Dynamic impedance matching is not really a concern; it's a question of available current. 25 W rms into a 4 ohm speaker requires 3.5 A peak. it's an unusual vacuum tube that will pass even 200 mA, so unless you parallel tubes, a transformer is required to boost the current by 17.5 times, for instance. The transformer of course will simultaneously divide the voltage by 17.5, resulting in a (largely irrelevant) impedance transformation of.306:1.
Transistors are really good at pumping out lots of current, but at high voltages transistors (particularly bipolar transistors) suffer from constraints that make design more difficult: loss of gain-bandwidth product and secondary breakdown. (For very high power audio amplifiers, even 4 ohm speakers have an impedance higher than would make amplifier design easiest.)
Whether you can see distortion products 30 dB down on an oscilloscope depends upon the nature of the distortion. Crossover distortion (a glitch near 0 volts due to a class AB circuit being too close to class B) is visible. Hard clipping is visible at 30 dB down if you look carefully.
He's not going into volume production and he can buy filaments rather than drawing and alloying his own with tungsten and thorium or whatever. This is not EPA territory.
Having done both, it's much easier to build a solid state amp than a tube amp from scratch. Neither requires a PCB, but tubes almost always have to be socketed and usually need a chassis. If the signal level is low. the filament voltage either needs to be DC or needs to be shielded from the signal path. Transistors and ICs don't have filaments. ICs provide more gain per stage, and a power amp can be made with just 1 or 2 ICs, depending on the gain required. Tube amps require an output transformer. Tubes use much higher voltages, increasing the risk of painful or deadly shocks.
Because tubes have to heat the cathode in order to eject electrons, at audio frequencies the best tubes are inherently more noisy than the best transistors. It's not a lot of noise - in a properly designed circuit it's unlikely to be noticeable - but physics does not allow a tube at audio frequencies to be as quiet as a transistor.
News organizations protect their sources because if their sources think they can be identified, they will stop being sources in order to not be identified. News organizations care very little about their sources except insofar as the sources provide information which allows the organizations to prosper.
Current technology of CMOS gates shows roughly the same energy per transition as a neuron (given similar levels of complexity.) Since AI uses so much power for such marginal results, this implies that computer hardware is not properly designed/optimized for intelligence work or that AI software is woefully wrong (or both).
Brain cells and brains as a whole aren't magical; they work by some mechanism. Equivalents of all mechanisms can be made by digital logic systems, but we don't know how to make the equivalent of a human brain / nervous system yet. When we do, we should expect equivalent performance.
A number of barriers exist.
We don't really know what's going on yet.
Brains are much more parallel than transistor systems, and replacing parallel with multiple serial is energy inefficient.
Brains are densely packed in 3D; transistors are densely packed in 2D and the third dimension by comparison is 100 to 1000 times larger (A neuron is about 1 micron. IC wafers are about 1 mm thick, although they can be shaved down.)
At least some long-term learning in brains occurs by rewiring, rewiring in digital circuits is unusual (reprogrammable FPGAs) and inefficient - digital circuits learn by storing data.
Not only is there a long way to go, it seems to me that a GPU is spectacularly inappropriate for AI work
Free markets and "throwing sackfuls of cash" at those who govern us to influence their actions, are mutually exclusive. Free means without restrictions, and government action frequently restricts.
Without the bribery from the cane sugar industry, HFCS would probably never have become an issue. Cane sugar manufacturers pushed for high tariffs so that they could be protected from foreign competition. Domestic sugar prices went up, causing sugar users to search for a cheaper alternative. They found fructose, which could be cheaper and is also slightly sweeter than cane sugar.
Both those offering bribes and those accepting them are doing wrong. Having a government with enough power to make bribery profitable is also a problem.
Rotted foods are usually less healthy to eat than foods not rotted. Two properties commonly bred for are shelf life and disease resistance, both of which result in more foods that are not rotted.
