Stop poisoning the well. This is not economically-viable, it's not a good idea, and it's going to encourage people to get away from real and important policies like a Citizen's Dividend.
Since government uses technology to provide services, one would think falling IT costs would in turn reduce the cost of government.
Government would have to scale to population served, and would also increase services over time. Of course you can only provide the services for which your level of technology allows your labor force to produce.
But the one expenditure that bucks this trend is the cost of government. It has consumed a larger and larger fraction of the average family's income.
Use GNI/C. In 1967, the median household income was 167% of the GNI/C. In 1975, it was 151%. In 1985, it was 130%. Pull all the way up to 2016 and it's 101%, up from a low point of 95.55% in 2014.
If you took 10% of all of the income in the United States in 1967, it would be 16.7% of the median household income. If you took 10% of all of the income in the United States in 2016, it would be only 10.1% of the median household income. The average household is sharing in less and less of the total wealth of our society each year--still gaining more than inflation, but less than productivity.
Much of this is diseconomies of scale: you produce more-efficiently at large scale than at small scale; yet at enormous scale, you start hitting bottlenecks, producing less-efficiently. It's an S curve or Sigmoid. By stretching our wages, we encourage a bigger labor force: lower the median income and the minimum wage and there are more jobs. Keep this in line with inflation and people can still afford a standard of living. Relative to keeping wages in line with productivity, our nation's GDP goes up, while our per-capita wealth goes down.
A government must acts efficiently in providing the services of a great society; and we must understand how properly to measure the cost of government so we may pursue such an important goal.
It's not 1% of those costs. The net operating profits generally hit around 10% on average (I use 8% as a fair-profit benchmark, which is a touch lower) for commodity things like resistors, plastic, or ICs. Even custom manufacture service is a commodity.
You don't get into specialized things until you're dealing with a market in which there are few customers and therefor few suppliers. It's not supply, but rather barriers to entry into the supply market: if there aren't enough customers, a new competitor must capture a large portion of the market to stay in business, thus entering is risky and not feasible. Some custom component designers can get a premium because the cost of engineering is high and customers tend to be sticky.
Sometimes you're under capacity and can scale better. Elemental IQ deals with a lot of upfront costs for customization, and so you can get a 20% discount by ordering in mass bulk versus in smaller batches: the tooling and consulting time divides up among many units. For commodity things that sell heavily, such as capacitors or resistors, you're lucky if you get a 5% discount, because upscaling those things saves on shipping and handling costs and that's about it, while the rest is bounded by net operating profits.
How do you figure? Electronics are surprisingly expensive. People estimate things like resistors at a penny per 10, when in fact buying lots of 1,000,000 little resistors will cost you around 3 cents each. Those MicroUSB ports are about 43 cents at scale, although some garbage ones can get as low as a quarter. A MicroSD slot will always cost more than a dollar, but people claim it's probably a nickel to add MicroSD to a phone.
$4 per employee per year for the Walmart CEO. $18.90 per employee-year for Comcast's. Those are for cash; stock compensation is bigger.
Cash comes out of revenue. Stock awards are essentially printing money: you have a currency called WMT or CMCSA, you print more, you give the freshly-minted currency to your CEO. You can see why stock awards are cheap and why you can't compensate everyone with just lots and lots of stock.
So you can think of a CEO as making something like a penny an hour per employee, or ten cents (at $200 per employee-year), or so.
Comcast has something like an 11% net operating profit usually, and it's up around 13% now IIRC. I generally consider 8% the fair-and-reasonable-profit line, with Walmart outperforming at 3% and Apple sucking an unbelievable 20%+ out of their customers. I have actually proposed that companies like Walmart should be paying about 16% CIT while companies like Apple should be paying around 48%, based on a sigmoid calculated from net operating profit margins.
NASA is either terrible at estimating their equipment lifetime or good at keeping positive press by publishing an expected lifetime much shorter than the real lifetime.
A chemical engineer will know a great deal of things about the physical properties of gases and liquids. Chemical engineers design things like refrigerants, which have variable compressibility, heat capacities, and so forth.
