Give me free electricity and compensation for every screw up and I'd gladly live next to a reactor.
Second that. I've been a long time green party voter, and as much as I like seeing solar panels on an ever increasing # of homes, reality is that solar + wind can't cover 100% of our energy needs right now. Period. Not unless / until the storage problem is solved. The sun doesn't shine at night, the wind doesn't always blow (and sometimes too hard!), and no amount of solar panels will fix that. Hydro could be used as backup, but has its own drawbacks & only possible in a few places. Geothermal etc is interesting, but again: far from practical everywhere.
So for filling in the gaps we NEED something else, no way around it. Between 'cheap' coal, oil, natural gas, or covering land masses with biofuel crops, a modern design nuclear plant isn't a bad option. Yes environmentalists may have speeded up investment in solar projects etc (and I applaud anyone for that no matter the reasons), but in resisting (modern) nuclear they've kinda lost sight that thus we're currently on an energy mix where fossil is still king. That could have been very different if modern nuclear plants were common today.
And no, nuclear waste isn't the be-all-end-all-problem it's made out to be. Right now it's choosing between evils, and btw nuclear waste: it's all about what exact substances, how much, stored how & where. The waste from eg. a fast breeder reactor is very different stuff than what comes out of another type of nuclear plant. Stuffing it in rockets & shooting it at the sun, has different risks & costs than burying inside a mountain. Material with 300 year half-life needs a different approach than material with a 30,000 year half-life. And so on.
CO2 in the atmosphere, and the world's CO2 output over a year, isn't the same thing. They're correlated, but with a long delay (in the order of decades or longer IIRC). The atmosphere itself, oceans, forests etc all act like buffers. So if the world (read: mankind's) CO2 output would drop to 0 instantly, CO2 in the atmosphere will stay high for a long time no matter what. Adding more CO2 just makes the problem worse. So a more accurate way is saying that the rate at which we're making the problem worse, has slowed down / flattened. We're still running, and still in the opposite direction of where we should be going, just our [running in the wrong direction] has slowed down.
Once atmospheric CO2 (and with that, average global temperatures) passes certain levels, all kinds of secondary effects may kick in: melting of permafrost areas, melting of oceanic methane ice (yeah I know not CO2 but still caused & contributing to same problem), forest fires due to extended droughts, etc, etc.
In the novel, people on a generation starship discover that salt and other toxins start building up quickly in the smaller scale of their ship.
That's a general rule seen in many contexts: the larger a system is, and the more varied its contents (preferably including many subsystems / groups that work independently from each other), the more stable the whole will be. And/or the bigger the chance at least some of its citizens will survive a catastrophe. Size & numbers matter. Especially if "numbers" can be read as "varieties" rather than a larger count of the same thing(s).
So in terms of passing on genes, a city sized spaceship would be a safer bet than an ISS sized capsule.
But perhaps we'd better take a clue from nature: rather than put all our eggs in a few baskets, it may be a better idea to send out relatively small ships to many destinations. Many won't survive if the journey is long enough or a destination's environment is hostile enough. But overall, there's a bigger chance at least some of those ships will hit fertile ground.
That's not saying as mankind we couldn't do both in a few centuries from now: send some city sized ships to promising destinations, and many small ships elsewhere just in case there's something there.
If you do the test with all the risks of the real mission you might just as well do the real mission.
Even with many of the risks of a real mission, it's still worth doing a dry run to find out what works and what doesn't, while avoiding the COST of a real mission.
... Or perhaps even better: sailors' personal diaries. Mankind has been exploring this planet for millennia. Every sea-faring nation sent out explorer ships at one time or another. Groups of men packed onto a boat with little privacy, not knowing whether they'd ever reach a destination or what they'd find. With the constant danger of disease, water / food / vitamin shortages, going overboard in a storm, fights with fellow sailors, etc etc. And no communication with the outside world for months.
Plenty of historic reports to pick from, plenty examples of how sailors would (or wouldn't?) cope on such journeys. Never mind that in comparison to an ocean-going vessel, a 'habitat' on some remote island is a pretty controlled environment.
Like it or not: the money that does buy laws that the recording industry wants, ultimately comes from us, consumers. So perhaps we shouldn't be buying DVD's etc, but use that money to buy politicians ourselves? Anyone for some crowdfunding actions?;-)
I love the whole "it's only 20 years if you travel at 20% of the speed of light!" part. It makes it sound so close.
What's a human lifetime, anyway? Insignificant.
