In June 2026, OpenAI published research highlighting a surge in non-developers using AI tools to perform engineering tasks, reshaping traditional job boundaries.
Between August 2025 and June 2026, usage of OpenAI’s Codex by non-developers jumped dramatically: individual user token consumption increased by 137 times, and group consumption rose 189 times.
By June, Codex was responsible for nearly all weekly output tokens, capturing 99.8% of them.
Employees in legal, finance, and recruiting departments started handling coding tasks that a year before would have required developers to be hired.
OpenAI’s survey found more than 80% of users delegated tasks equivalent to at least 30 minutes of human work to Codex, and over 25% used it for work totaling more than eight hours.
The company’s April 2026 “AI Jobs Transition Framework” estimated AI could theoretically automate around 90% of tasks in the most vulnerable roles. In practice, workers engaged AI tools much less, tapping into less than a quarter of this potential.
A June report revealed that Europe has a smaller share of jobs at high risk of automation than the US, a factor crypto firms should weigh when choosing operational locations.
Reliance on a single AI vendor presents risks as well: if Codex’s API experiences downtime, price changes, or output quality issues, projects dependent on it could face serious disruptions.