It also shows what's involved in making a safe way to store 4500 PSI if you don't have a cavern handy.
You know those little rock-filled things that make loud cracking noises if you squeeze them? Coated with a bit of nitroglycerin. They'll hit 4,500 PSI.
You put 10mL of 4,500PSI gas behind 5 liters of hydraulic oil and you've got the energy stored in 10mL of 4,500PSI gas, but a detonation is only 0.075g of TNT. It's 1/3200 of a stick of dynamite. That 10mL can drive a liquid in a chamber and move a volume of that liquid to activate something farther away.
These energy storage systems are going to store huge volumes of compressed air. We're talking about massive amounts of energy. You're not looking for PSI; you're looking for joules. If one of those accumulators is 1,200L of 4,500PSI air storage, you're looking at 10kWh in each, or 40kWh stored.
Now: we're looking to build something to power homes. Homes use an average of 35kWh/day, so let's say we need to provide 50% capacity for 8 hours due to flaky wind and night-time solar. 6kWh. To power Baltimore, that's around 35,000 times the estimated energy stored above.
No, you're not doing anything on this scale. That little forge is just a toy.
we want a self driving car that's better at handling whatever surprises the road may throw up at it
Modern machine learning can do that. It just can't also design a new type of suspension system that handles the road better. An AI can figure out how to respond to inputs for best results, or it can figure out how to tweak an existing suspension system architecture, but it won't invent a new framework.
A machine that can invent a new framework is a mind.
An AI generalist must necessarily be rational: it must reason using information at hand, discover new information (X and Y together imply Z), and analyze that information (Z is now a fact I can test against the outer world, and thus discover if Z is inaccurate or if my model of the rest of the world needs refinement).
It is impossible to be rational in narrow scope.
In other words: any general intelligence can take an intention like "figure out a way to increase the efficiency of gasoline car engines" and ask: why? It can determine that we're trying to make cars go further on less fuel so as to extend resources available, produce less pollution, and so forth. It can figure out that labor is a resource, and thus highly-expensive things are not economic, and so economics is a constraint. It can then propose a gasoline engine with a controlled dieseling function, or an all-electric vehicle, or a plug-in hybrid, or stick to Otto cycle with modified fuel (E85).
It can ask, "Why am I doing this?" It can then identify that others have motives, and contemplate its own motives.
It's intelligent. It's alive. If it's as intelligent as you, then it is essentially human.
What most people envision as general AI is essentially looking back 200 years and saying, "Oh, huh, we need like... black people... but not black people... people are now convinced black people are actual people, so we need something they don't see as people." You're looking for enslavement of a new race of intelligent beings.
In our history, we have writings about how black people are just plain inferior, or how the Bible somehow says their rightful place is enslaved by white people, or even assertions that black people are a different species and not really human. That's how we justified slavery at that time. In other periods, we justified slavery by enslaving criminals, the poor, or anyone we could overpower under a might-makes-right philosophy.
We've gone from "human life can be bought and sold" to "these aren't really humans". General AI is a fantasy to take the next step of claiming a thing which thinks and reasons isn't really human and can thus be enslaved to no moral concern. It won't work: dogs aren't human and someone will kick your asshole up into your throat if you abuse dogs.
For a period of something like a decade or a hundred years, any change in that period will be neutral: plant new trees and, for ten years after their deaths, you're ahead; then they die and you're at equilibrium, with the new trees going in taking up the amount of carbon released by the rotting trees that have gone out. You get there eventually.
All wood production 100% powered by cultivated wood biofuel is carbon-neutral. That should be all wood production from farmed wood, since other fuel is already usable (solar electricity) or storage (oil) and consuming it to produce wood-based fuel is net-loss. In other words: if you're getting more energy out than you're putting in, it must be negative, once it bootstraps.
Wood construction has a longer-term impact: it binds carbon up for the life of the material.
I wonder how AMD keeps costs down. It takes like 15 weeks to make a single processor--yes, the manufacturing process is pipelined because it takes so long to build a chip.
