YOU don't understand how statistics ACTUALLY work.
You don't seem to understand. The whole point of statistics is to give predictive power. If you can't predict what the 100,001st drive will probably do, then you're not using statistics.
Quantum "random" may be random to any observer in our Universe, but that's probably because you would need perfect information about the state of the entire Universe in order to predict it perfectly. This is still pseudo-random, just very chaotic.
In theory, 256bits of entropy is enough to be impossible to break. Pretty much a lifetime of "random" numbers can be created from 256bits, it doesn't take much. The problem is our algorithms and systems are not perfect and data gets leaked for various reasons. As long as you can seed the system with something like 256bits of strong entropy, you can slowly collect additional entropy over time, even if very slowly, and just keep mixing it in. All that matters is you acquire entropy some factor faster than you leak it.
The Universe is just varying degrees of pseudo-random. With the Avalanche effect, even small amounts of "entropy" mixed into psuedo-random. The biggest issues with making strong random numbers isn't creating high quality 50/50 distribution of 1s and 0s, but making sure the RNG isn't susceptible to some form of malicious "entropy" from a compromised entropy source or leaking data about the state of the RNG.
They are free to leave society. If they want to be in society, then they are obligated to give back. And it wasn't forced upon them. Their parents could have left them as a new born in a wilderness where they would have been to do whatever they wanted with their life.
Smart people have issues with stress and high school can be very stressful. All of my teachers said I was very smart, but I was getting Cs and Ds. I took 3 years off of school, just played Quake and watched TV for 3 years, and any self-education I did because I'm naturally inquisitive. Went back in 11th grade and suddenly jumped to the top of my class.
My other brother had a similar issue. He was failing his math and his teacher wanted to hold him back, but my mom threw up a shit storm and got him moved from freshman math into senior AP math, and suddenly he was getting all As. Graduated with high honors. Now he's in in a University, his adviser is grooming him for a PHD in AI, is getting all kinds of special treatment from the University for skipping many required classes, and is interning for high performance computing at a state research facility. He's going to be learning to program super computers, which he was told students almost never get allowed to do.
He and I get along well because I understand what he's saying and can contribute back. I also do a lot of high performance parallel programming at my job. I found it naturally easy.
Now that I'm older, I have all kinds of problems with anxiety. A bit of research and it seems it comes with the territory.
Of the literature that I've read about anti-oxidants and cancer, they did not increase the risk of getting cancer, only made cancer potentially worse one you got it.
Cellphones output less power and lower frequency than IR at room temperature. Human body temperature emits more than 400watts per square meter of skin in blackbody radiation. The only danger cellphones create is localized heating because they're not in thermal equilibrium, but the total energy is much lower.
Personally, I'm more worried about the localized heating caused by taking a hot shower. That is much more EM radiation than my cell phone and more deadly type of radiation, and it dries out my skin. A hot tub is right-out, along with exercising, and almost anything else that warms up your body more than a cell phone. The only thing more deadly is a lightbulb, of any kind. Very high frequency EM radiation compared to microwave, and much higher amounts of total radiation. Even possibly localized heating if you sit near an incandescent bulb.
Most of the specs are between the lines, the specs constantly change even if slightly, and the specs don't include future major features that could have easily been imagined. I've head to deal with a few SLA'd high feature creep projects, and 80% of my job was to ask questions about the proposed specs, and many times make educated guesses when I could not ask the customer.
Short of a "hello world" class project, creativity is very important if you don't want a brittle product.
80% of people who apply for programming courses at Universities fail in the first two semesters. Of the remaining 20%, over half of them can't create a proper mental model of what is going on and only get by with luck and determinism. Only 10% of people who even apply are useful, and those skills are power-curve distributed.
I've had a decent time with them, and so have my family, but they still have issues every night where YouTube, Netflix, or Hulu will randomly buffer. Mostly good, but periodic drops in performance. Starting a video stream also takes a bit to get going. I no longer have them, and I have fiber Internet and have no buffering anymore.
One annoying issue with Charter was the random monthly internet drop out. Worse is you try contacting customer support and they say your connection looks fine from their end. After 10-15 minutes of them trying to have you reboot your modem, computer, and light bulbs, they go "ohh, yeah, we can't contact your modem". WTF did you mean by "your connection looks fine" if you can't fk'n communicate with it?!
I guess you missed many of those/. posts about peak bandwidth over the years. Even if not "world wide", Netflix(~35%) and Youtube(~20%) make up 50% of peak USA internet usage and Netflix has claimed during FreeBSD conferences that they are near the 10Tb/s mark. It doesn't take much to figure out the USA's peak bandwidth is below 100Tb/s, or was several years back when this data was valid.
I do agree that those network devices are expensive, but when you're making $3bil+/month in revenue from the services those router provides, they are suddenly very cheap.
Only 900Tb/s? That's only 900,000 1Gb links and several times more bandwidth than peak world wide usage. If a single router has enough bandwidth to handle the entire world without congestion, then 10 of these should be able to be 10x faster.
