Alphabet paid ~$1.1B in taxes in Q1 (quarter of the year). It amounts to roughly ~11%. So while a fairly low rate, it's complete hyperbole to say they don't contribute *anything* to the Federal budget.
That doesn't take into account the payroll taxes, Medicare taxes and income taxes of their employees. Realistically, corporate taxes don't make sense. Tax the investors -- they can't move overseas.
The EU is getting mightily close with China these days. And trade with China would prove -- especially over the next decade or 2 -- way more lucrative.
A wholistic Asian-European-African alliance would be interesting. Russia acting as the military arm to provide security. China as the production and financial engine (supplanting the US). EU as the research, educational and cultural engine (supplanting the US). Africa as the raw materials and labor engine (China's already investing there to turn it into such).
The Americas would be left to their own devices. Which wouldn't be a bad thing if we didn't also piss off the other North American nations we do the most trade with....
It actually does lead to performance issues though. The SD card is orders of magnitude slower. And -- especially Samsung's software -- uses a lot of paging due to its high memory usage.
Not to mention apps loaded from the SD load really slow (noticeably).
The Meltdown demos (from the original researcher) was able to read secure passwords stored in the OS that were supposed to be secured. This wasn't some elaborate setup. This was the guy (the researcher's) personal computer that he just happened to run the code on.
1) Ridiculously difficult to implement. At the end of the day, you are fundamentally tickling the cache and timing the resultant reads to try to determine the content of that cache. Is there ANY reasonable way to "read" the contents of said cache and determine what context a blob of data means?!?
Not that difficult. It just takes a long time as you're relying on certain processor timings. And you also have to target it to the specific microarchitecture you want to attack. So in that respect, it's difficult in that you can't just release one set of code and expect it to work with everything. But if you were to target, say, a device with Snapdragon 845 or an Apple A11...
3) This requires the virus to be running ON your fucking computer!! If you are running ANY virus on your computer, you're hosed.
Every time you visit a website, you're running external Javascript code on your computer that you haven't vetted. Even scarier, if you rent an AWS session, you get a VM that you have all sorts of permissions to run whatever code you want on it.
4) Derived from 3), for the forseeable future ANY virus on your system is about 28Giga-times more likely to be a standard, run-of-the-mill virus. Meantime, everyone is running around wanting to burn their CPUs because they are "vulnerable".
The issue isn't so much your dad with his Windows ME machine being hacked. I mean, we all know that's going to happen. The issue is that previously thought to be super-secure computing domains (AWS, Azure, bank servers, The Pentagon, iOS walled garden) now has a flaw that no one really quite knows how to fully fix yet.
Cross-process isolation can be somewhat simpler. The problem with Spectre specifically is that many programs (your browser opening/. for example) are within the same process running external -- interpreted -- code).
IIRC, most semiconductors are not manufactured in China. The two biggest fab locations are either Taiwan (technically not China by Western standards) or the US...
The problem is that the population does not remain static. And with more mouths to feed (and house and clothe and educate so that they aren't a drag on the economy) year over year, you *need* that productivity increase just to keep things the same.
And don't forget, to have anything to buy used, someone at some point had to have bought it new.
That's also what Spectre (and Meltdown did). They timed cache accesses before and after speculative loads using secure data as the "forwarding address".
The other variants (BranchScope if you're interested) uses a similar technique except it trains the branch predictor using secure data bits and then times the execution time.
What exactly do you think the difference between prefetching and speculative execution is? Most prefetches use program patterns (some even go so far as runahead to guess addresses) to prefetch into cache. It's this exact behavior of populating the cache before permissions are resolved that is both fast (speedup) and insecure.
Half of 200k is more than 3/4 of 100k. But realistically, 200k is just salary. Compensation involves much more than that so you're talking about possibly *saving* 200k/year in equity+cash.
Do that for 5 years and you can go live without worry elsewhere.
If you're main concern is truly to allow non-hollywood-like control of IP. I can understand. Which also means you likely have little issue with China's rather...lax view of IP.
But from a US power and influence standpoint, it would've been a small price to pay (after all Hollywood is primarily an American phenomenon) to contain the only other superpower that could challenge the US.
Let's see, what are the main complaints right now about how China treats foreign businesses?
