Hot on the heels of the debunking of the previous Duffy presser, we have a new one to go over. Here’s a YouTube link until the official transcript is up — https://www.youtube.com/watch?v=ox-nQQZVSII.
Unofficial transcript: https://notes.rud.is/n/u2qopdj4qaaykp6vnz722jhcjm/
If you were anywhere near an airport Friday, you already know how that day went. Friday afternoon, Transportation Secretary Sean Duffy stood at a podium and told America the airspace shutdown was an old-age story: aging TDM/analog copper “got overloaded with data,” the “old hardware went down,” and — this being the part that mattered to the room — “all of the new equipment that we put in, none of that went down.” (None of it did. That’s the problem.)
By Saturday, that story was dead.
The replacement: L3Harris was doing a copper-to-fiber conversion at a remote FAA site, a (new) router installed during that work flooded the network with roughly a thousand connection attempts every four to five seconds, and the cascade took out 17 of 22 en-route centers. Duffy’s own framing of the open question — “whether we had incompetence… user error… or a faulty router” — is a wholesale rewrite of Friday’s bizarre narrative, and it quietly moves the suspect component from the legacy gear to the equipment the conversion crew just installed. Even the age of the implicated hardware drifted: “1980s equipment” on Friday became routers “manufactured in 2001” on Saturday. The technical narrative is still shifting under questioning, which is exactly what you’d expect when the first version was assembled for a press conference and the second one was assembled under it.
Both versions, though, end at the same place: give us money. That message is the singular takeaway vs. what we should be talking about in detail: that a single router can cascade into a national shutdown, that change control at the FAA runs on paper, and that this exact failure mode already happened in 2009. None of that got a sentence at either podium.
What Actually Checks Out
There is no evidence of a cyberattack or of AI involvement, and that claim is consistent across both briefings and all reporting; SMART (the “AI” tool rolled out around November 10) is a prompting/advisory layer, not something wired into control systems, so “no AI” is plausible and, so far, uncontradicted. The Saturday impact figures — 4,000 to 5,000 cancellations expected, about 12,000 of Friday’s 50,000 scheduled flights unrecovered, roughly 3,000 general aviation and 9,000 commercial — line up loosely with the FlightAware/Cirium-style counts for Friday (9,000–12,000 canceled, 25,000 delayed). Systems were rebooted and are stable, with full schedule integrity projected by end of day Sunday.
Bedford also admitted (to be fair, with commendable candor) that maintenance is logged on paper (paper!), which meant it require[ds] manual records searches just to figure out who touched what, and that the FAA still doesn’t know whether this was contractor error or faulty hardware. And then there’s: “We have one federal telephone infrastructure; we don’t have two. When the routers cannot communicate data, the redundancy is limited.” That is both the most damning and somewhat shocking admission of the weekend — the national airspace runs on a backbone with no real redundancy — buried in an answer about telecom plumbing.
The Money (Still) Never Adds Up
The administration’s dollar figures cannot all be true at once: a “$31B itemized total,” a “we need $10.1B more” (Bedford, Saturday), a $30.1B joint Duffy/Vought letter on November 13, an “I want $50 billion” from the same podium twelve hours earlier, and $12.5B already appropriated — of which only about 38% is even committed. Twelve and a half plus ten point one is $22.6B, not $31B. Three different asks in 72 hours, from the same administration, about the same system.
The “$50B” seems to be a (sad) attempt at an initial negotiating number, not a cost estimate, and anyone citing it as “what it takes to fix the system” is being misled, doing the misleading, or a moron. When Congress is told the solution is $10–50B while the thing that actually broke was process and oversight on money already appropriated, the funding request rapidly turns into (at best) a partisan agenda (at worst, it’s another grift attempt).
”It Wasn’t the Modernization Money”
The copper-to-fiber conversion is the FTI modernization L3Harris has been paid to perform since 2002 — the front line of the very upgrade the $12.5B pays for. Calling it “ordinary course-of-business sustainment” is how a program protects its image after its own work took down the national airspace on the Friday before Thanksgiving travel.
The Most Damning Answer of the Weekend
“It was done during the day because it was a remote site; it wouldn’t affect the wider system.” Yes, that’s their answer, and further demonstrates the management incompetence at hand: no model of blast radius, no containment, a Friday-afternoon change on the shared national backbone, followed by the discovery (their words) that “we did not know a single router could have this impact.” Bedford’s follow-up, “we’ll re-evaluate when we do this work,” is admittance that their change control procedures are woefully inadequate, especially with Thanksgiving weekend and a full aviation system still dependent on it.
2009 All Over Again
The parallel is brutal (something I emphasized on Mastodon at least 4 times on Friday): in 2009, Harris Corporation technicians misconfigured a router during a fiber conversion and knocked out 129 facilities, and a DOT inspector general flagged it. Same contractor lineage (Harris became L3Harris), same failure mode, 17 years later, on a Friday before Thanksgiving. Duffy’s promise to “hold people accountable” and his jab that “some contractors take our money and provide poor service” names no mechanism, no personnel, no debarment, no timeline — because the uncomfortable truth is that the FAA specifies, accepts, oversees, and runs change control on this exact work. If this were Halloween, it might be ok for a government contract-management failure donning a vendor-failure costume, but it’s not.
And the “copper was being discontinued, we had no choice” defense is only (barely) half true. Carrier TDM sunsets are real. But it omits that the FAA chose the cheapest, fastest path with no segmented test bed — which is what “the worst-kept secret in government” (Bedford’s own phrase) usually means: a known weakness, no redundancy, deferred replacement, blamed on Congress afterward.
What to Believe Until the Records Are Public
“We conclude it wasn’t cyber because the system recovered once we took the location offline” is absence of evidence in the first 36 hours presented as proof, and a potentially reasonable working conclusion presented as certainty (and one that quietly concedes they had no technical explanation at all on Friday). “We are back at full operations” was announced in the same breath as 4,000–5,000 cancellations; “full operations” means stable, not recovered. The progress numbers drift (68% in Vought’s letter, 69% on Saturday, “over 70%” elsewhere) and coexist with “we don’t yet know what broke” — the program is claiming rapid progress and structural failure at the same time, with no published root-cause report.
Friday blamed age; Saturday blamed a router installed during routine maintenance; both days used the crisis to demand money that was already partially appropriated. Until the root-cause report and the L3Harris change-control records are public, treat every official “we know what happened” sentence — from any official, on any day of the week — as (at best) provisional. And treat the $30B, the $50B, and the $10.1B as what they are: advocacy or grift, not true engineering estimates.
No comments yet.