When the Sky Said No: How the 737 MAX Crashes Reshaped How We Think About Software Safety
On October 29, 2018, Lion Air Flight 610 took off from Jakarta, Indonesia and never came back. Within minutes, all 189 people on board were gone. The cause wasn't pilot error or a freak storm — it was a software system called MCAS, a flight control automation designed to correct the plane's nose angle, that activated based on a single faulty sensor and couldn't be overridden in time. A piece of software, deployed without adequate pilot training or transparent documentation, brought down a plane full of people.
That's a devastating thing to sit with. And for those of us who build software for a living — whether it's an enterprise platform, a customer-facing app, or an internal automation tool — it's also a profound reminder of what's actually at stake in the decisions we make. The 737 MAX disaster exposed a culture where business pressure to ship fast had quietly overruled engineering transparency. Stakeholders didn't fully understand what the system did. Operators weren't properly informed. Assumptions were baked into the design that nobody had stress-tested against real-world edge cases. Sound familiar? It should — versions of this dynamic play out in tech teams every single week, just usually without the catastrophic consequences.
Here's the uncomfortable takeaway for anyone leading a software project: automation doesn't eliminate human responsibility, it concentrates it. When you build a system that makes decisions on behalf of users — whether it's an AI recommendation engine, an automated deployment pipeline, or a workflow tool — you are responsible for what it does when things go sideways. The engineers and leaders who worked on MCAS weren't villains; they were caught in a system that rewarded speed over clarity. Our job, every day, is to build the kind of culture that pushes back on that pressure — that asks "what happens when this fails?" before asking "when can we ship?" That question, asked consistently, is what separates good software from software that earns trust.