I'm curious if insect death scent freaks out mammals. If it doesn't, this might be a good repellant for homes with pets (which rules out a number of poisons that the pets might ingest).
That does require you to inhale them. If they're creating sufficiently long tubes with sufficient durability and weaving them together, it's not as much of a problem. After all, you rarely inhale your shirt.:-) That said, I have no doubt it will be less durable than promised and this will be an issue, but they won't bother with safety tests up front because there are no regulation requiring it yet.
And to be fair, in some cases the new materials have taken over. Most passenger jet designs are switching to carbon fiber bodies; the cost is high, but the lighter material means that the you need far less fuel on every trip, eventually paying for itself. (Yes, the 787 is having problems in production, but I suspect that's more a matter of poor coordination than any intrinsic weakness in the material.) And the GP ignored plastics, which relatively recently displaced all sorts of time tested materials in the construction of all manner of products. Who's to say that we won't find a way to produce carbon nanotubes cheaply in the next few years?
Of course, if technologically enabled warrantless snooping is okay, they could track which subway entrances were used to determine where you've been. In the linked case, tracking confirmed an alibi. But it could just as easily be used for fishing expeditions if not confined to the scope of a warrant based on probable cause.
Exactly. The Massachusetts decision makes sense: If you can show probable cause, you can intrude upon a person's privacy, but *only* if you show probable cause. Wisconsin decided that privacy is subordinate to police effectiveness. Problem is, you follow that track too far and you end up with a police state and no rights to speak of. The police don't *intend* to violate your rights, they simply do whatever is allowable to uphold their mandate (keeping the peace). If you don't restrict the range of allowable activities, and they can use technology to supplement their numbers, upholding their mandate most effectively requires them to scan every phone call, track every car, open all mail, etc.
Technology allows quantitative differences to become qualitative differences: Police can already tail anyone on a public street. But limited numbers mean they are only able to do so for a small number of people, so they tend to have good reasons when they do tail. But if you can track every car effortlessly and keep a database of movements, you can go on fishing expeditions. Someone dumped a body on the side of a highway? Quick, pull up the logs and find every person who passed that stretch of highway recently. Then demand DNA and fingerprint samples from all of them (assuming you haven't already collected them). It's effective, at the cost of invading everyone's privacy.
For the same reasons all OSes end up using more memory over time:
Larger memory space (Doesn't apply to NT4 vs. 32 bit XP, but it's part of the cost of XP x64 for instance) means larger pointers. Pointers are everywhere, and the cost of a pointer doubles from 16->32 and from 32->64.
Full unicode support (every part of the system upgraded to unicode doubles or quadruples the size of each string; NT4 had partial unicode support but many components didn't bother)
Graphics improvements. Moving from 16->24->32 bit color and increasing resolution from 640x480->1280x1024->1920x1200 means the resting memory cost just to maintain the current screen increases substantially (for example, a double-buffered 32 bit 1920x1200 display will cost you 17.5 MB of RAM just sitting there)
Improved security/stability. Allocating dynamic strings and doing proper bounds checking cost more than strcpy to a stack allocated byte array. Separating services and reducing shared memory means a crash or exploit of one is less likely to affect others, but it often means the two tasks are now using twice the memory. Driver layers that insulate the kernel from a failure cost additional memory too. An example: Every new release of the Server product increases the cost of a user session (for Remote Desktop and the like) because more programs are loaded in session private memory, to reduce the risk of users viewing each other's data, or a crash taking down another user's session.
Code reuse. NT4 was written largely in C; every subsequent release has introduced more C++, more COM controls, virtual machines, etc. It means less time spent writing and maintaining code, and the reuse often means the code is effectively better tested (since it is being reused in different ways by lots of components). But it also means that popping up a Yes/No dialog box loads large parts of the whole windowing framework even if the code isn't strictly necessary.
The last two are particularly nasty sources of memory bloat. And of course, there are illegitimate reasons for bloat; poor design can allocate excessive memory or leak memory, reimplement the same functionality despite the existence of a sharable code (meaning different parts of the same program might load effectively identical code at different places because one developer reinvented the wheel unnecessarily), etc.