Let's say we set the bar a few orders of magnitude lower. Say, 0.15% the speed of light. Leave around the time the ancient pyramids of Egypt were built, arrive today.
Now pick something in between. Say, 1% the speed of light. One-way trip ~425 years. Is it so hard to imagine that in a # of decades, we might have probes able to accelerate to that speed? Now replace 'probe' with 'city-sized starship'. Something big enough to allow generations of people to grow up & have offspring. Decades of technological progress not enough? How about a century from now? Or 2 centuries?
In other words: all we need is patience, and imagination. And (as mankind) not be stupid enough to blow ourselves up before those spaceships are on their way. As long as travel group can sit out the ride, who cares if the actual trip time is 20, 200 or 2000 years.
I'm guessing that time to live is more important than having everything looking pretty with your i's dotted and t's crossed.
Absolutely. If TTL is set too low, data packets won't make it back to NSA's servers. But for NSA peeps reading this: do make sure to avoid TTL in the electronics! It's lethal for your spying device battery life.
If 5G is all about short distances, why wouldn't people run their own cells? Kind of like running an open Wi-Fi spot.
For technologies that work over long(er) distances, it's -somewhat- logical that you can't put up your own cell tower. If everybody did that, some would stick to standards and some would not. And soon enough you'd have a free-for-all making the spectrum band(s) useless.
And thus we have (some) government regulation on who gets to use the spectrum & how. Auctioning it off to providers who rig up city- or nationwide networks. But what do you pay a provider for:
a) For maintaining the infrastructure. When everybody puts up their own 'micro-cell tower', no need to pay a 3rd party for maintaining it.
b) For connecting it to upstream (wired) infrastructure. But when those upstream connections have to run all the way to your front door anyway, you can do that yourself right? Again, same as in-home Wi-Fi routing to your internet connection.
c) For user-sharing on those networks, billing, network performance monitoring, etc. Again: when it's all short-distance anyway, no need for that, can be done decentralized by end users. Users that don't play by the rules, can only mess with the spectrum in their immediate area.
Yes you'd still need some standards to enable users to move from micro-cell to micro-cell seamlessly. And use the spectrum in a way that minimizes interference for users that are close to each other. But this is mostly a matter of putting some puzzle-pieces together & declare some de facto standards that every user can follow, right? (in the usual case, baked into consumer devices & their firmware).
'5G' coverage would then simply depend on how interested people in an area are in putting up their routers / antenna's etc. Or am I missing something here?
Following the general consensus, I'm sure the US Department of Defense has come to the same conclusion, and is re-directing their resources as we speak.
To anyone who lives / works near there: can you please look out the window & check if Twitter HQ is being bombed already? Thx for keeping us up to date!
a) Any attempt to access a non-existing page that results in showing a page anyway, is basically fooling the user. Some (ehm.. read: many) users may even think that page still exists even though the original is gone. From a UI perspective that's just wrong even if convenient in many cases.
b) Access to old / archived versions of pages often comes in handy. And that is what the Internet Archive is for. But sometimes pages (or sites) are pulled for a reason. Sometimes good reason(s). Not all information ever placed on the internet needs to be preserved forever, imho.
c) If every 404 leads to a request to the Internet Archive, can they handle the extra load? Even if so, would the extra bandwidth / CPU / disk IO etc be a good use of the IA's limited resources? I very much doubt that, and perhaps Firefox maintainers should answer that question first before activating such a feature by default.
As one of many add-ons: sure why not. As a default feature: bad idea imho.
Okay, we already knew that correct spelling isn't a job requirement for/. editors.
But FAILing to do a simple copy & paste of an article's title? Hell, even some 6y olds can handle that...
Can't/. editors take themselves out of the process, or something? Just write up a couple of scripts to automate the 'editing' and be done with it?
A 100000 year-old piece of code is not "outdated" if the original software is 4 billion years old. In fact, it is actually brand new.
Yeah... Debian experimental-style new. But a lot depends on your definition of "original"; quite a few patches have been made over time so perhaps little (if any) of the original code is left.
As for 100,000 year-old code: pfff... I keep a some local repositories around that see regular updates. Daily *and* nightly builds, a full suite of regression tests on real hardware, automated backups... the works.
On a side note: the hell with versioning or changelogs. Just kick out a stable release every couple of years, stick a funny name on it, and call it a day.
Who is going to buy all this stuff if they don't have jobs?
And there lies the heart of the problem: purchasing power is coupled to having a job.
