Jeju Air jet black boxes stopped recording before crash: transport ministry

As an example, what is the contribution of air travel to climate change and wildfire, such as those presently underway in Los Angeles and environs, now costing > 12,000 homes and 24+ deaths? Aviation overall is about 6% of global petroleum consumption as I recall, and safety systems would then be a pro-rated fraction of that.

Another lesson of my IT career has been that a frequent cause of outages and risks is safety equipment itself. Failover routers that fail, failover power supplies that wipe out power, load balancers which lock up, firewalls that wall legitimate traffic, logging systems that fill disks, alerting and monitoring systems that overwhelm pagers and staff, etc., etc. Every technological mitigation has both benefits and costs, often manifesting in different timescales and conditions. We tend most often to adopt measures which promise quick and manifest results, or whose negative consequences are temporally distant and vague, but this also means that we underutilise those which offer distant and vague benefits (exercise, healthy eating, as the classics), or whose costs are manifest and immediate (again, exercise, various medical interventions such as exams and vaccinations, many safety measures and mitigations).

Commercial jet aviation is remarkably safe, but that's come through an intense focus on safety at all levels, from aircraft to operators to air traffic to ground response and most critically to thorough evaluation of incidents. One consequence is that many notable recent events have come from one or more of those factors themselves: pilots intentionally crashing aircraft (GermanWings 4U9525, 2015; probably Malaysian Airlines 370; EgyptAir 990, 1999), by passengers (most notably the 9/11 attacks in the US, 2001), or by safety or compensatory equipment (most notoriously the Boeing 737 MAX MCAS system, and the crashes of Lion Air 610 and Ethiopian Air 302). As with many other technical and economic phenomena, arguments of marginal utility strongly trend to developments reaching a pain threshold where harms increase to balance benefits, making naive analysis difficult.

(A realm in which this is probably evident to many HN readers is the tendency for software and online services to enshittify, in Cory Doctorow's marvelous coinage, to the point that they (attempt to) perch delicately on the threshold of fatal annoyance. A chief problem with such tactics are that the threshold itself is a remarkably unstable point, and induces its own further interactions and harms.)