"x86" is kind of a generic name of a "line of" processors, similar to the System/360 "architecture" even though current machines go beyond the original System/360.
English is almost inherently lossy anyhow such that it's probably not worth debating. One can use "the x86 line" to hopefully indicate a more generic implication.
Sometimes choice is better than progress, or at least feels that way to humans.
Does universal logic dictate that progress is more important than choice? I'd like to see something close to a formal proof if you claim so. You know, pedantic debate fun.
But as long as we need flexibility in our processing environment, we will need general processing CPU architectures
Couldn't a general CPU be a fall-back option if specialized processors are not available at a given time or environment?
I'm no "chip expert", but I can envision at least 2 ways this could be done: one is by emulation. If the specialized CPU is not available, then the general CPU processes the task in emulation mode.
Another approach is to use the same machine language on the specialized chips, but optimize specific kinds of processing and/or instructions. For example, if AI uses a lot of the A and B instructions, but D and C are rarer in AI, then the "AI chip" would run instructions A and B really fast, but slack on D and C to save resources for the first two. If by chance the AI chip is not available, the general CPU can still process the AI-centric task, just not as fast.
bigger question is really if x86 should be the basis for a new processor initiative
China may be hedging its bets by investing in both kinds of CPUs: x86 and ARM.
Some argue x86 is dying, but x86 has more server-centric features than ARM, and is thus is still popular for server farms. (At least the x86 server features are more mature.)
I wonder if GPU's will overtake both of these, or at least push x86 and ARM into being mostly coordinators. Perhaps the market will shift to specialized chip-sets for AI, databases, graphics, etc., and x86 or ARM will mostly function as process coordinators which dish out specific tasks to specialized CPUs.
Good! AMD and Intel need competition in x86 chips. Duopolies usually provide narrow choices, and consumers can't do squat about it (big telecoms cough cough).
It was aimed using rays, not propelled. A reaction wheel and its spare failed, so they used Mr. Sunshine instead. Not as good as the original, but they were still able to do science.
There used to be a "thing" called efficiency experts. They did something called, generically, time and motion studies.
My father used to be one under the title "industrial engineer". He started with factories, but as they offshored, worked with hospitals. Some people resented being told how to do their job, so there were office politics fireworks at times.
He'd do it at home also, arguing that we wasted toilet paper; and we had strange arguments over the physics of wiping and average consistency of poop ("bar" charts, eeww). "Everyone is different, how do you know my consistency profile is the same as yours? You are guessing, admit it!"
I kind of have the same "gene" and argue that our development stack wastes too many keystrokes, eye movements, and has unnecessary code pattern redundancies as alleged insurance against unlikely future events. The primary architect doesn't like me. He uses bullshit buzzwords to try to counter me. The fight ain't over...
Anyhow, I'm not sure how it relates to most of the article, though. My dad's studies were not done in secret. Most of the article was not about training robots to be better or more efficient, but about faking AI by acting like or in place of a bot to trick investors and/or buy time when bots are not ready.
Well-known languages seem to collect cruft as they age. Look up and try to do reflection on C#'s nullable-types for example. Nullable types were an after-thought such that they ended up bastardizing dynamic types to implement them, making for ugly reflection. Granted, perhaps there were other ways to implement nullable types, but I'm sure those would uglify something else.
automation and the associated productivity will make life much better, just like it always has.
That's not true. The Luddites didn't riot because they hated new gizmos, but because they lost their jobs.
I agree automation has the potential to make most lives easier, but the distribution of the benefits and downsides is not even. Owners of the automation don't necessarily want to share the benefits. Inequality is increasing and shows no signs of slowing down.
Automation is why you aren't at the stream beating your dirty clothes against a rock to clean them.
If you don't have money because a machine took your job, then you may end up doing similar in your bathtub to save a buck.
Economists should be able to see the 4% unemployment
Economies are cyclical and the slumps appear to be lasting longer. Plus, we don't know the full extent of eventual automation. Extrapolating past patterns is an imperfect way to know the future of jobs. At least be somewhat prepared by asking tough questions up front. Why bash economists for asking questions?
Why is that infinitely valid? At some point there won't be new jobs that can be created.
That's the tricky question. Just because some pattern happened for 200 years does not mean it will continue indefinitely. But it's really hard to say when it will stop. It's kind of like a Moore's Law: It's not a law, just an observed pattern.
I look at it this way: machines are gradually growing ever smarter while humans stay roughly the same intelligence. Eventually machines will do too many jobs better than the dumbest humans, and gradually crawl up the intelligence ladder to medium intelligence, etc. But nobody really knows when the lines will meet.
Scooters pollute less and use less petroleum than cars. They should be encouraged. Sounds like more traffic and parking enforcement is needed for them, however. Add that into the licensing fees.
Slashdot Mirror
"x86" is kind of a generic name of a "line of" processors, similar to the System/360 "architecture" even though current machines go beyond the original System/360.