As technology marches forward, that coupling has to be let go. Or at least loosened. The majority of the population needs to have some purchasing power even if there's no job for them. Think basic income.
The alternative: (almost) everything automated, production equipment (including robots) in the hands of a few corporations & the billionaires at their top, with the rest of the population jobless / out of money (and in the extreme case: out of housing or food). Great recipe for say, a nice little civil war. As it has been several times in history.
The automation itself isn't a bad thing, it increases productivity so we can have more nice things or do fun stuff more of the time. But the fruits of that increased productivity should be divided somewhat evenly over people. If it ends up in the hands of a few you have a recipe for disaster.
It is even illegal to sell unpasteurized milk in most of EU.
Depends on how & where. For example: I'm pretty sure most dairy farmers in my area will be happy to have a meet & greet with one of their end users, tap a few litres into a bottle, and take ~3x the wholesale price they are getting from factory.
That's unpasteurized milk, full fat, straight from cow -> cooling tank -> end user's fridge (leave it there overnight to skim off the fat). As has been done for ages regardless what EU rules say about it. Thankfully EU bureaucrats haven't rotted everyone's brain.. yet...
As a user experience feature, yes. As a security feature, no.
1st: click-to-run moves the decision whether something is safe to run from system to the user. Which time and time again, in many different contexts, has been proven to... NOT WORK. "Do you want to run this random piece of crap from unknown / untrusted source?". "Yes, please!". In the vast majority of cases, users don't have enough info / can't be trusted / aren't knowledgeable enough to make that decision.
2nd: either you can guarantee that all content thrown at a plugin is safe to run, or you can't. If you can, then click-to-run is not needed (at least not for security reasons). If you can't, then click-to-run just opens up a security hole that shouldn't exist in the first place.
So if it's done to let users decide what obnoxious ads they want to see, or what web game to bog down their machine: fine.
As a way to enhance security, that's just security by obscurity. Not saying it doesn't help... but the choice presented to the user should be "do you want to play this?". NOT "do you think this content is safe?"
I'd be inclined to agree with you - for thinly populated areas.
You do realize that (for a wide frequency band) the EM spectrum is a shared resource? Like the air we breathe: if I start a fire, that removes oxygen from the air around it. And puts smoke in the air. Smoke that will be visible from a distance, and the combustion products may affect people in a wide area. Therefore (in most populated areas) people are not free to burn stuff out in the open as they please. Such activities may be regulated, and rightfully so.
Above certain EM frequencies (say, IR and up), the physical properties of signals make it pointless to try and regulate things. Below certain frequencies, lack of practical applications make regulation not-needed / pointless. But in between, we're talking about a shared (and limited!) resource. So some government regulation is quite appropriate.
Care to explain that? Open source can (and usually is) copyrighted. It has nothing to do with competition.
It does: you may regard the code itself as documentation. Describing a process, some method of calculation, a file format processed, etc. Which in turns makes it easy to write a competing implementation that does the same job.
For closed source software that is much more difficult. It doesn't even matter whether the code is open in the "libre" sense: as long as you can inspect the code, you can figure out what it does. Same with copyrights: that serves to give author(s) some control over copy & paste style use of the code. But it doesn't prevent others from writing a competing implementation.
Having code that's actually "libre" open source is still nice though for other reasons.
If the data has been accessed by unauthorised persons, there is no way to be 100% certain that it hasn't made it off premise (..)
There is: if the system(s) in question are air-gapped, or on a LAN that has no external network connections. Malware (ransomware included) could still make its way onto such systems. Let's say through an infected USB stick.
For real-world scenarios that's mostly a hypothetical case I suspect. While in theory that USB stick could compromise an air-gapped system, retrieve sensitive data, and then upload that data when it (later) gets plugged into another machine that does have internet access, that's more along the line of a highly targeted phishing attack.
In such a breach the amount of leaked data would be limited by the capacity of said USB stick (and perhaps also its write speed). For an air-gapped system or isolated LAN, if a breach occurs obviously it's worth investigating how that happened. Let's say that's done, a specific 'bad' USB stick is found and its whereabouts are known, then yes it may be possible to say with confidence: breach / infection occurred, but no data leaked out.
Disgruntled employee that carries infected USB stick & demands some Bitcoins? Sure, possible, but how likely? How often have you heard of such a case? So mostly hypothetical. 'Random' infection through internet, controlled by persons unknown in a far away country, is a much more common scenario I think. In such a case, "compromised" will imply "data may have leaked". An investigation of the capabilities of the specific malware could yield more clues, but 100% certaintly? Indeed not.