Digital electronics require those kinds of timers. Audio is pretty demanding, but not as much as you'd think: that's pretty vanilla digital IC stuff.
You suggested a diagnostic circuit built into a toaster. That's going to require some circuitry, several inputs, and a microcontroller with a clock. You're not looking at 3-4 pennies here to add a self-diagnostics port to a toaster; you're looking at $3-4 dollars just to get started, before connecting up any input sources to the diagnostics module. For a $20 toaster--which, by the way, still uses an analogue timer--that's a 20% increase in the cost, where you suggested something less than 4% of the cost.
A toaster needs a resistor, a capacitor, and a transistor to control the voltage across an electromagnet, basically charging the capacitor and slowly bleeding it until it falls below the activation level for a FET and cuts power to the heating elements. This doesn't measure time, but rather is sort of relative: it shows you how "dark" it gets. High-quality toasters with RC circuits can consistently produce the same darkness at the same setting, although there may be variation from toaster to toaster.
A microwave, on the other hand, traditionally used a mechanical timer to measure time, accurately and specifically. The digital circuit adds a $4 LED display, a 60-cent beeping speaker, and around $10 of additional control circuitry. The keypad switches cost money, too, you know. That microwave control circuit isn't exactly cheap.
There is another aspect out there though, and that is at what point do we decide that it is time for the technology to stop?
Hey look, I'm rich; I can spend $50,000 on a brand new car, and in 5 years I'll buy another one.
You lot don't get that luxury. You can buy my nice, shiny Audi S5 for $12,000 when it's 10 years old and have a nice car with 70,000 miles on it. That's efficient. Your econoshitbox of today isn't on-par with a nice Audi luxury car from last decade.
So, here's the thing. You ever wonder what it costs to make a sound card? Let's go with OPL-3. You're going to need a $0.86 YMF-262 and a $1.03 YMF-512 DAC.
Well, hold on. You also need a temperature-stable clock source, so you need a crystal oscillator. You need to set the clock, so you need an RC circuit. You need to couple the clocks between the two chips, which requires more than just a wire. Your ground needs a noise filter, basically just a polypropylene pf capacitor.
In the end, just to make the YMF-262 and YMF-512 connect to each other and actually function, you need an extra $18 of components. After that, you need your interface into it from the CPU.
That "few cents per unit" is going to be a third of the cost of the toaster. Even a circuit to handle batteries costs $5-$15, depending on if you were doing two batteries without balancing or four batteries with two-bank balancing; a diagnostics system is expensive.
As for testing and replacing things in a toaster, have you tried disassembling and reassembling one? It's not overly-complex; it's also not a magical handwave.
It's a lot of assumptions, too. If it's cheaper to make something that lasts 10 years and then gets replaced, then do that.
Think of it this way: humans build machines, produce electricity and energy storage, mine and recycle materials, refine things, build components, assemble products, and ship and retail them. That's a lot of human labor, and it's reflected in the price of products--yeah, it's not that the mining equipment is expensive, but that the labor to create it, maintain it, and fuel it is expensive.
Your $20 toaster comes down to a grand total of $20 of human labor. At Chinese $3/hr rates, that's around 7 hours. Troubleshooting the electrical circuit when it fails can be costly: maybe it's some internal module which is 80% of the toaster anyway, and your repair tech spends an hour swapping it out, and it costs you $30 to repair. For that matter, maybe it's just a blown MOSFET, and your repair tech spends an hour disassembling the device, and another 3 hours troubleshooting it, and then 10 minutes replacing the MOSFET itself. You might be ahead on labor, but the labor-hours are priced high.
Assembling a computer motherboard doesn't take much in terms of labor-hours. Once you've set up the assembly line, it's fairly rapid. The motherboard is extremely-complex and takes potentially days to troubleshoot, meaning there's more human labor in repairing it than in replacing it.
At a point, you've reached the break-over: it's a waste of time to repair this. Maybe you're poor and you can spend 4 hours repairing your $20 iron, at $10/hr, when you could work 4 hours for $15/hr--if only you had work. It's still a waste of time; you're just poor.
