Now you have a fair price, and if it becomes law we can all agree on using it as a standard.
I do agree that in a healthy free market the price is dictated by dynamic agreement by buyers and sellers. However, the last mile of telecom is a natural monopoly, so there can be no free market price established in this manner.
I'm a moderate-volume user, and I wouldn't mind it if the price was fair.
If 3TB is really $50, then 1TB would probably be $10-20. 100GB would be only $1-2. The typical email+web customer might pay 50 cents per month.
Sure, throw in a $5/month fixed billing charge like the utilities do. Your ISP might sell email for another $1/month/account, and webhosting/etc for $1/month for those cheapo file-only accounts.
The key is for the pricing to be fair - wholesale cost plus a few percent profit. The utility has zero risk, so it doesn't need much reward.
Net neutrality is a total non-issue, as the last mile just delivers ethernet frames to a regionally centralized point. You buy your connection to the internet at large from a competitive company. If your current ISP starts throttling VOIP then buy your net connection from somebody else. If local offerings get out of hand just about anybody can buy a line to a tier-1 carrier and start their own regional ISP with very minimal costs (maybe a few k per month, easily scaled to the amount of bandwidth you actually sell). I doubt it would even be necessary - if you have 25 competitors do you want to be the guy who slows down Vonage or Youtube?
Yup. The real problem is charging for an unmetered service, and then trying to somehow meter it.
Look, is it "unlimited" or not? If yes, then just live with what you promised. If not, come up with something reasonable.
The last mile of telecom is a natural monopoly, and price should be PUC regulated just like your water or electricity. Does the electric company publish a list of top-10 electric consumers? Of course not - those are its best customers.
I'm fine with paying by the GB, provided those rates are reasonable. Then everybody can use whatever they want to.
Probably the best free-market solution is to have the telco/cable co own the last mile, and charge PUC-regulated rates. They only provide data service to the central office, and they cannot sell "internet" service (email, etc). Then you buy your internet service from an ISP, who runs their own bandwidth to the CO and rents rack space at a regulated price there. That is no longer a natural monopoly, which means every town in America will likely have 3-50 of them to choose from. That means you'll probably get a fair price, and get to pick whether you want usenet, email, plain old routing only, or whatever. Your local telco just transmits raw ethernet frames or something like that, so it also means that IPv6 will be available as soon as some local ISP decides to offer it.
Also - if the telco provides service over a shared line, they could meter it. However, if the telco's technology uses dedicated lines (like DSL) then they would have to offer it uncapped. Prices would of course be tied to actual costs, and investment decisions/etc would be PUC-regulated.
This isn't rocket science - we've operated utilities for years...
As I said, measures and countermeasures, cat and mouse...
Sooner or later, somebody will find a way. Then they'll come up with a harder chip to crack, and then somebody will crack that...
From a theoretical standpoint, time is on the side of the crackers I think. You can't give somebody the ability to play a game in a self-contained system that they physically possess without also giving them the ability to copy the game, manipulate it, etc. Consoles aren't magic - they're machines.
Those designs have been around for quite some time, and there are no known workaround for the current generation of those chips AFAIK.
Again, no disputing any of this. However, sooner or later a solution will be found. If you just image the surface, remove one layer of atoms, image that, and keep going, sooner or later you'll have an atomic-level 3D map of the whole thing. No law of physics makes this impossible to do - we just lack the technology right now. Or, perhaps some technique will be designed to image the interior without cutting it open (x-rays, etc).
Agreed that there is no use for moore's law for breaking the crypto itself, but that is trying to go in the front door. Depending on the algorithm there might be a quantum solution at some point. However, computation isn't needed to just read the hardware right out of the ROM/flash.
Sooner or later the PS3 hardware will be cracked, most likely within our lifetimes.
Again, my point isn't that this is easy to do - only that it is possible to do. To convince me that it is impossible you'll need to show me a fundamental law of physics that prevents it - not just a lack of current technology.