Slashdot Mirror
Give me free electricity and compensation for every screw up and I'd gladly live next to a reactor.
Second that. I've been a long time green party voter, and as much as I like seeing solar panels on an ever increasing # of homes, reality is that solar + wind can't cover 100% of our energy needs right now. Period. Not unless / until the storage problem is solved. The sun doesn't shine at night, the wind doesn't always blow (and sometimes too hard!), and no amount of solar panels will fix that. Hydro could be used as backup, but has its own drawbacks & only possible in a few places. Geothermal etc is interesting, but again: far from practical everywhere.
So for filling in the gaps we NEED something else, no way around it. Between 'cheap' coal, oil, natural gas, or covering land masses with biofuel crops, a modern design nuclear plant isn't a bad option. Yes environmentalists may have speeded up investment in solar projects etc (and I applaud anyone for that no matter the reasons), but in resisting (modern) nuclear they've kinda lost sight that thus we're currently on an energy mix where fossil is still king. That could have been very different if modern nuclear plants were common today.
And no, nuclear waste isn't the be-all-end-all-problem it's made out to be. Right now it's choosing between evils, and btw nuclear waste: it's all about what exact substances, how much, stored how & where. The waste from eg. a fast breeder reactor is very different stuff than what comes out of another type of nuclear plant. Stuffing it in rockets & shooting it at the sun, has different risks & costs than burying inside a mountain. Material with 300 year half-life needs a different approach than material with a 30,000 year half-life. And so on.
Radation is deadly.
On spelling tests, it is indeed. But members of the Therefor people (whoever they are) would probably forgive you...
Is anyone surprised that if stuff was cheaper, people are less likely to take it for nothing?
I'm confused now. Let's say legitimate providers make content available for free. Then should we take that, or keep pirating the stuff anyway?
CO2 in the atmosphere, and the world's CO2 output over a year, isn't the same thing. They're correlated, but with a long delay (in the order of decades or longer IIRC). The atmosphere itself, oceans, forests etc all act like buffers. So if the world (read: mankind's) CO2 output would drop to 0 instantly, CO2 in the atmosphere will stay high for a long time no matter what. Adding more CO2 just makes the problem worse. So a more accurate way is saying that the rate at which we're making the problem worse, has slowed down / flattened. We're still running, and still in the opposite direction of where we should be going, just our [running in the wrong direction] has slowed down.
Once atmospheric CO2 (and with that, average global temperatures) passes certain levels, all kinds of secondary effects may kick in: melting of permafrost areas, melting of oceanic methane ice (yeah I know not CO2 but still caused & contributing to same problem), forest fires due to extended droughts, etc, etc.
In the novel, people on a generation starship discover that salt and other toxins start building up quickly in the smaller scale of their ship.
That's a general rule seen in many contexts: the larger a system is, and the more varied its contents (preferably including many subsystems / groups that work independently from each other), the more stable the whole will be. And/or the bigger the chance at least some of its citizens will survive a catastrophe. Size & numbers matter. Especially if "numbers" can be read as "varieties" rather than a larger count of the same thing(s).
So in terms of passing on genes, a city sized spaceship would be a safer bet than an ISS sized capsule.
But perhaps we'd better take a clue from nature: rather than put all our eggs in a few baskets, it may be a better idea to send out relatively small ships to many destinations. Many won't survive if the journey is long enough or a destination's environment is hostile enough. But overall, there's a bigger chance at least some of those ships will hit fertile ground.
That's not saying as mankind we couldn't do both in a few centuries from now: send some city sized ships to promising destinations, and many small ships elsewhere just in case there's something there.
If you do the test with all the risks of the real mission you might just as well do the real mission.
Even with many of the risks of a real mission, it's still worth doing a dry run to find out what works and what doesn't, while avoiding the COST of a real mission.
Plenty of historic reports to pick from, plenty examples of how sailors would (or wouldn't?) cope on such journeys. Never mind that in comparison to an ocean-going vessel, a 'habitat' on some remote island is a pretty controlled environment.
Like it or not: the money that does buy laws that the recording industry wants, ultimately comes from us, consumers. So perhaps we shouldn't be buying DVD's etc, but use that money to buy politicians ourselves? Anyone for some crowdfunding actions? ;-)
I love the whole "it's only 20 years if you travel at 20% of the speed of light!" part. It makes it sound so close.