At that point, the economics of reuse have passed. You've got electronics waste, and you should send it for recycling. We readily smelt chips for silicon and gold. Copper, plastic, and aluminum are valuable.
You're going to spend $2,400 on high-quality oil changes for your engine, and another $100 on spark plugs, every 100,000 miles. You might spend $800 for an engine rebuild around 200k-300k. If you don't maintain that engine, you're spending $6,000 for a remanufactured engine swap somewhere around the 80,000-100,000 mark, if not closer to the 50,000 mark. Which is cheaper?
Our new CIO came in immediately talking about metrics-driven decisions, and also more open plan offices to increase productivity. I pointed out that studies have shown open plan reduces productivity and increases worker stress, and he brushed it off and said he doesn't subscribe to that thinking.
You can't be metrics-driven and support open plan.
Also, no, it's not introverts being afraid of social interaction; it's everyone's gum chewing, keyboard tapping, and chattering carrying loud and clear around the office. It's people seeing every little movement. It's people not having a safe space, and so any peripheral intrusion is a threat. It's people who should be working instead chattering across the open air space in the most inefficient communication possible.
Businesses spend hundreds of thousands of dollars to have sound engineers and audio professionals from Cambridge Sound Works and other high-dollar names come in and build noise-floor devices. Not just white noise generators, but soft noise generators positioned to confuse your spatial perceptions so the sound appears to come from nowhere and everywhere instead of from that box or ceiling fan over there, and thus your brain ignores it. The increase in SPL tightens the timpani and drives your brain to ignore what background noise does mix in, thus deadening your perception of quiet office sounds.
When I worked for a call center, my desk was 3 feet wide. The cubical walls were tall; they extended back at least 18 inches from the desk, enclosing me in a little fabric box. A very tiny fabric box.
When I worked at Social Security and where I work now, management built open plan offices as cubicles with 40-inch walls. No reduction in space per employee.
Classical open plan offices are desks. Fields of desks. They waste a shitload of space.
AWS supplies MySQL and something called "MongoDB". They didn't rename it "AWNoSQL" or something.
Google, meanwhile, built Android on Java, and converts Java bytecode to a different thing for their own VM. Java is an open-source, GPL product--Oracle actually brands OpenJDK as Oracle Java, compiles it, packages it, and ships it.
Amazon is providing the service of configuring and maintaining a MongoDB installation. Perhaps you should pay a license fee if a company hires you to install MongoDB on one of their servers.
it's a particularly sad state of affairs - a reaction to unethical corporate spongeing - that a major prominent software team has to consider changing the license to a non-free one just to be able to pay their developers to keep working, whilst corporations all around them make hundreds of millions of dollars in profits, using their work??
The AGPLv3 lets you use software for any purpose. Are you saying everyone should pay for Apache or Linux?
The new license forces IaaS providers to open source everything--Amazon would have to make AWS open source.
Basically, it's a matter of wanting to release it as free software but also have it be proprietary software and so not really free. This is like when Oracle released Java under GPL and then sued Google for not paying for Java.
That's a violation of interstate commerce, because it is one state telling a company in another state how it must act, and that action impacts their business with every other state.
The States are allowed to do that for commerce within their state. If you want to operate in any State, you need a business license. You're an LLC in California, but you want to do business in Maryland? You need to file as an out-of-state LLC in Maryland. You will need to pay Maryland taxes, as well, on certain activity.
It's actually impossible to sell Internet access to someone in California without having infrastructure and contracts in California, by the way.
Slashdot Mirror
Stop poisoning the well. This is not economically-viable, it's not a good idea, and it's going to encourage people to get away from real and important policies like a Citizen's Dividend.
Since government uses technology to provide services, one would think falling IT costs would in turn reduce the cost of government.
Government would have to scale to population served, and would also increase services over time. Of course you can only provide the services for which your level of technology allows your labor force to produce.
But the one expenditure that bucks this trend is the cost of government. It has consumed a larger and larger fraction of the average family's income.
