Has the cost of building software just dropped 90%? (martinalderson.com)

A recent analysis suggests that the cost of software development may have dropped by as much as 90% due to the rise of agentic coding tools. These tools, equipped with advanced AI capabilities, can drastically reduce the time and labor required to develop applications. Previously, building an internal web app could take months involving coordination among teams, yet now, this process can be streamlined to mere hours with AI-driven tools capable of generating extensive unit tests and backend services. This shift not only reduces overhead but also aligns with Jevons Paradox, which posits that as production costs decline, demand typically increases. For the AI/ML community, this evolution signals a significant transformation in the industry landscape. While some might view this as a threat to traditional software development roles, the analysis argues that the true value now lies in domain expertise. The most effective use of these tools emerges from a partnership between knowledgeable business users and developers, allowing for rapid iteration and adaptability—essentially making software development more dynamic and less tied to manual coding. As AI tools continue to improve, the urgency for developers to embrace these changes is critical, as outdated perceptions could hinder progress in a rapidly evolving field.