Post

Conversation

Ari Schulman

@AriSchulman

Re-upping this:

Why didn't the FAA stop a crash that was "waiting to happen"?

From thenewatlantis.com

11:45 AM · Feb 25, 2025

26.7K

Views

Post your reply

Ari Schulman

@AriSchulman

Feb 25

A number of remarkable details to add here from a Feb 13 NYT article: First, National Airport workers were so worried about a collision between a helicopter and a plane that in 2018 "they formed safety groups in attempts to prevent a crash."

In July 2018, a plane landing at Reagan nearly crashed into three military helicopters. The FAA began using the incident "as a case study in a nationwide training course for air traffic controllers." It happened at the same spot as last month's collision.

There was even an internal report in 2019 — seemingly issued by that ad hoc working group? — warning "that recurring controller error increased 'the risk of a collision and loss of life'" at Reagan.

But the Times piece describes the discussion as focusing on controller error. I think this adds weight to my claim — the safety system was working against itself here, using process critiques to make it *harder* to see the obvious: that chopper path should not have been there.

Times piece is here:

‘Risk of a Collision and Loss of Life’: D.C. Crash Warnings Were Years in the Making (Gift Article)

Concerns that a deadly collision could occur at Reagan National Airport had long been building. But attempts to draw attention to potentially dangerous conditions sometimes went unheeded.

In case you were thinking "surely it's safer than ever to fly at Reagan after the crash, as they must be on hyper-alert now"...

I've already flown since witnessing the crash, fwiw, but I am going to be making the schlep out to Dulles for the foreseeable future.

DCA has core design, not just process, problems: the helicopter flight path (only partly shuttered, & with trial balloons about reopening already being floated), crossing runways... The public needs assurance on this more than greater penalties for erring pilots and controllers.

Evidence so far strongly suggests human error as a factor in both the DCA crash and the Midway near-miss this week: pilots misheard controller instructions. But any system where 1 single error of that kind is all that stands between safe and mass fatality is a design failure.

Video of yesterday's go-around at DCA:

0:33

From

Excellent piece ari. But I'm curious- you keep referring to how organizations can't see problems when they are so big that they're existential. What problem here was like that? If anything this seems like it was more on the military for insisting on doing the helicopter flights

Pretty classic case of “we should not optimize/improve a thing that should not exist.”

240

Discover more

Sourced from across X

Harrison Kinsley

@Sentdex

14h

My personal findings so far with 3.7 vs 3.5 sonnet: 3.5 needs it's hand held to be guided through what to do. 3.7 needs its hand held as a leash to stop it from building an entire SaaS business when you asked to create a database.