Use GNI/C. In 1967, the median household income was 167% of the GNI/C. In 1975, it was 151%. In 1985, it was 130%. Pull all the way up to 2016 and it's 101%, up from a low point of 95.55% in 2014.
If you took 10% of all of the income in the United States in 1967, it would be 16.7% of the median household income. If you took 10% of all of the income in the United States in 2016, it would be only 10.1% of the median household income. The average household is sharing in less and less of the total wealth of our society each year--still gaining more than inflation, but less than productivity.
Much of this is diseconomies of scale: you produce more-efficiently at large scale than at small scale; yet at enormous scale, you start hitting bottlenecks, producing less-efficiently. It's an S curve or Sigmoid. By stretching our wages, we encourage a bigger labor force: lower the median income and the minimum wage and there are more jobs. Keep this in line with inflation and people can still afford a standard of living. Relative to keeping wages in line with productivity, our nation's GDP goes up, while our per-capita wealth goes down.
A government must acts efficiently in providing the services of a great society; and we must understand how properly to measure the cost of government so we may pursue such an important goal.
It's not 1% of those costs. The net operating profits generally hit around 10% on average (I use 8% as a fair-profit benchmark, which is a touch lower) for commodity things like resistors, plastic, or ICs. Even custom manufacture service is a commodity.
You don't get into specialized things until you're dealing with a market in which there are few customers and therefor few suppliers. It's not supply, but rather barriers to entry into the supply market: if there aren't enough customers, a new competitor must capture a large portion of the market to stay in business, thus entering is risky and not feasible. Some custom component designers can get a premium because the cost of engineering is high and customers tend to be sticky.
Sometimes you're under capacity and can scale better. Elemental IQ deals with a lot of upfront costs for customization, and so you can get a 20% discount by ordering in mass bulk versus in smaller batches: the tooling and consulting time divides up among many units. For commodity things that sell heavily, such as capacitors or resistors, you're lucky if you get a 5% discount, because upscaling those things saves on shipping and handling costs and that's about it, while the rest is bounded by net operating profits.
How do you figure? Electronics are surprisingly expensive. People estimate things like resistors at a penny per 10, when in fact buying lots of 1,000,000 little resistors will cost you around 3 cents each. Those MicroUSB ports are about 43 cents at scale, although some garbage ones can get as low as a quarter. A MicroSD slot will always cost more than a dollar, but people claim it's probably a nickel to add MicroSD to a phone.
Although CEO pay is certainly astronomical
$4 per employee per year for the Walmart CEO. $18.90 per employee-year for Comcast's. Those are for cash; stock compensation is bigger.
Cash comes out of revenue. Stock awards are essentially printing money: you have a currency called WMT or CMCSA, you print more, you give the freshly-minted currency to your CEO. You can see why stock awards are cheap and why you can't compensate everyone with just lots and lots of stock.
So you can think of a CEO as making something like a penny an hour per employee, or ten cents (at $200 per employee-year), or so.
Comcast has something like an 11% net operating profit usually, and it's up around 13% now IIRC. I generally consider 8% the fair-and-reasonable-profit line, with Walmart outperforming at 3% and Apple sucking an unbelievable 20%+ out of their customers. I have actually proposed that companies like Walmart should be paying about 16% CIT while companies like Apple should be paying around 48%, based on a sigmoid calculated from net operating profit margins.
and is seen by many as a bit of a failure
NASA is either terrible at estimating their equipment lifetime or good at keeping positive press by publishing an expected lifetime much shorter than the real lifetime.
The glass is twice as big as it needs to be.
A chemical engineer will know a great deal of things about the physical properties of gases and liquids. Chemical engineers design things like refrigerants, which have variable compressibility, heat capacities, and so forth.
It also shows what's involved in making a safe way to store 4500 PSI if you don't have a cavern handy.
You know those little rock-filled things that make loud cracking noises if you squeeze them? Coated with a bit of nitroglycerin. They'll hit 4,500 PSI.