What's a human lifetime, anyway? Insignificant.
Let's say we set the bar a few orders of magnitude lower. Say, 0.15% the speed of light. Leave around the time the ancient pyramids of Egypt were built, arrive today.
Now pick something in between. Say, 1% the speed of light. One-way trip ~425 years. Is it so hard to imagine that in a # of decades, we might have probes able to accelerate to that speed? Now replace 'probe' with 'city-sized starship'. Something big enough to allow generations of people to grow up & have offspring. Decades of technological progress not enough? How about a century from now? Or 2 centuries?
In other words: all we need is patience, and imagination. And (as mankind) not be stupid enough to blow ourselves up before those spaceships are on their way. As long as travel group can sit out the ride, who cares if the actual trip time is 20, 200 or 2000 years.
I'm guessing that time to live is more important than having everything looking pretty with your i's dotted and t's crossed.
Absolutely. If TTL is set too low, data packets won't make it back to NSA's servers. But for NSA peeps reading this: do make sure to avoid TTL in the electronics! It's lethal for your spying device battery life.
He was bigger than most of us will ever be...
If 5G is all about short distances, why wouldn't people run their own cells? Kind of like running an open Wi-Fi spot.
For technologies that work over long(er) distances, it's -somewhat- logical that you can't put up your own cell tower. If everybody did that, some would stick to standards and some would not. And soon enough you'd have a free-for-all making the spectrum band(s) useless.
And thus we have (some) government regulation on who gets to use the spectrum & how. Auctioning it off to providers who rig up city- or nationwide networks. But what do you pay a provider for:
a) For maintaining the infrastructure. When everybody puts up their own 'micro-cell tower', no need to pay a 3rd party for maintaining it.
b) For connecting it to upstream (wired) infrastructure. But when those upstream connections have to run all the way to your front door anyway, you can do that yourself right? Again, same as in-home Wi-Fi routing to your internet connection.
c) For user-sharing on those networks, billing, network performance monitoring, etc. Again: when it's all short-distance anyway, no need for that, can be done decentralized by end users. Users that don't play by the rules, can only mess with the spectrum in their immediate area.
Yes you'd still need some standards to enable users to move from micro-cell to micro-cell seamlessly. And use the spectrum in a way that minimizes interference for users that are close to each other. But this is mostly a matter of putting some puzzle-pieces together & declare some de facto standards that every user can follow, right? (in the usual case, baked into consumer devices & their firmware).
'5G' coverage would then simply depend on how interested people in an area are in putting up their routers / antenna's etc. Or am I missing something here?
Following the general consensus, I'm sure the US Department of Defense has come to the same conclusion, and is re-directing their resources as we speak.
To anyone who lives / works near there: can you please look out the window & check if Twitter HQ is being bombed already? Thx for keeping us up to date!
No worries! Only some oxygen dihydride was detected in your drinking water, in generally safe levels.
A better reference:
The Matrix - virus scene
For several reasons:
a) Any attempt to access a non-existing page that results in showing a page anyway, is basically fooling the user. Some (ehm.. read: many) users may even think that page still exists even though the original is gone. From a UI perspective that's just wrong even if convenient in many cases.
b) Access to old / archived versions of pages often comes in handy. And that is what the Internet Archive is for. But sometimes pages (or sites) are pulled for a reason. Sometimes good reason(s). Not all information ever placed on the internet needs to be preserved forever, imho.
c) If every 404 leads to a request to the Internet Archive, can they handle the extra load? Even if so, would the extra bandwidth / CPU / disk IO etc be a good use of the IA's limited resources? I very much doubt that, and perhaps Firefox maintainers should answer that question first before activating such a feature by default.
As one of many add-ons: sure why not. As a default feature: bad idea imho.
(..) the DNC is one big corporate bride.
Okay, we already knew that correct spelling isn't a job requirement for /. editors.
But FAILing to do a simple copy & paste of an article's title? Hell, even some 6y olds can handle that...
Can't /. editors take themselves out of the process, or something? Just write up a couple of scripts to automate the 'editing' and be done with it?
A 100000 year-old piece of code is not "outdated" if the original software is 4 billion years old. In fact, it is actually brand new.
Yeah... Debian experimental-style new. But a lot depends on your definition of "original"; quite a few patches have been made over time so perhaps little (if any) of the original code is left.
