An OpenAI leader says AI is forcing companies to rethink their software playbook (www.businessinsider.com)

OpenAI’s Asia-Pacific go-to-market lead Andy Brown says AI is not only changing what products companies build but how they build them: teams are moving away from long sprint cycles toward an “always-on” product-release cadence that mirrors the rapid update tempo of AI firms. Brown pointed to OpenAI’s Agent Builder — a drag-and-drop developer platform built in about six weeks, where roughly 80% of the code was generated by a model — as evidence that model-assisted development is compressing timelines and enabling multiple releases per week. He urged engineering leaders to experiment with these tools as model capabilities now advance on the order of months rather than years. For the AI/ML community this signals a structural shift in software engineering and MLOps: faster iteration demands new pipelines for model-in-the-loop development, continuous integration/deployment, automated testing, monitoring, and governance to manage risk at velocity. The industry context is clear — OpenAI, Google and Anthropic have compressed their release cycles (GPT-4o updates, Google I/O announcements, Claude Sonnet 4.5 arriving four months after its predecessor) — underscoring competitive pressure to ship quickly. Practically, teams will need to rethink roles, toolchains and safety guardrails to harness model-generated code and rapid feature rollout without sacrificing reliability or compliance.