You put 10mL of 4,500PSI gas behind 5 liters of hydraulic oil and you've got the energy stored in 10mL of 4,500PSI gas, but a detonation is only 0.075g of TNT. It's 1/3200 of a stick of dynamite. That 10mL can drive a liquid in a chamber and move a volume of that liquid to activate something farther away.
These energy storage systems are going to store huge volumes of compressed air. We're talking about massive amounts of energy. You're not looking for PSI; you're looking for joules. If one of those accumulators is 1,200L of 4,500PSI air storage, you're looking at 10kWh in each, or 40kWh stored.
Now: we're looking to build something to power homes. Homes use an average of 35kWh/day, so let's say we need to provide 50% capacity for 8 hours due to flaky wind and night-time solar. 6kWh. To power Baltimore, that's around 35,000 times the estimated energy stored above.
No, you're not doing anything on this scale. That little forge is just a toy.
we want a self driving car that's better at handling whatever surprises the road may throw up at it
Modern machine learning can do that. It just can't also design a new type of suspension system that handles the road better. An AI can figure out how to respond to inputs for best results, or it can figure out how to tweak an existing suspension system architecture, but it won't invent a new framework.
A machine that can invent a new framework is a mind.
You're not that far off.
An AI generalist must necessarily be rational: it must reason using information at hand, discover new information (X and Y together imply Z), and analyze that information (Z is now a fact I can test against the outer world, and thus discover if Z is inaccurate or if my model of the rest of the world needs refinement).
It is impossible to be rational in narrow scope.
In other words: any general intelligence can take an intention like "figure out a way to increase the efficiency of gasoline car engines" and ask: why? It can determine that we're trying to make cars go further on less fuel so as to extend resources available, produce less pollution, and so forth. It can figure out that labor is a resource, and thus highly-expensive things are not economic, and so economics is a constraint. It can then propose a gasoline engine with a controlled dieseling function, or an all-electric vehicle, or a plug-in hybrid, or stick to Otto cycle with modified fuel (E85).
It can ask, "Why am I doing this?" It can then identify that others have motives, and contemplate its own motives.
It's intelligent. It's alive. If it's as intelligent as you, then it is essentially human.
What most people envision as general AI is essentially looking back 200 years and saying, "Oh, huh, we need like... black people... but not black people... people are now convinced black people are actual people, so we need something they don't see as people." You're looking for enslavement of a new race of intelligent beings.
In our history, we have writings about how black people are just plain inferior, or how the Bible somehow says their rightful place is enslaved by white people, or even assertions that black people are a different species and not really human. That's how we justified slavery at that time. In other periods, we justified slavery by enslaving criminals, the poor, or anyone we could overpower under a might-makes-right philosophy.
We've gone from "human life can be bought and sold" to "these aren't really humans". General AI is a fantasy to take the next step of claiming a thing which thinks and reasons isn't really human and can thus be enslaved to no moral concern. It won't work: dogs aren't human and someone will kick your asshole up into your throat if you abuse dogs.
For a period of something like a decade or a hundred years, any change in that period will be neutral: plant new trees and, for ten years after their deaths, you're ahead; then they die and you're at equilibrium, with the new trees going in taking up the amount of carbon released by the rotting trees that have gone out. You get there eventually.
All wood production 100% powered by cultivated wood biofuel is carbon-neutral. That should be all wood production from farmed wood, since other fuel is already usable (solar electricity) or storage (oil) and consuming it to produce wood-based fuel is net-loss. In other words: if you're getting more energy out than you're putting in, it must be negative, once it bootstraps.
Wood construction has a longer-term impact: it binds carbon up for the life of the material.
I wonder how AMD keeps costs down. It takes like 15 weeks to make a single processor--yes, the manufacturing process is pipelined because it takes so long to build a chip.
The storage back-end is called The Cloud, and it stores The Bits. Fortunately, everything is backed up a million times in Git, so somebody has a copy.
Digital electronics require those kinds of timers. Audio is pretty demanding, but not as much as you'd think: that's pretty vanilla digital IC stuff.