I meant to include a note at the top of that post that it was "just in case" someone was stupid enough to try it. Slashdot dropped the first paragraph though. Weird.
Mixing bleach and ammonia creates chlorine gas. In all likelihood, if you clean with one after the other, so only a few drops of one remain to react with the other, you'll only do a small amount of damage (though repeated exposure will cause progressively worsening effects). If you actually mix them in any serious quantities, particularly in a poorly ventilated space, you will gas yourself and you will die. So please don't mix bleach and ammonia, for cleaning purposes or anything else.
That said, the reactive properties of bleach and ammonia make bleaching one of the few ways of effectively removing the smell of cat urine from clothing, towels, etc. I've got a few towels at home that could not be used if I didn't bleach them to remove the cat urine odor (repeated non-bleach washings did little to nothing). Of course, they were originally blue, and now have large patches of pink due to the bleach, but I'm not picky about the appearance of my towels.:-)
It's not funny, it's true. I've worked at MS, and while I personally tested a whole bunch of install scenarios (for a specific bundled app), upgrade always got short shrift and had the most problems. Yes, the most egregious errors were addressed, but most of the intensive testing happens on clean installs. Back up your files and install clean, unless you're really interested in finding all the corner cases.
Interplay had little or nothing to do with the "legacy". Black Isle designed and wrote the both of the games you mentioned, Interplay just published them. Yes, Interplay owned the IP rights, but it doesn't mean they had some special attachment to them that no one could replace.
If you read the couple dozen earlier posts here, it's been pointed out repeatedly that it doesn't actually render them radioactively inert, just chemically inert and insoluble. If radioactive elements don't dissolve their barrels, and aren't soluble in water, then storage becomes a much easier problem.
You do realize that no one involved in developing Fallout (or the sequels) actually thought radiation could do the things it does in the game, right? The idea was to create a world based on what scientifically ignorant people in the late 40s, 50s and early 60s thought a post-nuclear apocalypse would be like. Attacking them for failure to adhere to real world physics/biology is missing the point. Virtually everyone with even a high school science education is aware that excessive radiation exposure kills, but for fictional purposes games and comics change the rules. Lighten up a bit.
Adding to the above point: These bacteria can't magically create compounds from random elements. While the article and summary discuss it in general terms ("heavy metals"), the only metal currently processed seems to be uranium. Presumably they would want to engineer bacteria to do the same thing to other soluble heavy metals, but the bacteria aren't magic. They can't tell the difference between a nonradioactive element and its radioactive isotopes, so selectively absorbing C14 isn't going to happen.
Beyond that, all bacteria everywhere make compounds out of carbon (carbon rings are the basis of most biologically active compounds). That doesn't mean they magically absorb the carbon from your body. If, for some reason, you had uranium as part of a chemical in your body, these bacteria still wouldn't do anything with it; they convert elemental uranium to a uranium compound, they don't strip it out of other compounds.
Actually, it is. If they nest in one office, then returning to that office (from any starting location) is innate. Training helps with speed of transit, but they do it of their own accord. Of course, to make the trip again, they need to be boxed up and shipped back to the starting office.
Well, keep in mind, full screen on the OLPC isn't as big as on other machines. And if their new hardware has a GPU of some sort, some of the decoding and stretching can be offloaded.
Controlled through fear? If I may ask, what are we being controlled to do? Watch more TV news, mostly, as far as I can tell.
Which is a problem in itself. TV news isn't actually informative to a large extent. From the weeks of TV news coverage given to swine flu, the actual relevant information could easily be condensed down to half a dozen bullet points (I'll append them to the end of this post). No, in the case of the flu the motives may be benign, so it's not "controlled" in the puppet sense, but the hysteria over swine flu clogs emergency rooms (costing money, time and just for good measure, increasing the spread as the few people with swine flu give it to the hypochondriacs and patients with other ailments), causes people to horde Tamiflu (creating shortages and/or inflating the price) and leads to innumerable overreactions by public and private entities due to the hysteria of a few (e.g. closing schools for weeks).