1. Require joint-venture. Oh right, US companies can now sue the TPP governments who try to pull this BS. 2. State aid towards domestic companies? See above. 3. No labor or environmental standards so that they're waaay more price competitive than a first-world nation? Taken care of.
This "corporations doing their corporationy things is corporate-bad" mentality some people have is confounding. Do you or do you not want American businesses to succeed?
A guy who has not relinquished control of his billion dollar business (for fear of conflict of interest) giving up 400k/year in income is something other than publicity stunt to you?
The TPP is an economic alliance with (originally) the US learning from its trade in China and specifically putting provisions in to try and stop the emerging SE Asian economies from repeating a China. Specifically:
1. They need environmental standards. No more maxing out pollution to be competitive. 2. Labor standards. No more practical slaves to be more competitive. 3. Investor-state dispute. People (especially hard-liberals) see companies-able-to-sue-governments and turn on their "burrrr corporations baddd!" brain. But this is exactly what would've been needed in all the cases where China stole US company IP. Or required joint-ventures. Or subsidized and/or spied for their own domestic companies.
You had 2 opponents of the TPP:
1. People who just think globalization is bad because reasons. 2. People who thought the pacific rim countries were just backwards sh*tholes nobody needed.
Of course, Malaysia and Vietnam are some of the fastest growing tech and manufacturing economies right now. Even China recognizes they can't compete and is moving to shift to services and to bind these countries under trade agreements.
The mouthbreathers (and their elected President) threw away the last chance we had to really contain China.
I mean, I'm ok with that I guess. Despotism isn't my favorite form of government but ruthless as he might be, Xi at least seem interested in stability and prosperity and is actually smart enough to make it happen.
You know, itâ(TM)s not randomly turning. Itâ(TM)s not able to handle a highway exit/split. Which is perfectly expected for, I quote the manual âoeauto lane keepâ.
It works perfectly fine when driving down a marked highway. And from anecdotal experience works better than I could at night and in rain.
And just like adaptive cruise control, itâ(TM)s a convenience feature meant to be used in the right conditions.
But that's binaries that needed a recompile. I'm questioning how hard it really is to just run it through LLVM/GCC/XCode to get an app running on a different architecture.
You'd be surprised at how much modern code doesn't do "calculations" at all. A browser session is literally just function redirect after function redirect.
Both of which fit perfectly in at least the last level cache of modern CPU's. So you have no idea what performance for workloads that don't fit in cache (all of them that matter) performs.
And both have fairly rudimentary hot loops that basically no modern software that users care about (mostly javascript and media creation software) cares about.
Except there isn't even one benchmark that rules them all. There's not even an agreement as to which matters more for what, not to mention how accurate it relates to any one user's actual use case.
If you play Civ 6 all day long, you give 2 shits about whether something scores higher in 3DMark (because Civ 6 happens to be AI limited).
If you do software development, you (again) give 2 shits about 3DMark or PCMark or Geekbench. Because being fast in GCC is a different workload (that perform differently on different architectures) compared to MS Office. Hell, it even varies between different compilers and different languages!
If you do a lot of content creation, you (again) give 2 shits about 3DMark. You *might* give 0.13 shits about Cinebench but even that doesn't correlate to all media tasks. Handbrake, for instance, runs entirely faster on some architectures than Cinebench because it has a large serialized component in pass2.
And all of this changes once the software is updated over time, with compilers targeting different optimization points that benefit/derail different architectures. And we're not talking a paltry 5-10% difference; a recompiled binary can make a 10x speedup difference at times.
Having a single number that never changes through the life of a processor to try to capture all of that? It'd be no more reliable than MAX EXTREME HADES YOUR MOM EDITION
As soon as someone gives me a definitive definition of what "real world performance" for a CPU/GPU is that doesn't change over time/software-version/user-care-ometer is, I might agree that it's feasible to use it to name models.
With modern toolchains, I question how much porting is really required for the vast majority of software (including Adobe) these days. It seems the things that actually need to be "ported" from ecosystem to ecosystem is the UI/UX flow (works differently on a small touchscreen than a laptop with keyboard/mouse).
I doubt there's much architectural-specific code in Photoshop these days...if any at all.
Slashdot Mirror
Looking at last Quarter's statement:
https://abc.xyz/investor/pdf/2...
Alphabet paid ~$1.1B in taxes in Q1 (quarter of the year). It amounts to roughly ~11%. So while a fairly low rate, it's complete hyperbole to say they don't contribute *anything* to the Federal budget.