It's fairly easy these days to get a custom chip design that incorporates the CPU, secure memory, and whatnot, so that whatever you do outside of the chip won't cause any security breaches.
Sure, but at no point did I restrict attacks to staying outside of the chip. What happens if you slice it open and image it at the atomic level?
Sure, that isn't easy to do, and it isn't possible to do yet with arbitrary accuracy. However, eventually this will be possible (within the limits of the uncertainty principle, which of course is also a lower-limit on the density of storage anyway so you can't hide behind this either).
My point is that these kinds of schemes have a fundamental flaw. It is a cat and mouse game, but one which those trying to hide data on an embedded system are certain to lose. Any system you devise will only be secure for a limited period of time. Now, from a practical standpoint nobody uses consoles more than 10 years after they're introduced (at least not in a way that is market-relevant). You can certainly delay at attack on your hardware as a result. However, you cannot prevent it forever, for in giving the customer a console and disc that works offline, you've given them the keys to the kingdom as it were.
I'd say their problem is that getting a machine to run some software and not other software is essentially an impossible problem to solve, when the person trying to defeat your protection has physical access to your machine.
Sure, you can make it hard, but sooner or later somebody will come up with a workaround. If your decryption keys are stored in the hardware, they can be retrieved from the hardware. If your verification keys are stored in the hardware, they can be modified. Your code can also be modified in arbitrary ways, as can any circuitry stored in the CPU or any other component in the system.
Who makes money from piracy? The ad revenue on a warez site is pretty small potatoes. If you're talking about some major producer in asia mass-producing discs, then they probably could do that already by copying the encrypted discs at the physical layer.
When you say security through obscurity, you usually mean "nobody is going to type in 'website.com/passwords' into the server!" The way you're using it, it makes it sound like any DRM even on a closed platform is doomed.
Uh, well, it is. Fundamentally DRM is unsound unless the code runs on a server you have physical control over. Sure, you can make it difficult, as was done with the PS3, but fundamentally you cannot prevent somebody from obtaining the decryption keys and firmware from RAM and either emulating the whole platform, or re-engineering the software to work without keys. If the CPU can execute the instruction, then somebody with physical access to the CPU can monitor it. Granted, that might require logic analyzers running at full clock speed soldered to 50nm leads, but it will always be possible.
Sure, but then there would be no evidence of unrelated life forms walking around today.
Could it have happened? Sure - that would be along the lines of the earlier post mentioning stuff arising and then getting gobbled up every other Tuesday.
I was referring more to the concept that there are creatures all around us that evolved completely independently.
My understanding is that it all comes down to what they can argue in court.
If you're a lowly peon in the big scheme of things, what are the chances that your former employer will even find out you violated the agreement?
If you're a big-wig, then the risk is higher, but chance are you got a golden parachute.
From what I've read, these kinds of agreements only stick if they are limited in scope and length (the more the better), and include compensation of some kind (you can't work at a competitor for a year, but they pay you 2 years' wages when they fire you, etc). They are much easier to enforce against high-paid executives than against ordinary employees.
Yup. At my workplace I think that the general policy is to just tell people that their services are no longer required, full stop. They actually give out very decent severance anyway, so it is unlikely that anybody would sue them (at least a part of the severance is contingent on signing a wavier).
When they do layoffs the beancounters are sure to count how many are gone from any protected class imaginable, so good luck with a discrimination suit. No doubt personal grudges factor into these kinds of things, but you'd never be able to prove it.
Oh, some things can be our problem. Search for the word bio-rhythm in that transcript and be prepared to be very amused. Apparently what amounts to astrology was an important consideration in the staffing of missile silos.
Agreed. Life arising, dying out completely, and arising again I could buy - if the earth went through conditions sufficient to sterilize everything.
Life starting 50X in parallel seems hard to fathom, when half the reason we're so sure that everything evolved is all the homologies. Where are all these creatures that aren't descended from a common ancestor? It should be completely evident in their biochemistry.
