs to someone else.

Anthropic, a smaller rival started by OpenAI defectors, has found runaway success with its programming agent, Claude Code. The product accounts for nearly a fifth of its business—more than 1 billion in annualized revenue, according to a person with direct knowledge of the matter. What gives?

“First to market is worth a lot,” Altman says finally. “We had that with ChatGPT.” But the time is right for OpenAI to lean into coding, he says. He thinks the company’s AI models are now good enough to power very capable coding agents. “It’s going to be a huge business—just the economic value of it, and then also the general-purpose work that coding can unlock,” Altman says. “I don’t throw this around lightly, but I think it’s one of these rare multitrillion-dollar markets.”

OpenAI’s first Codex effort in 2021 powered GitHub Copilot, but the team dissolved as focus shifted toward GPT-4 and later ChatGPT. OpenAI then spent much of 2023–2024 prioritizing multimodal AI models and agents instead of a dedicated coding product.

Anthropic, meanwhile, focused more deliberately on coding and trained on messier real-world repositories. Claude Sonnet 3.5 and later Claude Code helped cement Anthropic’s lead.

The article says OpenAI later formed multiple internal efforts around coding agents, including command-line-based approaches, and eventually merged them into a sprint to ship Codex. It also details failed acquisition attempts involving Cursor and Windsurf.

According to the piece, OpenAI’s newer momentum came from stronger reasoning models, command-line-native workflows, aggressive enterprise distribution, and subsidized usage to win over developers.

The article closes by arguing that the competition is now very tight, but that the bigger issue is no longer which coding agent wins—it is how rapidly these tools are changing expectations around white-collar work, software development, and automation.

Graph View