That doesn't take into account the payroll taxes, Medicare taxes and income taxes of their employees. Realistically, corporate taxes don't make sense. Tax the investors -- they can't move overseas.
The EU is getting mightily close with China these days. And trade with China would prove -- especially over the next decade or 2 -- way more lucrative.
A wholistic Asian-European-African alliance would be interesting. Russia acting as the military arm to provide security. China as the production and financial engine (supplanting the US). EU as the research, educational and cultural engine (supplanting the US). Africa as the raw materials and labor engine (China's already investing there to turn it into such).
The Americas would be left to their own devices. Which wouldn't be a bad thing if we didn't also piss off the other North American nations we do the most trade with....
CA, NY and TX have some kind of influence or power. The other 47 or so are pretty much failed states.
It actually does lead to performance issues though. The SD card is orders of magnitude slower. And -- especially Samsung's software -- uses a lot of paging due to its high memory usage.
Not to mention apps loaded from the SD load really slow (noticeably).
The Meltdown demos (from the original researcher) was able to read secure passwords stored in the OS that were supposed to be secured. This wasn't some elaborate setup. This was the guy (the researcher's) personal computer that he just happened to run the code on.
1) Ridiculously difficult to implement. At the end of the day, you are fundamentally tickling the cache and timing the resultant reads to try to determine the content of that cache. Is there ANY reasonable way to "read" the contents of said cache and determine what context a blob of data means?!?
Not that difficult. It just takes a long time as you're relying on certain processor timings. And you also have to target it to the specific microarchitecture you want to attack. So in that respect, it's difficult in that you can't just release one set of code and expect it to work with everything. But if you were to target, say, a device with Snapdragon 845 or an Apple A11...
3) This requires the virus to be running ON your fucking computer!! If you are running ANY virus on your computer, you're hosed.
Every time you visit a website, you're running external Javascript code on your computer that you haven't vetted. Even scarier, if you rent an AWS session, you get a VM that you have all sorts of permissions to run whatever code you want on it.
4) Derived from 3), for the forseeable future ANY virus on your system is about 28Giga-times more likely to be a standard, run-of-the-mill virus. Meantime, everyone is running around wanting to burn their CPUs because they are "vulnerable".
The issue isn't so much your dad with his Windows ME machine being hacked. I mean, we all know that's going to happen. The issue is that previously thought to be super-secure computing domains (AWS, Azure, bank servers, The Pentagon, iOS walled garden) now has a flaw that no one really quite knows how to fully fix yet.
Cross-process isolation can be somewhat simpler. The problem with Spectre specifically is that many programs (your browser opening /. for example) are within the same process running external -- interpreted -- code).
IIRC, most semiconductors are not manufactured in China. The two biggest fab locations are either Taiwan (technically not China by Western standards) or the US...
The problem is that the population does not remain static. And with more mouths to feed (and house and clothe and educate so that they aren't a drag on the economy) year over year, you *need* that productivity increase just to keep things the same.
And don't forget, to have anything to buy used, someone at some point had to have bought it new.
That's also what Spectre (and Meltdown did). They timed cache accesses before and after speculative loads using secure data as the "forwarding address".
The other variants (BranchScope if you're interested) uses a similar technique except it trains the branch predictor using secure data bits and then times the execution time.
What exactly do you think the difference between prefetching and speculative execution is? Most prefetches use program patterns (some even go so far as runahead to guess addresses) to prefetch into cache. It's this exact behavior of populating the cache before permissions are resolved that is both fast (speedup) and insecure.
You realize this flaw exists in almost every CPU built in the past 2.5 decades right? The newer CPUs are actually less susceptible...
Half of 200k is more than 3/4 of 100k. But realistically, 200k is just salary. Compensation involves much more than that so you're talking about possibly *saving* 200k/year in equity+cash.
Do that for 5 years and you can go live without worry elsewhere.
If you're main concern is truly to allow non-hollywood-like control of IP. I can understand. Which also means you likely have little issue with China's rather...lax view of IP.
But from a US power and influence standpoint, it would've been a small price to pay (after all Hollywood is primarily an American phenomenon) to contain the only other superpower that could challenge the US.
Let's see, what are the main complaints right now about how China treats foreign businesses?