Well, of course interactions with the world would need to be modeled (that is the data).
Cell-cell interactions are defined by the genome, egg cytoplasm, and the laws of physics.
Basically I'm just describing what you need to determine the initial state of your simulator. I'm not trying to suggest that you need model nothing else. That would be absurd, since otherwise all you have is a static snapshot of a fertilized egg.
Yup - I was lackluster at math and science in elementary school (sure, doing fine and I guess above average, but nothing distinctive). Fast-forward to high school and I was probably 3-4 standard deviations above the average. The main difference was that you spent first grade adding 1-digit numbers, 2nd grade adding 2-digit numbers, and so on. Actually, half of those grades were just reviewing the second half of the previous grades. Is it any wonder I didn't apply myself to it (and what was there to apply myself to - being the fastest multiplication table reciter in the school?)?
From my observations schools have only gotten worse at catering to the average - or even the below average (it is those kids who don't pass state tests that cause problems for schools). No child left behind and all that - better to focus on the 1 kid that is struggling than the 99 that are mediocre but with the potential to be better.
I have no doubt that some will still require medication, or would at least benefit from it. However, maybe if we didn't bore kids out of their minds that would work also.
An area this technology might be used in could be embedded controllers, which are not general-purpose devices.
If you're building a thrust vectoring system for a plane, and the servos have an accuracy of 1%, then it is more important to deliver more frequent servo updates than to deliver those updates with 0.0001% accuracy. If your device is attached to a sensor that has 5% manufacturing tolerances then you may not need even 8-bit precision on the math.
In the IS discrete-math world we tend to view numbers as precise figures. However, most numbers that computers deal with are actually not discrete quantities, and they can have substantial levels of error. If you can optimize a system to make it faster and cheaper and have it consume less power at the cost of a level of error that is still negligible compared to the error already in the system, then that is a good move.
However, I'm not sure this will ever make sense for GP computing, unless you can make it part of a standardized coprocoessor or something like that. If GP computing it is hard to know where error can be accepted, unless this is specified by the programmer.
GCC has -ffast-math, and I see this as something similar.
What, don't you realize that they're professionals, and they clearly must know more about raising children than your fiance?!!!
Now, the question is how many other parents just find a doctor to prescribe drugs to get the thought police off their backs. I'm sure you were threatened with neglect charges on more than one occasion.
I was sent to Catholic school at an early age, and had a lot of problems with boredom/etc. The teacher (privately) suggested public education due to the availability of ability-tailored curricula (gifted programs, tracked classes, etc). This definitely helped me, although I still struggled with poor performance until I started hitting more serious classes.
From what I've seen the trend is to treat all kids as equals, and to not put "smart" kids in separate classes (ugh, perish the thought that maybe some kids have more potential than others). Is it a wonder then that more kids can't be bothered to pay attention, when education is aimed at the mythical average (or below-average)?
A kid who exceeds test standards by 500% probably gets the school nothing. A kid who fails to meet them by 20% probably costs them all kinds of grief. Guess where the attention will get aimed? If that 500%-er turns into a 75%-er due to bad attitude the solution the school employs will be to drug them up to turn them into a 150%-er and not have to deal with them further.
Let's settle this via Slashdot. Whoever gets to score 5 first must be right...
I don't care enough to bother finding out, and I don't know that it really matters. Bottom line is that pumping kids with drugs is generally a bad thing, except in the unusual situations where it is a good idea. When half the population has ADHD it makes me think the diagnostic criteria need retooling.
I found the back-and-forth here amusing, however.:)
Oh, I agree that human brain simulation is still a ways off. I'm not sure you need to simulate every molecular orbital from first principles to model a brain, but looking only at gene sequence is obviously not enough (at least not unless we get a LOT better at the other stuff).
On the other hand, I'm not so sure that AI is necessarily decades off. You don't have to model a human brain to achieve AI. Just as the human brain is a marvelous example of emergent behavior from simpler parts, somebody may find a similar clever solution to AI. Predicting this, however, is a complete gamble at this point.