You suggested a diagnostic circuit built into a toaster. That's going to require some circuitry, several inputs, and a microcontroller with a clock. You're not looking at 3-4 pennies here to add a self-diagnostics port to a toaster; you're looking at $3-4 dollars just to get started, before connecting up any input sources to the diagnostics module. For a $20 toaster--which, by the way, still uses an analogue timer--that's a 20% increase in the cost, where you suggested something less than 4% of the cost.
A toaster needs a resistor, a capacitor, and a transistor to control the voltage across an electromagnet, basically charging the capacitor and slowly bleeding it until it falls below the activation level for a FET and cuts power to the heating elements. This doesn't measure time, but rather is sort of relative: it shows you how "dark" it gets. High-quality toasters with RC circuits can consistently produce the same darkness at the same setting, although there may be variation from toaster to toaster.
A microwave, on the other hand, traditionally used a mechanical timer to measure time, accurately and specifically. The digital circuit adds a $4 LED display, a 60-cent beeping speaker, and around $10 of additional control circuitry. The keypad switches cost money, too, you know. That microwave control circuit isn't exactly cheap.
There is another aspect out there though, and that is at what point do we decide that it is time for the technology to stop?
Hey look, I'm rich; I can spend $50,000 on a brand new car, and in 5 years I'll buy another one.
You lot don't get that luxury. You can buy my nice, shiny Audi S5 for $12,000 when it's 10 years old and have a nice car with 70,000 miles on it. That's efficient. Your econoshitbox of today isn't on-par with a nice Audi luxury car from last decade.
So, here's the thing. You ever wonder what it costs to make a sound card? Let's go with OPL-3. You're going to need a $0.86 YMF-262 and a $1.03 YMF-512 DAC.
Well, hold on. You also need a temperature-stable clock source, so you need a crystal oscillator. You need to set the clock, so you need an RC circuit. You need to couple the clocks between the two chips, which requires more than just a wire. Your ground needs a noise filter, basically just a polypropylene pf capacitor.
In the end, just to make the YMF-262 and YMF-512 connect to each other and actually function, you need an extra $18 of components. After that, you need your interface into it from the CPU.
That "few cents per unit" is going to be a third of the cost of the toaster. Even a circuit to handle batteries costs $5-$15, depending on if you were doing two batteries without balancing or four batteries with two-bank balancing; a diagnostics system is expensive.
As for testing and replacing things in a toaster, have you tried disassembling and reassembling one? It's not overly-complex; it's also not a magical handwave.
Those are blunt white noise generators. They're annoying. Ever noticed a waterfall or a PC fan doesn't sound the same? Different noise distribution.
You can pay engineers hundreds of thousands of dollars for all this, or you can put up cube walls.
It's a lot of assumptions, too. If it's cheaper to make something that lasts 10 years and then gets replaced, then do that.
Think of it this way: humans build machines, produce electricity and energy storage, mine and recycle materials, refine things, build components, assemble products, and ship and retail them. That's a lot of human labor, and it's reflected in the price of products--yeah, it's not that the mining equipment is expensive, but that the labor to create it, maintain it, and fuel it is expensive.
Your $20 toaster comes down to a grand total of $20 of human labor. At Chinese $3/hr rates, that's around 7 hours. Troubleshooting the electrical circuit when it fails can be costly: maybe it's some internal module which is 80% of the toaster anyway, and your repair tech spends an hour swapping it out, and it costs you $30 to repair. For that matter, maybe it's just a blown MOSFET, and your repair tech spends an hour disassembling the device, and another 3 hours troubleshooting it, and then 10 minutes replacing the MOSFET itself. You might be ahead on labor, but the labor-hours are priced high.
Assembling a computer motherboard doesn't take much in terms of labor-hours. Once you've set up the assembly line, it's fairly rapid. The motherboard is extremely-complex and takes potentially days to troubleshoot, meaning there's more human labor in repairing it than in replacing it.
