Nobody Cares! (Science & Tech Edition)

Argent Stonecutter · Apr 1, 2021

Innula Zenovka said:
Do Waze/Google maintain an up-to-date database of all local speed limits? I ask because I don't know, but since they're determined locally in many cases, I'd be worried about keeping it current.

It doesn't have to, since I'm not suggesting getting rid of the existing scheme of reading signs, just using it as the primary source where available. Also, probably, not allowing speeds higher than the default for unimproved roads without a high speed database entry.

And while allowing local signs in some circumstances to override it, as you suggest, might be a solution, would that be a sufficient legal shield in all jurisdictions? I doubt it.

I don't see how it provides more legal exposure than just going by image recognition, which is the problem at hand.

Innula Zenovka · Apr 1, 2021

Argent Stonecutter said:
It doesn't have to, since I'm not suggesting getting rid of the existing scheme of reading signs, just using it as the primary source where available. Also, probably, not allowing speeds higher than the default for unimproved roads without a high speed database entry.

I don't see how it provides more legal exposure than just going by image recognition, which is the problem at hand.

Yes, but the question is whether it provides sufficient legal cover to make the driverless cars a viable commercial proposition in the first place.

That's what worries me -- quite apart from speed limits, somehow the AI is going to have to handle the trolly problem in a manner that satisfies both the law wherever the system is sold and used and the companies who offer third-party cover to drivers.

It's also going to have to reassure potential customers who are worried that their car might decide that the lesser of two evils would be deliberately to plough into a wall at speed and kill the driver in order to avoid ploughing into a queue at a bus stop and killing them.

Argent Stonecutter · Apr 1, 2021

Innula Zenovka said:
Yes, but the question is whether it provides sufficient legal cover to make the driverless cars a viable commercial proposition in the first place.

I'm not going anywhere near that. My point is that using stored speed limits as the preferred source of truth is the only sane choice from a legal standpoint.

the AI is going to have to handle the trolly problem

No it's not.

Applying trolley problems to autonomous vehicles is a fallacy. If the car is ever in a situation where it is has the potential for creating a fatal accident, it's already so far out of its design rules that there is no chance it has enough control over the situation to effect a choice. Either its got no agency (someone has overridden it) or the programming was fatally flawed and can't be trusted.

In your example, if the car is travelling fast enough that it can't stop before hitting a group of people, it is already outdriving its sensors. That means it's already in a state that it could have just as easily plowed into a wall at speed. Or a bollard. Or a piece of equipment. It's got no potential mechanism to "choose to" plow into anything else because it's operating outside its design parameters in the first place.

Innula Zenovka · Apr 1, 2021

Argent Stonecutter said:
I'm not going anywhere near that. My point is that using stored speed limits as the preferred source of truth is the only sane choice from a legal standpoint.

No it's not.

Applying trolley problems to autonomous vehicles is a fallacy. If the car is ever in a situation where it is has the potential for creating a fatal accident, it's already so far out of its design rules that there is no chance it has enough control over the situation to effect a choice. Either its got no agency (someone has overridden it) or the programming was fatally flawed and can't be trusted.

In your example, if the car is travelling fast enough that it can't stop before hitting a group of people, it is already outdriving its sensors. That means it's already in a state that it could have just as easily plowed into a wall at speed. Or a bollard. Or a piece of equipment. It's got no potential mechanism to "choose to" plow into anything else because it's operating outside its design parameters in the first place.

I think insurers will need a lot of convincing before they agree to underwrite that risk, though.

Veritable Quandry · Apr 1, 2021

It seems that someone authoritative would need to be responsible or a speed limit database. DoT or another Federal agency would be an obvious choice, but insurance companies may need to do it themselves. Google on their own should not, but if they worked with an industry group alongside other parties working on autonomous vehicles their resources would be useful.

Any party that maintains a speed limit database for road use would have to be seen as both authoritative and neutral, and be willing to take on the role and responsibility.

Sovereignty · Apr 6, 2021

How immunologists amuse themselves.

Dakota Tebaldi · Apr 24, 2021

Argent Stonecutter · Apr 24, 2021

#occupypluto

Saturn's rings are askew. It would have been better to tilt Saturn 90 degrees since it is anyway. Also this configuration is not gravitationally stable.

danielravennest · Apr 24, 2021

Argent Stonecutter said:
#occupypluto

Saturn's rings are askew. It would have been better to tilt Saturn 90 degrees since it is anyway. Also this configuration is not gravitationally stable.

Right, we all know this would immediately collapse into a larger Jupiter, because it has more mass than the rest put together. But that's not the intent. It's to show the distances between planets are very large. We see them as small dots in the kind of telescope you might personally look through, but if brought close, they would all fit between Earth and Moon, a distance you can at least imagine while looking at the Moon. An in that location they would be overwhelmingly large, with Jupiter filling 90 degrees of the sky.

Essence Lumin · Apr 24, 2021

Argent Stonecutter said:
#occupypluto
Also this configuration is not gravitationally stable.

Duct tape fixes everything. Well, that and wd40.

Argent Stonecutter · Apr 24, 2021

Essence Lumin said:
Duct tape fixes everything. Well, that and wd40.

Curses, I forgot duct tape.

Veritable Quandry · Apr 24, 2021

I think this one calls for bailing wire.

danielravennest · Apr 30, 2021

China just launched the first section of their new space station. Unfortunately they left a 20 ton upper stage in a decaying orbit. It will likely come down in the next two weeks, somewhere between 41 N and 41 S.

You can track it's current position and orbit parameters here

Currently a 176 x 368 km orbit (low and high points). When it gets down to ~130 km, it will likely be on it's last orbit.

[UPDATE] May 1st, 6:27 pm - now 175 x 359 km. The way orbit mechanics works, a change in speed affects the *other* side of an orbit. Since the air is much denser at the low point in the orbit, its losing the most from the high end. Eventually it will be the same height all around, at which point it just spirals in evenly.

May 2nd, 5:16 pm - 174 x 344 km
May 3rd, 4:33 pm - 172 x 333 km
May 4th, 3:54 pm - 169 x 310 km Orbit decay speeding up. Only a few days left before it falls down.
May 5th,11:05 am - 166 x 294 km
May 6th, 4:34 pm - 164 x 274 km
May 7th, 6:50 pm - 157 x 241 km - Current re-entry prediction is 04:19 UTC on 9 May +/- 8 hours (5 orbits), so soon.
May 8th, 10:21am - 150 x 215 km - Predicted re-entry is 03:22 UTC on 9 May +/- 4 hours

Innula Zenovka · May 4, 2021

Innula Zenovka · May 4, 2021

danielravennest said:
China just launched the first section of their new space station. Unfortunately they left a 20 ton upper stage in a decaying orbit. It will likely come down in the next two weeks, somewhere between 41 N and 41 S.

You can track it's current position and orbit parameters here

Currently a 176 x 368 km orbit (low and high points). When it gets down to ~130 km, it will likely be on it's last orbit.

[UPDATE] May 1st, 6:27 pm - now 175 x 359 km. The way orbit mechanics works, a change in speed affects the *other* side of an orbit. Since the air is much denser at the low point in the orbit, its losing the most from the high end. Eventually it will be the same height all around, at which point it just spirals in evenly.

May 2nd, 5:16 pm - 174 x 344 km
May 3rd, 4:33 pm - 172 x 333 km