The genome of a creature, plus the cytoplasm contents of an egg, plus a complete understanding of the laws of physics should in fact be all that you need in order to fully simulate a human being. Granted, you'd need to simulate it sequentially from conception to adulthood before you get anything useful out of it, which might take more or less than the biological time required depending on the power of your simulator.
Humans are deterministic, after all - we're just a bunch of atoms and molecules. Granted, there is the effect of random quantum effects, so three simulations with the same input might not come up with the same output if this is genuinely taken into account. However, all three would be plausible outcomes if we were talking about a real person with a real brain.
The part that is being left out is the little caveat: "plus a complete understanding of the laws of physics."
Here is an illustration. A jpeg of a rendition of the Mandelbrot set might take 20k of space. A mathematical description might take well under 1kb of code. That description might even be enough to fully simulate its behavior. That description is certainly not sufficient to UNDERSTAND its behavior.
Also, don't discount the cytoplasm. Proteins don't fold the same in buffer as they do in a cell, and simply adding non-specific protein doesn't always do the trick either. Gene regulation doesn't work without epigenetics, and epigenetics doesn't happen without regulatory proteins, and those proteins don't get there without translation from gene transcripts. DNA alone without capturing the initial state of the machine is as useful as a memory dump without the CPU status dump on a CPU with 43 million registers. The last I heard things like centrioles can't be replicated except in the presence of another centriole.
The bottom line is that there is nothing "magical" about human cells. However, to estimate their total information content at only 2GB or so is probably a gross underestimate.
Uh, citation? I don't see a $5/yr option. I do see a 90-day-free certificate (which sounds like a royal pain in the neck - mainly designed to get you to spend more I guess).
Slashdot Mirror
Sure - just a matter if definition.
Fair = min(your-cost,avg(cost-of-others-doing-same-thing)*1.2)*1.03
Unfair = > Fair.
Now you have a fair price, and if it becomes law we can all agree on using it as a standard.
I do agree that in a healthy free market the price is dictated by dynamic agreement by buyers and sellers. However, the last mile of telecom is a natural monopoly, so there can be no free market price established in this manner.
Yup, which was my whole point. The model has to change so that it isn't the opposite situation.
I'm a moderate-volume user, and I wouldn't mind it if the price was fair.
If 3TB is really $50, then 1TB would probably be $10-20. 100GB would be only $1-2. The typical email+web customer might pay 50 cents per month.
Sure, throw in a $5/month fixed billing charge like the utilities do. Your ISP might sell email for another $1/month/account, and webhosting/etc for $1/month for those cheapo file-only accounts.
The key is for the pricing to be fair - wholesale cost plus a few percent profit. The utility has zero risk, so it doesn't need much reward.
Net neutrality is a total non-issue, as the last mile just delivers ethernet frames to a regionally centralized point. You buy your connection to the internet at large from a competitive company. If your current ISP starts throttling VOIP then buy your net connection from somebody else. If local offerings get out of hand just about anybody can buy a line to a tier-1 carrier and start their own regional ISP with very minimal costs (maybe a few k per month, easily scaled to the amount of bandwidth you actually sell). I doubt it would even be necessary - if you have 25 competitors do you want to be the guy who slows down Vonage or Youtube?
Yup. The real problem is charging for an unmetered service, and then trying to somehow meter it.
Look, is it "unlimited" or not? If yes, then just live with what you promised. If not, come up with something reasonable.
The last mile of telecom is a natural monopoly, and price should be PUC regulated just like your water or electricity. Does the electric company publish a list of top-10 electric consumers? Of course not - those are its best customers.
I'm fine with paying by the GB, provided those rates are reasonable. Then everybody can use whatever they want to.