At a point, you've reached the break-over: it's a waste of time to repair this. Maybe you're poor and you can spend 4 hours repairing your $20 iron, at $10/hr, when you could work 4 hours for $15/hr--if only you had work. It's still a waste of time; you're just poor.
At that point, the economics of reuse have passed. You've got electronics waste, and you should send it for recycling. We readily smelt chips for silicon and gold. Copper, plastic, and aluminum are valuable.
You're going to spend $2,400 on high-quality oil changes for your engine, and another $100 on spark plugs, every 100,000 miles. You might spend $800 for an engine rebuild around 200k-300k. If you don't maintain that engine, you're spending $6,000 for a remanufactured engine swap somewhere around the 80,000-100,000 mark, if not closer to the 50,000 mark. Which is cheaper?
Our new CIO came in immediately talking about metrics-driven decisions, and also more open plan offices to increase productivity. I pointed out that studies have shown open plan reduces productivity and increases worker stress, and he brushed it off and said he doesn't subscribe to that thinking.
You can't be metrics-driven and support open plan.
Also, no, it's not introverts being afraid of social interaction; it's everyone's gum chewing, keyboard tapping, and chattering carrying loud and clear around the office. It's people seeing every little movement. It's people not having a safe space, and so any peripheral intrusion is a threat. It's people who should be working instead chattering across the open air space in the most inefficient communication possible.
Businesses spend hundreds of thousands of dollars to have sound engineers and audio professionals from Cambridge Sound Works and other high-dollar names come in and build noise-floor devices. Not just white noise generators, but soft noise generators positioned to confuse your spatial perceptions so the sound appears to come from nowhere and everywhere instead of from that box or ceiling fan over there, and thus your brain ignores it. The increase in SPL tightens the timpani and drives your brain to ignore what background noise does mix in, thus deadening your perception of quiet office sounds.
They could just put up cubicles.
Where do you come up with this shit?
When I worked for a call center, my desk was 3 feet wide. The cubical walls were tall; they extended back at least 18 inches from the desk, enclosing me in a little fabric box. A very tiny fabric box.
When I worked at Social Security and where I work now, management built open plan offices as cubicles with 40-inch walls. No reduction in space per employee.
Classical open plan offices are desks. Fields of desks. They waste a shitload of space.
AWS supplies MySQL and something called "MongoDB". They didn't rename it "AWNoSQL" or something.
Google, meanwhile, built Android on Java, and converts Java bytecode to a different thing for their own VM. Java is an open-source, GPL product--Oracle actually brands OpenJDK as Oracle Java, compiles it, packages it, and ships it.
Amazon is providing the service of configuring and maintaining a MongoDB installation. Perhaps you should pay a license fee if a company hires you to install MongoDB on one of their servers.
There have been times when parts of licenses and other contracts have been invalidated.
We can do better than a fork.
We can turn the MongoDB query language into a spec, and then release mongolite as MIT-licensed.
I'm pretty sure that's unenforceable. You'd have to make the entire OS and all of its shit AGPLv3.
I guess it's time somebody fork MongoDB from its current AGPLv3 source.
it's a particularly sad state of affairs - a reaction to unethical corporate spongeing - that a major prominent software team has to consider changing the license to a non-free one just to be able to pay their developers to keep working, whilst corporations all around them make hundreds of millions of dollars in profits, using their work??
The AGPLv3 lets you use software for any purpose. Are you saying everyone should pay for Apache or Linux?
The new license forces IaaS providers to open source everything--Amazon would have to make AWS open source.
Basically, it's a matter of wanting to release it as free software but also have it be proprietary software and so not really free. This is like when Oracle released Java under GPL and then sued Google for not paying for Java.
That's a violation of interstate commerce, because it is one state telling a company in another state how it must act, and that action impacts their business with every other state.
The States are allowed to do that for commerce within their state. If you want to operate in any State, you need a business license. You're an LLC in California, but you want to do business in Maryland? You need to file as an out-of-state LLC in Maryland. You will need to pay Maryland taxes, as well, on certain activity.
It's actually impossible to sell Internet access to someone in California without having infrastructure and contracts in California, by the way.