Another property bred for is insect resistance, but that's a two-edged sword: insect resistance without harm to people is probably harder to achieve.
The fact is, sugar is addictive, and that's why we like it
That's awfully close to being a circular argument.
We've evolved to like sugar because it's an indicator of ripeness, an indicator that fruit is ready to eat. Some unripe fruits are bitter or sour or hard, and would be unpleasant and possibly less healthful to eat.
Causing withdrawal symptoms when consumption ends is one characteristic of a substance being addictive, and usually ending sugar consumption doesn't cause pain or depression, only a thought of "I'd like something sweet." Calling it an addiction is too strong, it's more like a bad habit.
Mars is the second-nearest big thing in space that can be walked around on without protection against an extremely toxic and corrosive environment. Mars has at least some potential for terraforming, in that it can hold an atmosphere much better than Earth's moon. Details of the moon are easy to see with a good telescope, but Mars requires more effort to get equivalent information. Other places potentially available for (difficult) colonization are farther away (Jovian moons?, etc.)
Prices vary, but drilling out and capping an abscessed tooth is about $1200; I assume an extraction is less.
Here's an important difference: I pay for the dentist and his staff, and I'm paying for the damage that my inadequate care of my teeth caused. You're paying (on average) for the dentist and his staff and for a government bureaucrat. It's less efficient; the country as a whole is paying more than it would without the bureaucracy. Also, morality is being short-circuited: careless people pay little for the damage they cause; careful people are forced to pay for services they don't use.
Not to deny your main point, but the chance of infection increases with the quantity of the infecting agent. Most people are unlikely to get sick inhaling a couple of anthrax spores, but eat a pint of them and you're toast.
Anthrax can survive hundreds of years of dormancy. There's a documented example of 370 years here: http://www.standard.co.uk/news/crossrail-work-stopped-after-human-bones-found-on-site-6759649.html
Slashdot Mirror
Dynamic impedance matching is not really a concern; it's a question of available current. 25 W rms into a 4 ohm speaker requires 3.5 A peak. it's an unusual vacuum tube that will pass even 200 mA, so unless you parallel tubes, a transformer is required to boost the current by 17.5 times, for instance. The transformer of course will simultaneously divide the voltage by 17.5, resulting in a (largely irrelevant) impedance transformation of .306:1.
Transistors are really good at pumping out lots of current, but at high voltages transistors (particularly bipolar transistors) suffer from constraints that make design more difficult: loss of gain-bandwidth product and secondary breakdown. (For very high power audio amplifiers, even 4 ohm speakers have an impedance higher than would make amplifier design easiest.)
Whether you can see distortion products 30 dB down on an oscilloscope depends upon the nature of the distortion. Crossover distortion (a glitch near 0 volts due to a class AB circuit being too close to class B) is visible. Hard clipping is visible at 30 dB down if you look carefully.
In RF amplifiers, the harmonics are filtered out. The transistors themselves are not particularly linear.
First you have to mine iron ore so that you cane make the shovel to mine copper ore.
He's not going into volume production and he can buy filaments rather than drawing and alloying his own with tungsten and thorium or whatever. This is not EPA territory.
Having done both, it's much easier to build a solid state amp than a tube amp from scratch. Neither requires a PCB, but tubes almost always have to be socketed and usually need a chassis. If the signal level is low. the filament voltage either needs to be DC or needs to be shielded from the signal path. Transistors and ICs don't have filaments. ICs provide more gain per stage, and a power amp can be made with just 1 or 2 ICs, depending on the gain required. Tube amps require an output transformer. Tubes use much higher voltages, increasing the risk of painful or deadly shocks.
Are you referring to dynamic range compression or data compression?
Because tubes have to heat the cathode in order to eject electrons, at audio frequencies the best tubes are inherently more noisy than the best transistors. It's not a lot of noise - in a properly designed circuit it's unlikely to be noticeable - but physics does not allow a tube at audio frequencies to be as quiet as a transistor.