Probably the best free-market solution is to have the telco/cable co own the last mile, and charge PUC-regulated rates. They only provide data service to the central office, and they cannot sell "internet" service (email, etc). Then you buy your internet service from an ISP, who runs their own bandwidth to the CO and rents rack space at a regulated price there. That is no longer a natural monopoly, which means every town in America will likely have 3-50 of them to choose from. That means you'll probably get a fair price, and get to pick whether you want usenet, email, plain old routing only, or whatever. Your local telco just transmits raw ethernet frames or something like that, so it also means that IPv6 will be available as soon as some local ISP decides to offer it.
Also - if the telco provides service over a shared line, they could meter it. However, if the telco's technology uses dedicated lines (like DSL) then they would have to offer it uncapped. Prices would of course be tied to actual costs, and investment decisions/etc would be PUC-regulated.
This isn't rocket science - we've operated utilities for years...
As I said, measures and countermeasures, cat and mouse...
Sooner or later, somebody will find a way. Then they'll come up with a harder chip to crack, and then somebody will crack that...
From a theoretical standpoint, time is on the side of the crackers I think. You can't give somebody the ability to play a game in a self-contained system that they physically possess without also giving them the ability to copy the game, manipulate it, etc. Consoles aren't magic - they're machines.
Those designs have been around for quite some time, and there are no known workaround for the current generation of those chips AFAIK.
Again, no disputing any of this. However, sooner or later a solution will be found. If you just image the surface, remove one layer of atoms, image that, and keep going, sooner or later you'll have an atomic-level 3D map of the whole thing. No law of physics makes this impossible to do - we just lack the technology right now. Or, perhaps some technique will be designed to image the interior without cutting it open (x-rays, etc).
Agreed that there is no use for moore's law for breaking the crypto itself, but that is trying to go in the front door. Depending on the algorithm there might be a quantum solution at some point. However, computation isn't needed to just read the hardware right out of the ROM/flash.
Sooner or later the PS3 hardware will be cracked, most likely within our lifetimes.
Again, my point isn't that this is easy to do - only that it is possible to do. To convince me that it is impossible you'll need to show me a fundamental law of physics that prevents it - not just a lack of current technology.
It's fairly easy these days to get a custom chip design that incorporates the CPU, secure memory, and whatnot, so that whatever you do outside of the chip won't cause any security breaches.
Sure, but at no point did I restrict attacks to staying outside of the chip. What happens if you slice it open and image it at the atomic level?
Sure, that isn't easy to do, and it isn't possible to do yet with arbitrary accuracy. However, eventually this will be possible (within the limits of the uncertainty principle, which of course is also a lower-limit on the density of storage anyway so you can't hide behind this either).
My point is that these kinds of schemes have a fundamental flaw. It is a cat and mouse game, but one which those trying to hide data on an embedded system are certain to lose. Any system you devise will only be secure for a limited period of time. Now, from a practical standpoint nobody uses consoles more than 10 years after they're introduced (at least not in a way that is market-relevant). You can certainly delay at attack on your hardware as a result. However, you cannot prevent it forever, for in giving the customer a console and disc that works offline, you've given them the keys to the kingdom as it were.
I'd say their problem is that getting a machine to run some software and not other software is essentially an impossible problem to solve, when the person trying to defeat your protection has physical access to your machine.
Sure, you can make it hard, but sooner or later somebody will come up with a workaround. If your decryption keys are stored in the hardware, they can be retrieved from the hardware. If your verification keys are stored in the hardware, they can be modified. Your code can also be modified in arbitrary ways, as can any circuitry stored in the CPU or any other component in the system.
Who makes money from piracy? The ad revenue on a warez site is pretty small potatoes. If you're talking about some major producer in asia mass-producing discs, then they probably could do that already by copying the encrypted discs at the physical layer.
When you say security through obscurity, you usually mean "nobody is going to type in 'website.com/passwords' into the server!" The way you're using it, it makes it sound like any DRM even on a closed platform is doomed.
