Smart devices do a whole lot of updating, etc, and it's over encrypted channels now. If you can spool, you can bury the traffic with the legitimate and it'll be very hard to detect-- only by reversing the device itself can you really know.
> But the smart speaker things are supposed to be sending all their recordings to a server somewhere.
Yes, after the wake-phrase. And your phone is just supposed to send it during a call or when a permitted app is doing it. And your computer only under certain circumstances, too. E.g. in each case there are policies for when they're allowed to use the microphone and to send its data over the network that protect privacy if they are followed, but it's extraordinarily difficult to know for sure if they really are.
> Yes, but you'd have to actually enable "OK Google". Who does that besides my parents because they think it's "nifty"?
So you know the device is capable of doing always-on listening and keyword detection, without a big battery penalty. And you know there's a software switch where you can turn off the phone from responding to the "OK Google" phrase, and *this* is the bit that convinces you that you're not being monitored-- that you have that switch turned off?
So it's funny-- otherwise capable engineers who go and solve weird problems every day, then in turn look at their cellphone and say "no way could this be surveiling me".
But on the other hand, I have a feeling if you worked on the device, and knew-- OK, here's standby power, here's how often the phone would need to wake up to drain DMA from audio, here's how much it costs to compress and store audio, here's the longevity of flash, here's the amount of flash memory the user won't miss, and here's how short of a burst of wifi traffic it would be to send the audio back--- you could solve the problem and do it.
I don't think (most) phones are spooling up audio and chirping it back to the mothership, but it wouldn't be that hard to do and usually evade detection.
If you're too short, you might not be able to drive a random car. You might not be able to reach the microwave in many houses, etc. It's a completely disjoint different kind of suck.
Apple's P/E is only 19. That compares to e.g. Lenovo sitting at 19.3. Samsung's not great at 12 but has tons of capital tied up in things that are unexciting (fabs, shipbuilding, etc).
The fourth amendment mostly doesn't apply, sadly. After all, the fourth is about searches, and there was already a strong tradition in the common law of receiving ordinary business records, (including things one holds for other parties) by the weaker process of subpoena instead of search order. 19th century case law established firmly that if you put the information in the care of another (non-privileged) party that it must not be that sensitive anyways.
This is one of the assumptions with the constitution and its interpretation that no longer hold up so well in a digital age.
Another is the amount of information we allow police officers to gather in "plain sight." All kinds of privacy concerns aren't terribly significant if human constables have to do it; you can't have that many of them and presumably they will concentrate their efforts sanely. But this doesn't hold so well for things like license plate cameras, which can gather data in aggregate cheaply that is hugely invasive.
The real question is whether they can produce enough units at a price that people want to stop just chewing through capital at an insane rate.
It's relatively easy to dump money into growth and get lots of people who want your (subsidized) product. It's another thing to have the degree of operations acumen to keep costs low, and to have people like your product at the "natural price point" it actually shows up at.
Tesla makes a great product people want-- there's no doubt. Questions are just about the sustainability of the demand, the competitive pressures they're bound to face, the operational efficiency they'll need to reach, and the ultimate economics of their space.
What happened was, Google made a version of AlphaGo that beat Lee Se-dol. Call this AlphaGo One. It won, but it was at least close.
Then Google updated it and had it play lots and lots of top players, and it trounced them all. Call this AlphaGo Two.
Then they did this new version, AlphaGo Zero. Zero, early in training, beat AlphaGo One 40-0. Late in training it defeated AlphaGo Two 90% of the time.
Sure, it's bad at protecting minority interests-- and this is just another instance of it. Anyways, it seems you've completely failed to defend your original point.
Unions are great at screwing the best employees, the newest employees, and anyone notably different from the "typical" worker protected by the union whenever their interests diverge. It can be considerably better than a single powerful employer, but when there's many employers and the union becomes more powerful than any of them the union becomes the tyranny to be scared of.
> Well, it's you and your dad's choice to be good little Calvinists for corporate benefit, but the "unions promote mediocrity" line is and always has been bullshit. Nothing about unions prevents good workers from making more money or bad workers from being fired for cause.
Hey, there's plenty of good things about unions, but... contract negotiations with a union are always about what will keep the majority of the people in the union happy. Provisions to allow the best 10% to be paid a bunch more (presumably at the expense of the other 90%) are not likely to be it.
Likewise, provisions that improve job stability that are somewhat at the expense of the average quality of the workforce are likely to be popular.
I think the biggest concern with all these stories about being late and having trouble ramping production is that they consume lots of capital and put profitability further out of reach.
Equity markets love TSLA right now but will not give them endless sips of new capital.
Yah, it's clearly not OK but I can't help but have some sympathy for the guy. He was targeted with illegitimate reviews; fought one round in the courts and prevailed; and then maybe couldn't afford/stomach another battle.