Use an AQM to dynamically apply QoS based on flow usage. All flows get the same priority, but flows that use more get less priority. Done at a microsecond level. Look up Cake, they can do this with a small fixed amount of state, but scales to an infinite number of flows with no increase of state storage. Torrent all you want, VoIP won't even notice. You don't even need to classify traffic. Automagtically works, only need to know your bandwidth.
That's pretty bad. The few times we had wells, we had about 100 gallon pressured water tanks. You could take a bath and still have enough pressure to drink. Municipal Water Supply has several water towers here. The primary is 2mil gallons, but the smaller ones are around 500k gallons. Not too bad for a small city of around 30k.
They also have water pumps around the city for normal usage. They've fitted all home water pipes with a long range radio that can give real time water usage from around the city. This has saved them a lot of energy costs in running the pumps, dropping our bills a few cents per unit.
You need both ways. Some people are intuitively good at something and they think about things a different way. Don't try to change how they think, it's clearly working for them. At the same time, most people don't find everything to be intuitive, and they need to actually understand the problem, not just memorize stuff.
I like to say "Pattern Anti-Pattern". Any time you use code without understanding what it's doing, you're doing Cargo Cult programming. Most people like to say high level languages mean you don't need to know what's going on, that just means they're Cargo Cult programmers.If you ever created a program that did not work as expected, it's probably because of Cargo Cult programming. Computers are logical systems that can be reasoned. For the most part, they work exactly as expected.
Takes all kinds. I know I like to work on infrastructure related work. In an overly simplified description, I am a force multiplier and enable other programmers, but without them, I am not nearly as useful. Symbiotic.
apply logic to questions and use evidence to arrive at conclusion
When you start running into hard problems, you can't use evidence because the act of measuring changes the results. When you've reached this point of programming, then you need a really good understanding of theory and an excellent mental model. Try debugging higher performance multi-threaded async code where many errors have no stack traces.
Even more important is being able to debug hypothetical designs in your head to quickly weed out bad designs before you get stuck debugging them.
Slashdot Mirror
They're using all of the drives the same way and one brand fails 3x+ more than others. It doesn't matter if they're using the harddrives "wrong".
YOU don't understand how statistics ACTUALLY work.
You don't seem to understand. The whole point of statistics is to give predictive power. If you can't predict what the 100,001st drive will probably do, then you're not using statistics.
Quantum "random" may be random to any observer in our Universe, but that's probably because you would need perfect information about the state of the entire Universe in order to predict it perfectly. This is still pseudo-random, just very chaotic.
In theory, 256bits of entropy is enough to be impossible to break. Pretty much a lifetime of "random" numbers can be created from 256bits, it doesn't take much. The problem is our algorithms and systems are not perfect and data gets leaked for various reasons. As long as you can seed the system with something like 256bits of strong entropy, you can slowly collect additional entropy over time, even if very slowly, and just keep mixing it in. All that matters is you acquire entropy some factor faster than you leak it.
Intel has a high quality HW-RNG, but no one trusts it. That's the bigger issue with anything hardware, it can't be verified, only software can.
The Universe is just varying degrees of pseudo-random. With the Avalanche effect, even small amounts of "entropy" mixed into psuedo-random. The biggest issues with making strong random numbers isn't creating high quality 50/50 distribution of 1s and 0s, but making sure the RNG isn't susceptible to some form of malicious "entropy" from a compromised entropy source or leaking data about the state of the RNG.
They are free to leave society. If they want to be in society, then they are obligated to give back. And it wasn't forced upon them. Their parents could have left them as a new born in a wilderness where they would have been to do whatever they wanted with their life.
Smart people have issues with stress and high school can be very stressful. All of my teachers said I was very smart, but I was getting Cs and Ds. I took 3 years off of school, just played Quake and watched TV for 3 years, and any self-education I did because I'm naturally inquisitive. Went back in 11th grade and suddenly jumped to the top of my class.
My other brother had a similar issue. He was failing his math and his teacher wanted to hold him back, but my mom threw up a shit storm and got him moved from freshman math into senior AP math, and suddenly he was getting all As. Graduated with high honors. Now he's in in a University, his adviser is grooming him for a PHD in AI, is getting all kinds of special treatment from the University for skipping many required classes, and is interning for high performance computing at a state research facility. He's going to be learning to program super computers, which he was told students almost never get allowed to do.
He and I get along well because I understand what he's saying and can contribute back. I also do a lot of high performance parallel programming at my job. I found it naturally easy.
Now that I'm older, I have all kinds of problems with anxiety. A bit of research and it seems it comes with the territory.
Of the literature that I've read about anti-oxidants and cancer, they did not increase the risk of getting cancer, only made cancer potentially worse one you got it.
