This IEEE Spectrum article argues that recent multitrillion-dollar losses from software failures are not the fault of unpredictable bugs; they stem from repeatedly neglecting well-known risks.
Veteran analyst Robert N. Charette points out that over decades, organizations keep making the same errors: underestimating complexity, setting unrealistic budgets and schedules, skipping thorough testing, and trusting vendor promises without accountability. These shortcomings not only doom large IT rollouts but also erode trust, from payroll systems for tens of thousands of workers to critical infrastructure and public services.
Part of the problem is that software project managers rarely face legal liability for failures. In sectors such as medical devices, strict regulation forces proper testing and accountability, but for most enterprise IT projects, there are no such constraints; failures are treated as mere cost overruns.
That leads to a dangerous mindset: cutting corners feels acceptable because consequences feel abstract or distant. When problems arise, cascading system crashes, faulty paychecks, or long-term instability, end users feel the impact, but organizations seldom internalize the full cost.
By framing recent large-scale disasters as avoidable and repeating because of human error and institutional negligence, the article urges a rethink: software should be treated as critical infrastructure, with proper planning, realistic timelines, rigorous testing, and accountability, not as a quick, disposable commodity.