News organizations protect their sources because if their sources think they can be identified, they will stop being sources in order to not be identified. News organizations care very little about their sources except insofar as the sources provide information which allows the organizations to prosper.
Current technology of CMOS gates shows roughly the same energy per transition as a neuron (given similar levels of complexity.) Since AI uses so much power for such marginal results, this implies that computer hardware is not properly designed/optimized for intelligence work or that AI software is woefully wrong (or both).
Brain cells and brains as a whole aren't magical; they work by some mechanism. Equivalents of all mechanisms can be made by digital logic systems, but we don't know how to make the equivalent of a human brain / nervous system yet. When we do, we should expect equivalent performance.
A number of barriers exist.
Not only is there a long way to go, it seems to me that a GPU is spectacularly inappropriate for AI work
ln -s /usr/bin/xpdf /usr/bin/acrobat
If it walks like a duck and quacks like a duck...
Rocky Mozell can sell to a billion fools.
Let's start with Bill Nye.
Free markets and "throwing sackfuls of cash" at those who govern us to influence their actions, are mutually exclusive. Free means without restrictions, and government action frequently restricts.
Without the bribery from the cane sugar industry, HFCS would probably never have become an issue. Cane sugar manufacturers pushed for high tariffs so that they could be protected from foreign competition. Domestic sugar prices went up, causing sugar users to search for a cheaper alternative. They found fructose, which could be cheaper and is also slightly sweeter than cane sugar.
Both those offering bribes and those accepting them are doing wrong. Having a government with enough power to make bribery profitable is also a problem.
Rotted foods are usually less healthy to eat than foods not rotted. Two properties commonly bred for are shelf life and disease resistance, both of which result in more foods that are not rotted.
Another property bred for is insect resistance, but that's a two-edged sword: insect resistance without harm to people is probably harder to achieve.
Saccharose is an obsolete name; sucrose is preferred.
That's awfully close to being a circular argument.
We've evolved to like sugar because it's an indicator of ripeness, an indicator that fruit is ready to eat. Some unripe fruits are bitter or sour or hard, and would be unpleasant and possibly less healthful to eat.
Causing withdrawal symptoms when consumption ends is one characteristic of a substance being addictive, and usually ending sugar consumption doesn't cause pain or depression, only a thought of "I'd like something sweet." Calling it an addiction is too strong, it's more like a bad habit.
Mars is the second-nearest big thing in space that can be walked around on without protection against an extremely toxic and corrosive environment. Mars has at least some potential for terraforming, in that it can hold an atmosphere much better than Earth's moon. Details of the moon are easy to see with a good telescope, but Mars requires more effort to get equivalent information. Other places potentially available for (difficult) colonization are farther away (Jovian moons?, etc.)
You completely miss the point of the (apocryphal) quote of Franklin, who was greatly interested in all scientific endeavors.
Cows on Mars! Yes! Not only will no human lives be at risk, we can test the hypothesis that methane causes global warming.
And if cow colonization is successful...
All hail our new bovine Martian overlords!
Prices vary, but drilling out and capping an abscessed tooth is about $1200; I assume an extraction is less.
Here's an important difference: I pay for the dentist and his staff, and I'm paying for the damage that my inadequate care of my teeth caused. You're paying (on average) for the dentist and his staff and for a government bureaucrat. It's less efficient; the country as a whole is paying more than it would without the bureaucracy. Also, morality is being short-circuited: careless people pay little for the damage they cause; careful people are forced to pay for services they don't use.
Not to deny your main point, but the chance of infection increases with the quantity of the infecting agent. Most people are unlikely to get sick inhaling a couple of anthrax spores, but eat a pint of them and you're toast.
Anthrax can survive hundreds of years of dormancy. There's a documented example of 370 years here: http://www.standard.co.uk/news/crossrail-work-stopped-after-human-bones-found-on-site-6759649.html
Aleppo is the absence of leppo.