Cellphones output less power and lower frequency than IR at room temperature. Human body temperature emits more than 400watts per square meter of skin in blackbody radiation. The only danger cellphones create is localized heating because they're not in thermal equilibrium, but the total energy is much lower.
Personally, I'm more worried about the localized heating caused by taking a hot shower. That is much more EM radiation than my cell phone and more deadly type of radiation, and it dries out my skin. A hot tub is right-out, along with exercising, and almost anything else that warms up your body more than a cell phone. The only thing more deadly is a lightbulb, of any kind. Very high frequency EM radiation compared to microwave, and much higher amounts of total radiation. Even possibly localized heating if you sit near an incandescent bulb.
Most of the specs are between the lines, the specs constantly change even if slightly, and the specs don't include future major features that could have easily been imagined. I've head to deal with a few SLA'd high feature creep projects, and 80% of my job was to ask questions about the proposed specs, and many times make educated guesses when I could not ask the customer.
Short of a "hello world" class project, creativity is very important if you don't want a brittle product.
Side question. The Oxygen has to go somewhere. The atmosphere? Chemical reactions?
80% of people who apply for programming courses at Universities fail in the first two semesters. Of the remaining 20%, over half of them can't create a proper mental model of what is going on and only get by with luck and determinism. Only 10% of people who even apply are useful, and those skills are power-curve distributed.
I've had a decent time with them, and so have my family, but they still have issues every night where YouTube, Netflix, or Hulu will randomly buffer. Mostly good, but periodic drops in performance. Starting a video stream also takes a bit to get going. I no longer have them, and I have fiber Internet and have no buffering anymore.
One annoying issue with Charter was the random monthly internet drop out. Worse is you try contacting customer support and they say your connection looks fine from their end. After 10-15 minutes of them trying to have you reboot your modem, computer, and light bulbs, they go "ohh, yeah, we can't contact your modem". WTF did you mean by "your connection looks fine" if you can't fk'n communicate with it?!
Citation needed on that peak worldwide number.
I guess you missed many of those /. posts about peak bandwidth over the years. Even if not "world wide", Netflix(~35%) and Youtube(~20%) make up 50% of peak USA internet usage and Netflix has claimed during FreeBSD conferences that they are near the 10Tb/s mark. It doesn't take much to figure out the USA's peak bandwidth is below 100Tb/s, or was several years back when this data was valid.
I do agree that those network devices are expensive, but when you're making $3bil+/month in revenue from the services those router provides, they are suddenly very cheap.
Only 900Tb/s? That's only 900,000 1Gb links and several times more bandwidth than peak world wide usage. If a single router has enough bandwidth to handle the entire world without congestion, then 10 of these should be able to be 10x faster.
Use an AQM to dynamically apply QoS based on flow usage. All flows get the same priority, but flows that use more get less priority. Done at a microsecond level. Look up Cake, they can do this with a small fixed amount of state, but scales to an infinite number of flows with no increase of state storage. Torrent all you want, VoIP won't even notice. You don't even need to classify traffic. Automagtically works, only need to know your bandwidth.
How much you use does not matter, it's when you use it. Unless you're using your phone at 2am, you're not reducing congestion by spreading it out.
Gravity doesn't pull because that would imply acceleration. Gravity is an emergent property of curved spacetime, but is not a force.
That's pretty bad. The few times we had wells, we had about 100 gallon pressured water tanks. You could take a bath and still have enough pressure to drink. Municipal Water Supply has several water towers here. The primary is 2mil gallons, but the smaller ones are around 500k gallons. Not too bad for a small city of around 30k.
They also have water pumps around the city for normal usage. They've fitted all home water pipes with a long range radio that can give real time water usage from around the city. This has saved them a lot of energy costs in running the pumps, dropping our bills a few cents per unit.
You need both ways. Some people are intuitively good at something and they think about things a different way. Don't try to change how they think, it's clearly working for them. At the same time, most people don't find everything to be intuitive, and they need to actually understand the problem, not just memorize stuff.
I like to say "Pattern Anti-Pattern". Any time you use code without understanding what it's doing, you're doing Cargo Cult programming. Most people like to say high level languages mean you don't need to know what's going on, that just means they're Cargo Cult programmers.If you ever created a program that did not work as expected, it's probably because of Cargo Cult programming. Computers are logical systems that can be reasoned. For the most part, they work exactly as expected.
Takes all kinds. I know I like to work on infrastructure related work. In an overly simplified description, I am a force multiplier and enable other programmers, but without them, I am not nearly as useful. Symbiotic.
Divas are a dime a dozen, but excellent programmers are unicorns.
apply logic to questions and use evidence to arrive at conclusion
When you start running into hard problems, you can't use evidence because the act of measuring changes the results. When you've reached this point of programming, then you need a really good understanding of theory and an excellent mental model. Try debugging higher performance multi-threaded async code where many errors have no stack traces.
Even more important is being able to debug hypothetical designs in your head to quickly weed out bad designs before you get stuck debugging them.