English is almost inherently lossy anyhow such that it's probably not worth debating. One can use "the x86 line" to hopefully indicate a more generic implication.
Sometimes choice is better than progress, or at least feels that way to humans.
Does universal logic dictate that progress is more important than choice? I'd like to see something close to a formal proof if you claim so. You know, pedantic debate fun.
Couldn't a general CPU be a fall-back option if specialized processors are not available at a given time or environment?
I'm no "chip expert", but I can envision at least 2 ways this could be done: one is by emulation. If the specialized CPU is not available, then the general CPU processes the task in emulation mode.
Another approach is to use the same machine language on the specialized chips, but optimize specific kinds of processing and/or instructions. For example, if AI uses a lot of the A and B instructions, but D and C are rarer in AI, then the "AI chip" would run instructions A and B really fast, but slack on D and C to save resources for the first two. If by chance the AI chip is not available, the general CPU can still process the AI-centric task, just not as fast.
China may be hedging its bets by investing in both kinds of CPUs: x86 and ARM.
Some argue x86 is dying, but x86 has more server-centric features than ARM, and is thus is still popular for server farms. (At least the x86 server features are more mature.)
I wonder if GPU's will overtake both of these, or at least push x86 and ARM into being mostly coordinators. Perhaps the market will shift to specialized chip-sets for AI, databases, graphics, etc., and x86 or ARM will mostly function as process coordinators which dish out specific tasks to specialized CPUs.
The titel is cuased by a Pentuim flaoting piont erorr. Its why we nead competion.
(Slahsdot just fixed the erorr, by the way.)
Good! AMD and Intel need competition in x86 chips. Duopolies usually provide narrow choices, and consumers can't do squat about it (big telecoms cough cough).
It was aimed using rays, not propelled. A reaction wheel and its spare failed, so they used Mr. Sunshine instead. Not as good as the original, but they were still able to do science.
Until this theory is fully tested in space, perhaps craft should use ceramic bearings. What are their drawbacks?
But electricity is needed to keep Earth's flat disk together.
My father used to be one under the title "industrial engineer". He started with factories, but as they offshored, worked with hospitals. Some people resented being told how to do their job, so there were office politics fireworks at times.
He'd do it at home also, arguing that we wasted toilet paper; and we had strange arguments over the physics of wiping and average consistency of poop ("bar" charts, eeww). "Everyone is different, how do you know my consistency profile is the same as yours? You are guessing, admit it!"
I kind of have the same "gene" and argue that our development stack wastes too many keystrokes, eye movements, and has unnecessary code pattern redundancies as alleged insurance against unlikely future events. The primary architect doesn't like me. He uses bullshit buzzwords to try to counter me. The fight ain't over...
Anyhow, I'm not sure how it relates to most of the article, though. My dad's studies were not done in secret. Most of the article was not about training robots to be better or more efficient, but about faking AI by acting like or in place of a bot to trick investors and/or buy time when bots are not ready.
bleep blurp glips
He's, living, long, and, ..... prospering.
Warren is relatively stingy with his kids even now. Why would that change when he dies? Although, I do agree that many charities are fronts.
Well-known languages seem to collect cruft as they age. Look up and try to do reflection on C#'s nullable-types for example. Nullable types were an after-thought such that they ended up bastardizing dynamic types to implement them, making for ugly reflection. Granted, perhaps there were other ways to implement nullable types, but I'm sure those would uglify something else.
Correction, "formerly known as"
"Welcome to Creepy Airlines, formally known as Crying Baby Airlines."
for a reason
From you.
Explains the hair
Or "One Flew Over the Cuckoo's Nest".
That's not true. The Luddites didn't riot because they hated new gizmos, but because they lost their jobs.
I agree automation has the potential to make most lives easier, but the distribution of the benefits and downsides is not even. Owners of the automation don't necessarily want to share the benefits. Inequality is increasing and shows no signs of slowing down.
If you don't have money because a machine took your job, then you may end up doing similar in your bathtub to save a buck.
Economies are cyclical and the slumps appear to be lasting longer. Plus, we don't know the full extent of eventual automation. Extrapolating past patterns is an imperfect way to know the future of jobs. At least be somewhat prepared by asking tough questions up front. Why bash economists for asking questions?
That's the tricky question. Just because some pattern happened for 200 years does not mean it will continue indefinitely. But it's really hard to say when it will stop. It's kind of like a Moore's Law: It's not a law, just an observed pattern.
I look at it this way: machines are gradually growing ever smarter while humans stay roughly the same intelligence. Eventually machines will do too many jobs better than the dumbest humans, and gradually crawl up the intelligence ladder to medium intelligence, etc. But nobody really knows when the lines will meet.
Conservatives would rather wear drag to church than consider that. It just won't fly in the USA. I'm just the messenger.
Scooters pollute less and use less petroleum than cars. They should be encouraged. Sounds like more traffic and parking enforcement is needed for them, however. Add that into the licensing fees.
Conservatives are usually the ones for big full prisons.