I'm thinking about people in general, not technically inclined individuals like those at Slashdot. Slashdot's audience is far from general, so assume people are tech illiterate, since that is a target demographic for this device.

I suppose it would mean enabling other websites to make purchasing ebooks on the device a trivial task. Maybe add a selection menu of shops (that can be modified of course) in place of the way you buy on amazon, similar to how MS was forced to let people choose which browser to use in the EU.

I can't help but think some of the higher ups will have a problem with this. The EU has started making a push for "digital sovereignty" so it seems like it would be a massive conflict to turn over control of key infrastructure like communications to a foreign company.

However, I could see this going through if the deal is for one or more EU companies to use them to bounce their own connection signals. Relying on foreign companies would still represent a potential single point of failure but this also creates the possibility of developing your EU satellites while building out the local infrastructure and accruing users. Using starlink/amazon satellites on a temporary basis would buy EU officials the time to make the economic case to develop EU owned satellites. It would effectively solve the "horse before the cart" issue that comes with deploying expensive satellites with no supporting infrastructure/users.

Security and capability, maybe... but resilience? the only way this is improving resilience is if it's merely a fallback option. Satellite communications are a poor option for dependable consumer communications as it's impacted by the weather and easily jammed. It can only improve resilience if it's solely used as a fallback option.

Kinda sounds like the online businesses need to start being financial contributors to ensure they are not relying on flawed software.

Besides, bugs are finite.

Because only morons try to make an argument that's binary all or nothing.

Alright, so first you need machine learning to model a specific area in a stable condition on people and then you can identify them... only in that specific area because all areas need to be trained differently. Furthermore, signal interference can impact results.

It's good work but it's a non-threat.

The Antarctic Treaty (of 1959) designated Antarctica as a scientific preserve and bans military activity, nuclear testing, and territorial claims, while promoting scientific research and international cooperation. As such, to travel to Antarctic you must obtain the permits through a recognized tour operator or your country's government. This means it non-scientific activities in Antarctica aren't really a thing... unless you want to go to jail for being a dumbass.

So toilet paper == Voyager mission entering the heliopause?

I've never gone to a grocery store that sells space missions exploring the solar system. Where do you shop?

You've got some weird fantasies/fetishes, dude.

Also solid advice, which doesn't require bulk public surveilance.

Oh piss off with your "think of the children" nonsense. If you're so worried about your daughter there are countless ways to help keep her safe that don't involve the government warrantlessly tracking the entire population. Teaching her how to properly knee someone in the groin is far more effective anyway.

Spot on. I share the same assumptions as you, that it's not that hard for someone who does such things for a living to figure out how to get the model to recognize "hey, there's water there" and "hey, there's a shitload of water there, probably shouldn't barrel through it".

My guess? I doubt it saw or recognized the intent of the hand gesture, but it almost certainly recognized the flashing red. I assume the "thought" process was "well, nobody else is going. We all stopped at roughly the same time. Yeehaw." but who knows. Doesn't Tesla have some sort of "playback" feature where it can show you what it saw? Or is that only a real-time view?

Fun, spectacularly interesting, and terrifying, all at the same time.

This, right here. Fun story from last week: I had a guy in front of me heading up an onramp. He literally drove OFF the ramp, three wheels in the grass, before he jerked back onto the ramp. Once he got on the highway, did the same thing twice. 80% into another lane before violently jerking back into the lane he was supposed to be in. He was turned completely around in the driver's seat digging for something in the back. I went around the dickhead, gave him a little "howbout you drive?" honk on the way by. The dude literally followed me for a mile with his middle finger hanging out the window. Then promptly went back to digging for whatever was more important than the safety of him and his fellow drivers.

Do autonomous "drivers" fuck up? Of course they do. But they can be "taught" to not do it again. And they certainly aren't turned completely around in their drivers seat, digging for something in the back of their car, while they are driving down the freeway. I'll take the AI driver, thank you very much.

My assumption is that the geo-fencing is a short-term "implement it immediately" kind of thing, while a "teach the Borg to not drive into the water" is a much longer lead time. If the sensor suite is capable of detecting water (which I have no idea what sensors they even have on them, nor their capabilities) I assume it's a relatively easy fix. Just like every other "don't do X" fix that's been implemented after something bad happens. If the sensors can't see water, that's a way different problem.

We don't directly pay licensing for OTA here. It's mainly supported by "OTA Fees" paid by cable/satellite subscribers. (I think, anyway. I know for sure there's a fee for cable/satellite, but no clue on the "mainly" part of my comment.)

Anyway, the prime reason for my OTA setup is for Football season (the American kind). I spent a couple hundred bucks once to set it up, a dual Silicon Dust IP tuner, good antenna, and plex-pass on sale. Now I can watch the games with no additional cost.