False flag is one of 3 big possibilities I can think of. Another is something that was intended to be a more subtle operation of some kind that was badly fucked up. Last, there could be rogue elements in Cuba's intelligence apparatus that would rather relations not improve.
I mean, if you're giving/selling someone a product, on some level they have all the blueprints to make it.
But bitstream security in FPGAs / program security in microcontrollers makes it significantly less likely someone makes a gate-for-gate copy of what you're doing a month after launch. Just by making cloning a little more difficult/inconvenient your market can be protected for a reasonable amount of time.
Sweden did OK... Denmark and Norway were both invaded and occupied. Norway in particular had the vast majority of their economic output raided. and then had entire cities torched to the ground by retreating German forces at the end of the war.
It means 32 layers of cells in the chip, which themselves are probably TLC. 3D-NAND/VNAND are not new, and there's products with up to 64 layers in the market.
Man, why did NASA keep assisting in the development of all those fighter plane prototypes and research programs like X-53, X-31, etc then?:P A huge part of what the Dryden (now Armstrong) Flight Research Center does is military R&D, not to mention significant work at other centers.
Let's say "we're at the top," and one day in the future we have enough resources for there to simulate a universe for whatever reason, and do so, and it has enough fidelity to simulate another one within, eventually. Of the three universes' dwellers asking themselves "are we in a simiulation?" the answer is "yes" for 2 of them and "no" for 1. This ignores the possibility for branching and deeper nesting.
Now take it one step further, and say "we don't know if we're at the top." We already know that most universes are simulated... how would you bet?
Of course, there's probably a very outermost, "real" reality. But most universes with beings dwelling in them are simulations, so...
I feel like you don't understand what CI is, or are taking potshots at a particularly extreme view of it (e.g. everyone checks in every day).
Having a degree of automated build and lightweight test on every check-in can save you from getting stuck when there's breakage-on-breakage-on-breakage from putting multiple developers work together. In turn, by not getting stuck in integration hell at the end of the cycle, you can have more time for actually-robust-test and cleanup.
Slashdot Mirror
Smart devices do a whole lot of updating, etc, and it's over encrypted channels now. If you can spool, you can bury the traffic with the legitimate and it'll be very hard to detect-- only by reversing the device itself can you really know.
> But the smart speaker things are supposed to be sending all their recordings to a server somewhere.
Yes, after the wake-phrase. And your phone is just supposed to send it during a call or when a permitted app is doing it. And your computer only under certain circumstances, too. E.g. in each case there are policies for when they're allowed to use the microphone and to send its data over the network that protect privacy if they are followed, but it's extraordinarily difficult to know for sure if they really are.
> Yes, but you'd have to actually enable "OK Google". Who does that besides my parents because they think it's "nifty"?
So you know the device is capable of doing always-on listening and keyword detection, without a big battery penalty. And you know there's a software switch where you can turn off the phone from responding to the "OK Google" phrase, and *this* is the bit that convinces you that you're not being monitored-- that you have that switch turned off?
Interesting bit of logic there.
So it's funny-- otherwise capable engineers who go and solve weird problems every day, then in turn look at their cellphone and say "no way could this be surveiling me".
But on the other hand, I have a feeling if you worked on the device, and knew-- OK, here's standby power, here's how often the phone would need to wake up to drain DMA from audio, here's how much it costs to compress and store audio, here's the longevity of flash, here's the amount of flash memory the user won't miss, and here's how short of a burst of wifi traffic it would be to send the audio back--- you could solve the problem and do it.
I don't think (most) phones are spooling up audio and chirping it back to the mothership, but it wouldn't be that hard to do and usually evade detection.
If you're too short, you might not be able to drive a random car. You might not be able to reach the microwave in many houses, etc. It's a completely disjoint different kind of suck.
Hauling around 3 gallons of water all day is no big deal. It's doing it with your arms, and repeatedly, that sucks.
Apple's P/E is only 19. That compares to e.g. Lenovo sitting at 19.3. Samsung's not great at 12 but has tons of capital tied up in things that are unexciting (fabs, shipbuilding, etc).
The fourth amendment mostly doesn't apply, sadly. After all, the fourth is about searches, and there was already a strong tradition in the common law of receiving ordinary business records, (including things one holds for other parties) by the weaker process of subpoena instead of search order. 19th century case law established firmly that if you put the information in the care of another (non-privileged) party that it must not be that sensitive anyways.
This is one of the assumptions with the constitution and its interpretation that no longer hold up so well in a digital age.
Another is the amount of information we allow police officers to gather in "plain sight." All kinds of privacy concerns aren't terribly significant if human constables have to do it; you can't have that many of them and presumably they will concentrate their efforts sanely. But this doesn't hold so well for things like license plate cameras, which can gather data in aggregate cheaply that is hugely invasive.
Man, I didn't know that the United Auto Workers were so good at rocketry. (I suspect you mean USA)
The real question is whether they can produce enough units at a price that people want to stop just chewing through capital at an insane rate.
