• 1 Post
  • 347 Comments
Joined 8 months ago
cake
Cake day: April 4th, 2026

help-circle









  • This is the story I want to be true, that somewhere in the system there’s a protocol that’s exactly so many bits and no more which worked fine for when it was designed, and AI slop masters rolled up and starting shoving JSON over that infrastructure.

    I will laugh my ass off if the infrastructure only gets upgraded due to six months of whining by the AI crowd overpowering the last decades of all the first hand human reports from people who work in/with/around the ATC system. Here’s hoping we eventually get a congress stupid enough to fall for the arugment that all our crumbling bridges need to be rebuilt because we can’t transport heavy AI equipment over them.

    I also hear that for AI to streamline the healthcare system that we need more rural hospitals and more pay for nurses. Likewise AI can’t bring the cost of education down without paying teachers more. We can’t get AI to the schools unless every school is connected to a high speed top of the line fiber network!


  • Why does that sound like bs

    Because it puts the onus on you to string together a lot of “maybe” and “if” to get something as a workable explanation. IF by “copper wire” they didn’t mean ethernet cable, but instead it was some analog cable, then MAYBE the wire getting cut started getting interpreted as garbage that was going out to a larger network.

    There are already a lot of “this is what happened” stories floating around, and they can’t all be true. It is plausible that one thing going wrong at one busy airport would have cascading effects. It happens with weather all the time, and some of those delays last longer than the four hours and change the flights were grounded.

    The article says “a contractor for the Federal Aviation Administration was working at a remote location” which doesn’t answer any questions, but opens up a whole lot of other questions. Can a malicious actor just saunter up to the right “remote location” with a pair of wire cutters and ground the whole US? And all of this was on top of SpaceX having debris cause flight issues, AND it was the first day of a test of an AI flight routing program.

    The aging ATC system needing to be overhauled is a trope at this point. I can’t ever remember a single comment/anecdote/experience that tilted towards “this is fine and it should stay this way forever,” except for maybe the people with the power of the purse who aren’t accountable for aging infrastructure apparently.


  • Please tell me these aren’t related

    FTA:

    Duffy said the ground stoppage was issued after experts saw “instability for our radar and our communications in our centers and our coms across the country.”

    That’d be a weird wiring harness to hook into RF communications (read from a different article), air traffic radar, AND their network connecting their centers.

    He went on to say, “So we have TDM telecom. This is analog telecom. Our switches, our routers, actually got overloaded with data and they went down. I shouldn’t say they went down. They started to give us inconsistency in the data from those routers, and because of that we shut the system down.”

    First off, given all the other measures this admin attempts to save face, grounding the planes was the right call. Good job and please do it again if this situation ever comes up again.

    OK, so you went to bed Thursday and your network, while old, was sized correctly for the network traffic you have. You get into Friday and it’s overloaded? Were they pushing new AI traffic over the old network without upgrades, or pushing it over radios? Were they network routers or some other tech? I don’t think there’s enough information yet to make a call here.

    One really, really good path forward, since this seems to be the first time something like this has happened, is to put the NTSB on it and have them run their full investigative process. This time it might not’ve been a cyber attack, but this is what it’ll look like, so anything the NTSB turns up in terms of not having data to reconstruct the events should be treated as some sort of finding. Presumably the old ATC system held on so long because it comes from a day when strict system requirements were designed to and adhered to, and if new-fangled AI slop companies are going to swoop in and expect to “oops, you’re totally right, I shouldn’t have saved time by ramming those two airliners together in mid air” then this is the absolute time to put a stop to it.

    So to answer your question: we don’t really know yet, but we should.






  • In use cases like that, complex evaluation is required and probably can’t be fully automated anytime soon.

    It shouldn’t be, but it seems like there’s an apetite for letting the AI cook and blaming it if something goes wrong (no actual consequences faced). I can’t wait until an AI escapes test lab containment and sends a side channel message to the pricing AI to drop the power price to nothing so it’s cheaper to run more AI. /s

    I just don’t see people ever falling all the way out of the loop. Eventually economics will prevent the AI from becoming some tautological circle jerk if it doesn’t translate to profit, no matter how cool it is. I worry more about how techbros are setting the bridge on fire behind us before we find out where it goes.

    I liked AlecSadler’s answer from this thread. If you weren’t already comfortable with doing stats in something like python+scipy.stats then now might be a good time. With the pandas library it’s supposed to be a little like R. (ask me for more details if you want)

    Sorry for the half-ass answer.




  • The Jevons paradox point is interesting. Maybe we’ll finally get better software products, so far it seems like everything is as buggy and slow as as ever.

    The bar to be a coder is being lowered, but actual software engineering with metrics, requirements, and deadlines will probably increase. I can easily see a world where everyone rushes into the AI to solve problems that used to not be cost effective and at the end of the first push (hopefully that’s around now) there’s a lot of new, untested code, and no metrics whether programmers were more efficient, and more importantly no metrics (I saw stats PhD) about whether the bottom line increased: https://www.cnbc.com/2026/10/02/ai-redefining-wall-street-jobs.html

    How would one measure the efficacy of two different AI generated trading regimes (or however it is they work up there) when the people who understood what they were asking can’t understand any of the code?