As for 100,000 year-old code: pfff... I keep a some local repositories around that see regular updates. Daily *and* nightly builds, a full suite of regression tests on real hardware, automated backups... the works.
On a side note: the hell with versioning or changelogs. Just kick out a stable release every couple of years, stick a funny name on it, and call it a day.
Who is going to buy all this stuff if they don't have jobs?
And there lies the heart of the problem: purchasing power is coupled to having a job.
As technology marches forward, that coupling has to be let go. Or at least loosened. The majority of the population needs to have some purchasing power even if there's no job for them. Think basic income.
The alternative: (almost) everything automated, production equipment (including robots) in the hands of a few corporations & the billionaires at their top, with the rest of the population jobless / out of money (and in the extreme case: out of housing or food). Great recipe for say, a nice little civil war. As it has been several times in history.
The automation itself isn't a bad thing, it increases productivity so we can have more nice things or do fun stuff more of the time. But the fruits of that increased productivity should be divided somewhat evenly over people. If it ends up in the hands of a few you have a recipe for disaster.
It is even illegal to sell unpasteurized milk in most of EU.
Depends on how & where. For example: I'm pretty sure most dairy farmers in my area will be happy to have a meet & greet with one of their end users, tap a few litres into a bottle, and take ~3x the wholesale price they are getting from factory.
That's unpasteurized milk, full fat, straight from cow -> cooling tank -> end user's fridge (leave it there overnight to skim off the fat). As has been done for ages regardless what EU rules say about it. Thankfully EU bureaucrats haven't rotted everyone's brain.. yet...
As a user experience feature, yes. As a security feature, no.
So if it's done to let users decide what obnoxious ads they want to see, or what web game to bog down their machine: fine.
As a way to enhance security, that's just security by obscurity. Not saying it doesn't help... but the choice presented to the user should be "do you want to play this?". NOT "do you think this content is safe?"
Backups are for pussies. Real men just upload their shit to some FTP server... ehm, never mind.
I'd be inclined to agree with you - for thinly populated areas.
You do realize that (for a wide frequency band) the EM spectrum is a shared resource? Like the air we breathe: if I start a fire, that removes oxygen from the air around it. And puts smoke in the air. Smoke that will be visible from a distance, and the combustion products may affect people in a wide area. Therefore (in most populated areas) people are not free to burn stuff out in the open as they please. Such activities may be regulated, and rightfully so.
Above certain EM frequencies (say, IR and up), the physical properties of signals make it pointless to try and regulate things. Below certain frequencies, lack of practical applications make regulation not-needed / pointless. But in between, we're talking about a shared (and limited!) resource. So some government regulation is quite appropriate.
Care to explain that? Open source can (and usually is) copyrighted. It has nothing to do with competition.
It does: you may regard the code itself as documentation. Describing a process, some method of calculation, a file format processed, etc. Which in turns makes it easy to write a competing implementation that does the same job.
For closed source software that is much more difficult. It doesn't even matter whether the code is open in the "libre" sense: as long as you can inspect the code, you can figure out what it does. Same with copyrights: that serves to give author(s) some control over copy & paste style use of the code. But it doesn't prevent others from writing a competing implementation.
Having code that's actually "libre" open source is still nice though for other reasons.
If the data has been accessed by unauthorised persons, there is no way to be 100% certain that it hasn't made it off premise (..)
There is: if the system(s) in question are air-gapped, or on a LAN that has no external network connections. Malware (ransomware included) could still make its way onto such systems. Let's say through an infected USB stick.
For real-world scenarios that's mostly a hypothetical case I suspect. While in theory that USB stick could compromise an air-gapped system, retrieve sensitive data, and then upload that data when it (later) gets plugged into another machine that does have internet access, that's more along the line of a highly targeted phishing attack.
In such a breach the amount of leaked data would be limited by the capacity of said USB stick (and perhaps also its write speed). For an air-gapped system or isolated LAN, if a breach occurs obviously it's worth investigating how that happened. Let's say that's done, a specific 'bad' USB stick is found and its whereabouts are known, then yes it may be possible to say with confidence: breach / infection occurred, but no data leaked out.
Disgruntled employee that carries infected USB stick & demands some Bitcoins? Sure, possible, but how likely? How often have you heard of such a case? So mostly hypothetical. 'Random' infection through internet, controlled by persons unknown in a far away country, is a much more common scenario I think. In such a case, "compromised" will imply "data may have leaked". An investigation of the capabilities of the specific malware could yield more clues, but 100% certaintly? Indeed not.