It's relatively easy to dump money into growth and get lots of people who want your (subsidized) product. It's another thing to have the degree of operations acumen to keep costs low, and to have people like your product at the "natural price point" it actually shows up at.
Tesla makes a great product people want-- there's no doubt. Questions are just about the sustainability of the demand, the competitive pressures they're bound to face, the operational efficiency they'll need to reach, and the ultimate economics of their space.
What happened was, Google made a version of AlphaGo that beat Lee Se-dol. Call this AlphaGo One. It won, but it was at least close.
Then Google updated it and had it play lots and lots of top players, and it trounced them all. Call this AlphaGo Two.
Then they did this new version, AlphaGo Zero. Zero, early in training, beat AlphaGo One 40-0. Late in training it defeated AlphaGo Two 90% of the time.
Sure, it's bad at protecting minority interests-- and this is just another instance of it. Anyways, it seems you've completely failed to defend your original point.
Unions are great at screwing the best employees, the newest employees, and anyone notably different from the "typical" worker protected by the union whenever their interests diverge. It can be considerably better than a single powerful employer, but when there's many employers and the union becomes more powerful than any of them the union becomes the tyranny to be scared of.
> Well, it's you and your dad's choice to be good little Calvinists for corporate benefit, but the "unions promote mediocrity" line is and always has been bullshit. Nothing about unions prevents good workers from making more money or bad workers from being fired for cause.
Hey, there's plenty of good things about unions, but... contract negotiations with a union are always about what will keep the majority of the people in the union happy. Provisions to allow the best 10% to be paid a bunch more (presumably at the expense of the other 90%) are not likely to be it.
Likewise, provisions that improve job stability that are somewhat at the expense of the average quality of the workforce are likely to be popular.
I think the biggest concern with all these stories about being late and having trouble ramping production is that they consume lots of capital and put profitability further out of reach.
Equity markets love TSLA right now but will not give them endless sips of new capital.
This is the one time that vertical video is OK, but even so-- most of the time you would still be better off with a horizontal framing of the subject.
Yes, losing the bottom sucks. But compared to giant black bars when anyone not on a phone views it, it's the lesser sin.
Yah, it's clearly not OK but I can't help but have some sympathy for the guy. He was targeted with illegitimate reviews; fought one round in the courts and prevailed; and then maybe couldn't afford/stomach another battle.
e.g. http://www.nordnorge.com/en/wa...
At some point you've been corrected enough and should educate yourself, rather than spouting crap and expecting others to teach you.
It's against Cuba's apparent interest, yes.
False flag is one of 3 big possibilities I can think of. Another is something that was intended to be a more subtle operation of some kind that was badly fucked up. Last, there could be rogue elements in Cuba's intelligence apparatus that would rather relations not improve.
Yes.
I mean, if you're giving/selling someone a product, on some level they have all the blueprints to make it.
But bitstream security in FPGAs / program security in microcontrollers makes it significantly less likely someone makes a gate-for-gate copy of what you're doing a month after launch. Just by making cloning a little more difficult/inconvenient your market can be protected for a reasonable amount of time.
> But not Sweden, Denmark, and Norway
Sweden did OK... Denmark and Norway were both invaded and occupied. Norway in particular had the vast majority of their economic output raided. and then had entire cities torched to the ground by retreating German forces at the end of the war.
It means 32 layers of cells in the chip, which themselves are probably TLC. 3D-NAND/VNAND are not new, and there's products with up to 64 layers in the market.
When we talk about strong AI, there's plenty of scary scenarios in the (not-near, but unclear how distant) future.
https://wiki.lesswrong.com/wik...
Man, why did NASA keep assisting in the development of all those fighter plane prototypes and research programs like X-53, X-31, etc then? :P A huge part of what the Dryden (now Armstrong) Flight Research Center does is military R&D, not to mention significant work at other centers.
That's actually the whole concept though.
Let's say "we're at the top," and one day in the future we have enough resources for there to simulate a universe for whatever reason, and do so, and it has enough fidelity to simulate another one within, eventually. Of the three universes' dwellers asking themselves "are we in a simiulation?" the answer is "yes" for 2 of them and "no" for 1. This ignores the possibility for branching and deeper nesting.
Now take it one step further, and say "we don't know if we're at the top." We already know that most universes are simulated... how would you bet?
Of course, there's probably a very outermost, "real" reality. But most universes with beings dwelling in them are simulations, so...
We had automated check-in before face rec was deployed in it...
I feel like you don't understand what CI is, or are taking potshots at a particularly extreme view of it (e.g. everyone checks in every day).
Having a degree of automated build and lightweight test on every check-in can save you from getting stuck when there's breakage-on-breakage-on-breakage from putting multiple developers work together. In turn, by not getting stuck in integration hell at the end of the cycle, you can have more time for actually-robust-test and cleanup.