Or more simply, because that's what "controversial" (from the prompt) means: something few believe that most don't.

I hear these critiques a lot, and whilst I'm inclined to agree, I have never had an answer to: how do you test that ? Basically a captcha most humans can pass, that an LLM never will ?

We could call it the

The Universal Reasoning In Novel Groundings

test

gyroscopes drift over time, that's why typically you need a kalman filter to combine the strengths of various sensors, magnetometers, where available, are bad at resolution, but don't drift.

MMs have their own set of issues though and calibrating them and knowing how the environment can influence them is a requirement for being able to use them productively.

true, sensor fusion of accel/gyro/mm only gets you so far, though of the 3, the mm is the only one that doesn't require integration, so while they are terrible at finding magnetic north in a given environment, the environment's big chunks of metal tend to not move, so with gravity, is the only external referential. That's why to do this stably you need something more, lidar/cameras+slam, the IR trick here, gnss (ideally with rtk/ppk/ppp).

> while they are terrible at finding magnetic north in a given environment

That's actually quite doable. The problem is that when the environment changes you will drift and then it becomes very hard to distinguish between actual heading changes and external influences. But if you can make a full spin you can (usually!) obtain magnetic North again. For drones this is a problem because they will drift and spin all the time. When you're on the ground and you have encoders on your axis you can do a lot better than with a drone.

RTK GPS is very nice to have but may not always be available (GPS itself may not always be available). I have some stuff in the works just for that. It does well in the lab but I still have (serious) problems in the field.

even with the calibration procedure, i've had nearly no success without going into industrial grade components, and as you point out, it's impractical.

gps also won't give you orientation really anyway, and indeed, is mostly outdoors, and rtk is needed for this resolution but also rather inaccessible to amateurs.

If you're working on indoor navigation, that's very cool, when I used to do it, in my context it was easier because we controlled the environment of the robots (markers on walls of a known colour type stuff). We also explored passive radar type arrangements for picking up wifi repeaters, but it became complicated to eg. ignore bluetooth.

Though in the case of this video, the wii sensor bar style approach seems fine, assuming you control reflections

Both indoor and outdoor. I have the relative positioning and heading working fine, it is the absolute positioning that is giving me a lot of trouble. I'll get there though, sooner or later or I'll die trying :)

It's insanely complex stuff, I've had to do a lot of re-studying math and this whole project has given me more respect than I already had for the engineers behind modern positioning systems.

the zoom scrolling and camera navigation in general, here at least (desktop mac chrome), seems like no one even tried it

If your experience was actually worse than my astonishingly janky mobile experience, that would be pathetic. Pretty bad. It looked cool for a second, and then when I tried to zoom in to where I live, I was instantly too irritated to bother.

Yes, and I insisted quite a bit because I was curious about a coast. I wonder if this is some kind of UI Neuralese, maybe the simulated scroll/touch attempts made sense to only it.

I’d guess it was a) an amalgam of a few good ideas that are pretty stupid together, or b) some hand-rolled techniques that got clobbered by a library imported for something else.

But also there's very little need today, it's very cheap to do, just complicated, a barrier that can be easily solved if there appears demand

because it's the unit of addressable memory that C is concerned by with sizeof, otherwise its values still depend on CHAR_BIT

I would agree if overflow on those types wasn't undefined behaviour, or unpredictable

I did buy it, all three, loved it. But somehow, broodwar is the one coming back on the scene more than it

I got one from a microwave, it opened vents. It is incredibly strong, also, takes soo much power and is so slow - although i suppose near heat it must work great.