Uh, well, it is. Fundamentally DRM is unsound unless the code runs on a server you have physical control over. Sure, you can make it difficult, as was done with the PS3, but fundamentally you cannot prevent somebody from obtaining the decryption keys and firmware from RAM and either emulating the whole platform, or re-engineering the software to work without keys. If the CPU can execute the instruction, then somebody with physical access to the CPU can monitor it. Granted, that might require logic analyzers running at full clock speed soldered to 50nm leads, but it will always be possible.
Sure, but then there would be no evidence of unrelated life forms walking around today.
Could it have happened? Sure - that would be along the lines of the earlier post mentioning stuff arising and then getting gobbled up every other Tuesday.
I was referring more to the concept that there are creatures all around us that evolved completely independently.
My understanding is that it all comes down to what they can argue in court.
If you're a lowly peon in the big scheme of things, what are the chances that your former employer will even find out you violated the agreement?
If you're a big-wig, then the risk is higher, but chance are you got a golden parachute.
From what I've read, these kinds of agreements only stick if they are limited in scope and length (the more the better), and include compensation of some kind (you can't work at a competitor for a year, but they pay you 2 years' wages when they fire you, etc). They are much easier to enforce against high-paid executives than against ordinary employees.
Yup. At my workplace I think that the general policy is to just tell people that their services are no longer required, full stop. They actually give out very decent severance anyway, so it is unlikely that anybody would sue them (at least a part of the severance is contingent on signing a wavier).
When they do layoffs the beancounters are sure to count how many are gone from any protected class imaginable, so good luck with a discrimination suit. No doubt personal grudges factor into these kinds of things, but you'd never be able to prove it.
Oh, some things can be our problem. Search for the word bio-rhythm in that transcript and be prepared to be very amused. Apparently what amounts to astrology was an important consideration in the staffing of missile silos.
Agreed. Life arising, dying out completely, and arising again I could buy - if the earth went through conditions sufficient to sterilize everything.
Life starting 50X in parallel seems hard to fathom, when half the reason we're so sure that everything evolved is all the homologies. Where are all these creatures that aren't descended from a common ancestor? It should be completely evident in their biochemistry.
Well, of course interactions with the world would need to be modeled (that is the data).
Cell-cell interactions are defined by the genome, egg cytoplasm, and the laws of physics.
Basically I'm just describing what you need to determine the initial state of your simulator. I'm not trying to suggest that you need model nothing else. That would be absurd, since otherwise all you have is a static snapshot of a fertilized egg.
Yup - I was lackluster at math and science in elementary school (sure, doing fine and I guess above average, but nothing distinctive). Fast-forward to high school and I was probably 3-4 standard deviations above the average. The main difference was that you spent first grade adding 1-digit numbers, 2nd grade adding 2-digit numbers, and so on. Actually, half of those grades were just reviewing the second half of the previous grades. Is it any wonder I didn't apply myself to it (and what was there to apply myself to - being the fastest multiplication table reciter in the school?)?
From my observations schools have only gotten worse at catering to the average - or even the below average (it is those kids who don't pass state tests that cause problems for schools). No child left behind and all that - better to focus on the 1 kid that is struggling than the 99 that are mediocre but with the potential to be better.
I have no doubt that some will still require medication, or would at least benefit from it. However, maybe if we didn't bore kids out of their minds that would work also.
I think the concept is a good one.
An area this technology might be used in could be embedded controllers, which are not general-purpose devices.
If you're building a thrust vectoring system for a plane, and the servos have an accuracy of 1%, then it is more important to deliver more frequent servo updates than to deliver those updates with 0.0001% accuracy. If your device is attached to a sensor that has 5% manufacturing tolerances then you may not need even 8-bit precision on the math.
In the IS discrete-math world we tend to view numbers as precise figures. However, most numbers that computers deal with are actually not discrete quantities, and they can have substantial levels of error. If you can optimize a system to make it faster and cheaper and have it consume less power at the cost of a level of error that is still negligible compared to the error already in the system, then that is a good move.
However, I'm not sure this will ever make sense for GP computing, unless you can make it part of a standardized coprocoessor or something like that. If GP computing it is hard to know where error can be accepted, unless this is specified by the programmer.