1. Require joint-venture. Oh right, US companies can now sue the TPP governments who try to pull this BS.
2. State aid towards domestic companies? See above.
3. No labor or environmental standards so that they're waaay more price competitive than a first-world nation? Taken care of.
This "corporations doing their corporationy things is corporate-bad" mentality some people have is confounding. Do you or do you not want American businesses to succeed?
A guy who has not relinquished control of his billion dollar business (for fear of conflict of interest) giving up 400k/year in income is something other than publicity stunt to you?
The TPP is an economic alliance with (originally) the US learning from its trade in China and specifically putting provisions in to try and stop the emerging SE Asian economies from repeating a China. Specifically:
1. They need environmental standards. No more maxing out pollution to be competitive.
2. Labor standards. No more practical slaves to be more competitive.
3. Investor-state dispute. People (especially hard-liberals) see companies-able-to-sue-governments and turn on their "burrrr corporations baddd!" brain. But this is exactly what would've been needed in all the cases where China stole US company IP. Or required joint-ventures. Or subsidized and/or spied for their own domestic companies.
You had 2 opponents of the TPP:
1. People who just think globalization is bad because reasons.
2. People who thought the pacific rim countries were just backwards sh*tholes nobody needed.
Of course, Malaysia and Vietnam are some of the fastest growing tech and manufacturing economies right now. Even China recognizes they can't compete and is moving to shift to services and to bind these countries under trade agreements.
The mouthbreathers (and their elected President) threw away the last chance we had to really contain China.
I mean, I'm ok with that I guess. Despotism isn't my favorite form of government but ruthless as he might be, Xi at least seem interested in stability and prosperity and is actually smart enough to make it happen.
You know, itâ(TM)s not randomly turning. Itâ(TM)s not able to handle a highway exit/split. Which is perfectly expected for, I quote the manual âoeauto lane keepâ.
It works perfectly fine when driving down a marked highway. And from anecdotal experience works better than I could at night and in rain.
And just like adaptive cruise control, itâ(TM)s a convenience feature meant to be used in the right conditions.
But that's binaries that needed a recompile. I'm questioning how hard it really is to just run it through LLVM/GCC/XCode to get an app running on a different architecture.
You'd be surprised at how much modern code doesn't do "calculations" at all. A browser session is literally just function redirect after function redirect.
Both of which fit perfectly in at least the last level cache of modern CPU's. So you have no idea what performance for workloads that don't fit in cache (all of them that matter) performs.
And both have fairly rudimentary hot loops that basically no modern software that users care about (mostly javascript and media creation software) cares about.
Except there isn't even one benchmark that rules them all. There's not even an agreement as to which matters more for what, not to mention how accurate it relates to any one user's actual use case.
If you play Civ 6 all day long, you give 2 shits about whether something scores higher in 3DMark (because Civ 6 happens to be AI limited).
If you do software development, you (again) give 2 shits about 3DMark or PCMark or Geekbench. Because being fast in GCC is a different workload (that perform differently on different architectures) compared to MS Office. Hell, it even varies between different compilers and different languages!
If you do a lot of content creation, you (again) give 2 shits about 3DMark. You *might* give 0.13 shits about Cinebench but even that doesn't correlate to all media tasks. Handbrake, for instance, runs entirely faster on some architectures than Cinebench because it has a large serialized component in pass2.
And all of this changes once the software is updated over time, with compilers targeting different optimization points that benefit/derail different architectures. And we're not talking a paltry 5-10% difference; a recompiled binary can make a 10x speedup difference at times.
Having a single number that never changes through the life of a processor to try to capture all of that? It'd be no more reliable than MAX EXTREME HADES YOUR MOM EDITION
As soon as someone gives me a definitive definition of what "real world performance" for a CPU/GPU is that doesn't change over time/software-version/user-care-ometer is, I might agree that it's feasible to use it to name models.
"Macs will be associated with computer illiterate people only."
When, in the history of Apple, has this ever not been true? It's literally their slogan: "it just works".
With modern toolchains, I question how much porting is really required for the vast majority of software (including Adobe) these days. It seems the things that actually need to be "ported" from ecosystem to ecosystem is the UI/UX flow (works differently on a small touchscreen than a laptop with keyboard/mouse).
I doubt there's much architectural-specific code in Photoshop these days...if any at all.