"Hey Siri, find me gas stations along my route" routinely fails now. That worked great for years. Total clown show.
329 karma · joined July 3, 2026
In some sense it felt safer than the 172 because I could fly a higher, tighter pattern. It was easier to lose a bunch of altitude on final by putting it sideways in an aggressive slip. The 172's small rudder didn't let it get as sideways, so it was a little more constrained. The 172 did have functioning instruments and dual radios, not just a handheld. It also had flaps. I think it was probably 10-15yr newer? Still a relatively "old" airframe. I probably had 5x or more as many hours in the Clipper so all that could just be biased towards familiarity.
Point being, old plane does not mean unsafe plane. It's not like you're riding around in some rusted out hooptie in the sky. There are rigorous mandatory inspections and maintenance schedules for every single component, which results in mechanical failures being really rare even in aircraft that are nearly 100 years old.
None of that obviates the need to know what you're doing, of course. The pilot in command needs to have control of the plane. I'm pretty skeptical of systems that kind of try to smooth over that sharp edge, I can understand the altruistic motivation behind it, but there's likely a side effect of reducing the perceived risk. That's a problem, because it makes people complacent. So in terms of overall safety, it might not be a win? You're trading on average higher complacency and lower situational awareness for maybe some better outcomes in rare situations? Probably not the best trade.
But I also wouldn't want anything to do with the new glass cockpit stuff, way too easy to just lose all your instruments at once. I'll take properly maintained, independent mechanical or electromechanical systems over digital ones every time for things that actually need to work.
Digital systems always seem to accumulate tons of complexity ("features") rendering them extremely brittle. Until we can figure out how to stop doing that, I prefer to avoid depending too heavily on them for important stuff.
The Cirrus parachute system seems pretty simple and trustworthy, that approach feels more directionally correct. Not much more complex than an automotive airbag.
I was pretty confident barring some catastrophic mechanical failure I could probably get it on the ground or into the water (or trees) in one piece. The big question remains though: what now? If you're knocked unconscious (likely--lap belt, big flat instrument panel at face height) then two things can happen:
(1) if in the water, you drown before you can get out.
(2) if you're on land, the fuel tank right underneath that instrument panel or the ones up in the wings might torch you before you can get out.
It's probably about right to compare the risk profile to motorcycle riding.
I don't fly anymore, but I see similar behaviors developing on the water. People who use chartplotters for navigation don't have a very good understanding of what's going on around them. If you're just interactively following a computed lay line or a plotted course you don't really need to know what the currents are doing. You don't need to know what the hazards are. You're not actively engaged in navigating the boat.
Then what are you supposed to do when your GPS craps out on approach in low visibility and heavy weather?
Obviously the answer is "fly the airplane". Which means you look at your artificial horizon, set trim and power for max climb, get to a safe altitude and figure it out.
If you're paying attention you already have your chart folded to the airport you're landing at, so it shouldn't be too hard to orient yourself and figure out what's next. Either you'll be making another attempt at this airport without GPS or you'll be going somewhere else.
There's never an excuse for failing to fly the airplane. That's basic airmanship. You need to be able to do it regardless of whether your GPS is working.
Even in the absence of GPS jamming, you can effectively lose GPS for all kinds of reasons in a plane--a connector failing somewhere on the back of your panel, a wire chafing through, etc. GPS is nice, but it's probably a lot less reliable in most light aircraft than the other options.
It's been 20yr since I've flown a plane, so maybe attitudes have changed, but my instructors always stressed the importance of always being able to fall back quickly and seamlessly to a paper chart, your watch, flight computer, and basic instruments. Because you will lose everything else at some point.
I made a lot of use of the TI-89 era CAS in college. But IMO the TI-83 era manuals taught me more about both math and computers than the subsequent generations could have.
Sailing would be more fun, IMO. Might take a quarter or half a year though.
> "For a submarine to swim as well as a fish it has to be as light as fish".
These are false equivalences. The post you're responding to defined intelligence as learning rate. LLMs unequivocally do not learn. You can disagree with OP or agree, but what you have done is out of bounds. You're implicitly claiming that LLMs learn, albeit differently from humans. This is categorically false, unless you count training as some kind of "learning".
That's of course absurd, almost no user of an LLM also trains it. Instead they rely on queries submitted to pre-trained LLMs ("inference"). And don't bother yapping about context windows, it's just not anything like learning.
So now the problem is highly visible. There's only one solution I can think of, and that's to hold individuals (not corporations) responsible for professional malfeasance.
So having the prompt available would be useful even in the degenerate case.
The rational choice would be to cut your losses and stop reading at that point. Once you realize zero effort went into the prompt, there's no longer any reason to read the output. The age old truism still applies: garbage in, garbage out.
If you think it's worthwhile, close the issue with a comment: "please rework this and show your work next time". Otherwise just close it without commenting and move on.
For a fun rabbit hole to dive down, check out Doble steam car technology or the Besler airplane engine.
You just have to trust your colleagues won't edit the tests in such a way that vandalizes the invariant. That's why you can never trust an LLM to edit a test. They're notorious for tweaking tests such that they pass.
[edit] preemptively, if you're going to claim "but this isn't art so irrelevant" then I claim bull fucking shit. It's the same problem no matter how you slice it whether it be engineering, support, art, or medicine. Get real. Look inward. Touch grass. If you think AI is a good thing for your profession--whatever that profession might be--you're probably a delusional psychotic. It's OK, you'll thank me later.
\m/_(>.<)_\m/[edit] This may need to be an integration test to be effective, there is a certain peril to mocking that could bite you here. But that's fine, we have the technology.