United Airlines grounds flights after system meltdown
allchronology.com
allchronology.com
There are plenty of real sources for this story:
https://www.wsj.com/business/airlines/united-airlines-halts-...
https://www.nytimes.com/2025/08/06/business/united-airlines-...
https://www.reuters.com/world/us/united-airlines-flights-res...
2. Didn't exist before 2023 (https://web.archive.org/web/20250000000000*/allchronology.co...)
3. Faces on https://allchronology.com/about/ look like they were taken from https://thispersondoesnotexist.com/
100k-ai-faces-3-1.jpg :)
But also factual errors, such as quoting the supposedly Democrat chair of the Senate Committee on Commerce, Science, and Transportation.
It seemed unlikely that a Democrat was left in that post, and I googled it, and indeed they weren't.
Lots of boilerplate platitudes, especially towards the end of the story.
Some of the quotes appear to be fabricated. I can't find the "aviation analyst" tweet, and I'm pretty sure Maria Cantwell hasn't commented (unless it was video/audio only, and this is the only outlet that printed it). She's also no longer committee chair, being a member of the minority party in the Senate.
It very closely follows the AP article on this - like they copied their homework but changed some words - https://apnews.com/article/united-airlines-flights-grounded-...
I figure that there has to be a forum out there where the nature of the software glitch is the focus, but I couldn’t find anything. If this story is AI generated, then I would very much like to read the sources it’s lifted from.
I did find a 2007 article discussing another Unimatic glitch that caused a stoppage which mentioned that the software dates from 1988:
https://www.aviationpros.com/home/news/10387920/computer-fai...
> The computer system, known as Unimatic, is essential to the airline's operation, providing flight plans for pilots, updates on maintenance information and crew schedules, among other flight information. United jets worldwide cannot take off unless it is operating. The original Unimatic system dates back to at least 1988, but it is updated "all the time," United spokeswoman Robin Urbanski said.
Which sadly leaves a gap in the market for AI slop. A few times recently I've tried to find a good news article to link to and had to chose between paywall sites, obvious AI slop or second-tier publications plastered with ads. I usually pick the third category.
https://onemileatatime.com/insights/highest-paid-airline-ceo...
He must know something, right? And it seems he knows more and more as the years go by.
It’s a perverted reward hack by leadership, and nobody - governments, shareholders, boardrooms, regulators, etc - seems to give a shit.
For example, because of a holiday, the system would be confronted with more customers than it can handle, and any little change (like a failing plane or changing flight-staff) could lead to a crash of the system. IIRC this was a problem they had some years ago.
Agreed.
Frequently the actual issue related to these big system outages is cost-cutting/outsourcing/lack of maintenance.
During previous similar outages it has taken days to track down staff and equipment once the systems are up again. During this time delays and cancellations continue. It's possible the system is much faster to bootstrap given correct locations of things from the start.
You wouldn't want to fly an airplane whose weight and balance has been miscalculated (or maybe properly calculated but can't vet if it has)
https://en.wikipedia.org/wiki/Cubana_de_Aviación_Flight_0972
>The system manages flight information like crew scheduling, weight-and-balance calculations, and aircraft movement logs.
But your other comment seems more likely: keep the people/equipment where they are to reduce the recovery time down the road.
Obviously air traffic control wants to know the exact location of the plane but they have radar and the airplane has a transponder for that purpose. I don't know if the airline cares so much, other than the estimated arrival time to account for delays.
Obviously planes already leaving the gates are fueled up so that's most likely not the case here but that's one example of how those systems can be still integrated. Now, today - with far more computing power - it is very well possible that that whole system runs on the plane side, but the amount of external data and various exceptions and almanac information that was pulled in for those computations was pretty impressive. Most likely that sort of thing is now done on an iPad or something similar.
Thus, the only reason to recall a plane that's pushed back already would be to avoid problems that would occur only after it has landed. I would be surprised if the external systems were needed for the flight itself.
The 777 has automated presentation of some relevant electronic checklists but these definitely are NOT automated because the whole point is human-attended verification.
It's no longer possible to not be a nightmare. Companies no longer employ devs for life to keep the system functional and up-to-date, and (most probably) no devs would stay in such a role for their entire career.
When nobody knows how the system works, it can only be a nightmare: they keep it barely functional with marginal patches, accumulating debt until it finally crashes in such a way that it can't be fixed, then a new team comes in, designs and builds a new system, patches some major flaws, and then leave, and the situation repeats for another 10 years.
Why would a 30-year-old car be flying a plane? That makes no sense.
While i sympathize, the world does rely on high quality 30+ year old software. I think it's time, as an industry, to stop seeing software as disposable and start designing for longevity.
