OpenAI published a paper claiming 98% of its employees now use Codex, its AI coding agent, up from 40% in August 2025. Non-developer usage reportedly grew 137x and active users grew fivefold in the first half of 2026. However, every metric in the paper is self-reported by OpenAI, a company with direct financial incentive to promote Codex ahead of a potential Q4 IPO. Critics note the data lacks independent verification, doesn't account for whether employees are incentivized to use the product, and measures input volume (tokens, task requests) rather than actual productivity outcomes. The paper is framed as evidence of a broader industry shift from chatbots to autonomous agents, but the gap between OpenAI's near-universal internal adoption and external adoption rates of under 1% raises questions about how representative the data is.

4m read timeFrom thenextweb.com
Post cover image