GCC has -ffast-math, and I see this as something similar.
What, don't you realize that they're professionals, and they clearly must know more about raising children than your fiance?!!!
Now, the question is how many other parents just find a doctor to prescribe drugs to get the thought police off their backs. I'm sure you were threatened with neglect charges on more than one occasion.
I was sent to Catholic school at an early age, and had a lot of problems with boredom/etc. The teacher (privately) suggested public education due to the availability of ability-tailored curricula (gifted programs, tracked classes, etc). This definitely helped me, although I still struggled with poor performance until I started hitting more serious classes.
From what I've seen the trend is to treat all kids as equals, and to not put "smart" kids in separate classes (ugh, perish the thought that maybe some kids have more potential than others). Is it a wonder then that more kids can't be bothered to pay attention, when education is aimed at the mythical average (or below-average)?
A kid who exceeds test standards by 500% probably gets the school nothing. A kid who fails to meet them by 20% probably costs them all kinds of grief. Guess where the attention will get aimed? If that 500%-er turns into a 75%-er due to bad attitude the solution the school employs will be to drug them up to turn them into a 150%-er and not have to deal with them further.
Let's settle this via Slashdot. Whoever gets to score 5 first must be right...
I don't care enough to bother finding out, and I don't know that it really matters. Bottom line is that pumping kids with drugs is generally a bad thing, except in the unusual situations where it is a good idea. When half the population has ADHD it makes me think the diagnostic criteria need retooling.
I found the back-and-forth here amusing, however. :)
Oh, I agree that human brain simulation is still a ways off. I'm not sure you need to simulate every molecular orbital from first principles to model a brain, but looking only at gene sequence is obviously not enough (at least not unless we get a LOT better at the other stuff).
On the other hand, I'm not so sure that AI is necessarily decades off. You don't have to model a human brain to achieve AI. Just as the human brain is a marvelous example of emergent behavior from simpler parts, somebody may find a similar clever solution to AI. Predicting this, however, is a complete gamble at this point.
I'd probably take an intermediate point of view.
The genome of a creature, plus the cytoplasm contents of an egg, plus a complete understanding of the laws of physics should in fact be all that you need in order to fully simulate a human being. Granted, you'd need to simulate it sequentially from conception to adulthood before you get anything useful out of it, which might take more or less than the biological time required depending on the power of your simulator.
Humans are deterministic, after all - we're just a bunch of atoms and molecules. Granted, there is the effect of random quantum effects, so three simulations with the same input might not come up with the same output if this is genuinely taken into account. However, all three would be plausible outcomes if we were talking about a real person with a real brain.
The part that is being left out is the little caveat: "plus a complete understanding of the laws of physics."
Here is an illustration. A jpeg of a rendition of the Mandelbrot set might take 20k of space. A mathematical description might take well under 1kb of code. That description might even be enough to fully simulate its behavior. That description is certainly not sufficient to UNDERSTAND its behavior.
Also, don't discount the cytoplasm. Proteins don't fold the same in buffer as they do in a cell, and simply adding non-specific protein doesn't always do the trick either. Gene regulation doesn't work without epigenetics, and epigenetics doesn't happen without regulatory proteins, and those proteins don't get there without translation from gene transcripts. DNA alone without capturing the initial state of the machine is as useful as a memory dump without the CPU status dump on a CPU with 43 million registers. The last I heard things like centrioles can't be replicated except in the presence of another centriole.
The bottom line is that there is nothing "magical" about human cells. However, to estimate their total information content at only 2GB or so is probably a gross underestimate.
They can look to prior precedent to determine how far exactly that is.
Sounds like a great deal for accessibility consultants and lawyers!
Pay us to help make you compliant, or pay your lawyers to try to prove your innocence. Or, most likely, do both...
Uh, citation? I don't see a $5/yr option. I do see a 90-day-free certificate (which sounds like a royal pain in the neck - mainly designed to get you to spend more I guess).