EDIT: My comment was specific to the 1995 toyota being a far simpler vehicle than current models, hence likely to be easier to fix....but, of course, the flipside is that potentially the older the vehicle, the likely less availability of certain parts.
And if I had enough disposable income to upgrade to something with side airbags, it’d compensate for my lack of trust in others on that drive too! (little tongue in cheek here, just voting for letting some money leave a bank account for reasonably modern safety features—even some ultra-high net worth folks seem not to care at all about those)
A well maintained 30-year-old car can drive cross country. There's a big difference between car thats old with a ton of known defects that the driver works around and a car thats old that had it's defects fixed when they came up.
That being said, the reality is that requirements change. Load changes. The world around the software changes. Systems need to be resilient, yet flexible enough to be maintained but not replaced over decades.
Whilst it would be a major upheaval to switch to a clean room engineered implementation using 2025 best practices it would at least increase the talent pool that can work on it effectively.
There does likely come a point where it is cost effective to rebuild it both in reduced unplanned downtime and reduced maintenance costs.
The good news is there's always a contractor willing to promise the world and deliver something that doesn't work in 5-10 years. Your internal team, who could have finished the original job in a couple years if they'd been funded, will be the ones that end up making that delivered system actually work. But you'll tell everyone that it was the contractor's high quality output that did the trick, because saying they failed would hurt your career. In 15-30 years the system will get replaced and your successor will hear about how great <contractor> did the first time, and they'll get another shot at failure.
[1]: https://www.planespotters.net/airframe/boeing-767-300-n641ua...
The main difference between software and physical objects like cars is that they degrade with the passage of time (due to wear, corrosion, etc...). If we would magically be able to get a brand new 30-year-old car, it would make absolute sense to use one for a trip where reliability is paramount, as the failure modes of such a car are better understood compared to a brand new design, and can be mitigated.
Also, how old are the planes?
Stop trying to put blame on labor.
The issue here is United penny pinching for decades and now they are stuck with their in house created mess. The right move for them could be to buy one of the smaller upstarts with less issues and then just roll out whatever the small company is using and retire the creaking old mess that just failed them.
Any project to slowly modernize or update that is doomed to fail.
You see the same in the fintech world where small relatively new banks are running circles around their older competitors.
In the energy world, a British company called Octopus is actually licensing their platform (Kraken) to lots of energy companies around the world. One of the interesting things about companies that move to that platform is that they become a much more attractive target for M&A as well; because the job of migrating all the old customers is easier if both companies that are merging use the same, modern platform. That's the hard risky part of any merger. Apparently, in some cases the decision to move to Kraken was motivated as a preparation for this.
I'm mentioning this because that's something that could work in the airline industry. Or United can just zombie on for a few more years and then end up getting absorbed into some other, more successful company.
This stuff must be costing them at all levels in the company. I can't imagine it's very efficient. And the airline industry is pretty cut-throat at this point.
United* was one of the first major airlines to embrace modern tech. Their phone app was lightyears ahead of anyone else (I haven't revisited others in the past couple years, maybe other airlines caught up, I dunno)
When I use to spend way too much time on frequent flyer boards it seemed United was the only one spending money and embracing their tech center to innovate and update old systems, both software and hardware. It takes time to modernize, some changes also have to happen in all the many airports they do business in, not to mention all of the other fragmented systems they were working to better integrate. Since they are not a tech company I don't expect -- or want -- them to "move fast and break things".
Singling them out as penny pinching on tech is simply disingenuous.
*United & the *A group are my preferred airlines; bias is likely as well as a keen interest in what they are doing.
Let’s hold companies accountable for safety practices, not just the calendar age of their fleets.
Except maintenance of software looks more like continuous architectural review and operation of an electrical grid. Calling it maintenance is a little narrow.
Now they're all deregulated and they are highly incentivized by their shareholders not to overbuild capacity because it costs money and provides zero return on investment.
The upshot is that these industries operate with zero excess capacity. They work fine when everything goes as planned, but the minute something breaks or a storm happens, everything cascades into a steaming pile of shit.
In the case of the airlines, competitive market forces previously provided some incentive to invest in reliability but so many airline mergers have been allowed to happen that there's effectively no airline competition in the US any more.
Those shareholders are their actual customers. Air traffic is just an annoying necessity.
Because software doesn't show wear. On the contrary - a system that's been running stable for 30 years probably has fewer bugs than one that's been running since last week. They are running old software on new computers.
Just wait till I find out how old this SABRE is!
https://en.m.wikipedia.org/wiki/Sabre_(travel_reservation_sy...
I interviewed to work on the C version about 35 years ago.
Anybody else tired of this broken hypercapitalistic system that only serves to benefit the few? Enshittification is a symptom of a much more disgusting disease.