Beyond the direct results of this sort of hysteria, there are indirect costs, namely a failure to properly evaluate risk. The money and time spent worrying about swine flu is vastly disproportionate to its danger. Inflate your tires, gets your brakes checked and don't drive drunk. You'll reduce your risk of death substantially more than wearing a face mask and stockpiling Tamiflu.
If something like SARS or influenza really did mutate and kill millions, it would not be a surprise, historically speaking, or biologically speaking. Seems to be worth letting people know about.
Well you're aware of it, right? I find it hard to believe that anyone (or at least, non-hermits in the U.S.) is *not* aware of the return of the swine flu. Until it actually does show signs of increasing mortality, why bother announcing it? Do we make a special announcement every time someone at a convention comes down with the seasonal flu? Again, disproportionate weight is being assigned to a negligible risk.
Swine flu bullet points:
Lethality appears equivalent to the regular flu
Due to prior exposure to related viruses, Baby Boomers (and earlier generations) have some resistance
Due to more powerful immune systems, teenagers and 20 somethings *may* have a slightly higher risk due to immune overreaction, but the additional risk is small
Aside from age, all the risk factors for seasonal flu apply to swine flu equally
The *only* thing that makes the swine flu measurably riskier is the lack of vaccine, a problem which is already being addressed and should be available in late September/early October.
Even without a vaccine, the flu season in the southern hemisphere was mild; barring a rapid mutation we're likely to have the same experience (but muted by the eventual availability of the vaccine)
Good lord, you can't even quote your own source correctly. It says "People who die after getting swine flu are 100 times more likely, compared to seasonal flu, to have been killed by the virus itself rather than secondary causes." Which is a rather odd way of putting it to start with (by "dying from the virus itself" they really mean "dying due to overzealous immune reaction"), and not helpful from assessing mortality rates in general. If 1000 people die of swine flu, and another 1000 die of seasonal flu with complications, it's not really that important from a lethality standpoint. If the seasonal group's secondary infections would not have occurred absent the seasonal flu, then the seasonal flu was still responsible for the death.
Secondly, the article itself notes that "Both seasonal influenza and the new A(H1N1) virus that has swept the globe since May appear to have roughly the same mortality rate of one-to-five per 1,000 infections," though it notes that the swine flu figures are sketchy.
Finally, your source is suspect to start with. It's a single scientist, and the results haven't actually been published, just posted to a research sharing website (no real vetting has occurred). In case you've forgotten, science reporting goes for sensation over accuracy.
Being small costs a lot. Netbooks are cheap because they use relatively cheap components, usually underpowered and ordered in bulk to keep costs low. And they're still large enough that they can use some commodity components; hard drives, memory, adapters, etc., can all be borrowed from larger laptops. The problem is that miniaturization costs more as you get smaller and smaller. Once you drop below the netbook size range, into the sub-7" screens, it's hard to fit everything into the package without:
Overheating/Running out of power too quickly (both are linked to hardware with poor performance per watt)
Being massively underpowered
Being expensive
Sacrificing versatility
One way or another, you'll have to compromise on the design; a general purpose phone that performs well without overheating will require expensive components. The phone can be much cheaper if it's less versatile and/or underpowered (which is how regular cell phones are kept in the $100 range). Similarly, a short battery life can keep costs down. Few phones compromise on this, but a lot of laptops and older MP3 players keep costs down this way. They require more frequent charging, allowing the use of either cheaper, more wasteful CPUs and storage or smaller/cheaper batteries.
Unlike the iPhone though, that's the whole cost. The iPhone's "real" price is buried in the mandatory contract. Outside of the U.S., you don't have a phone bundled with the plan, but the plan itself is cheaper. So yes, it costs $650 (or equivalent local currency), but the plan only runs $20-40/month, not $60-70/month. Over two years, you will have paid quite a bit more for the iPhone. And if you choose not to upgrade after two years, the savings over the iPhone accrue even faster.
It seems to indicate that playkeys would be per file. And the cost to store a key maxes out at about half a KB (for an RSA prime number based system); substantially less if it uses either a private key style encryption system or an elliptic curve based public key system. So for your files, that would be around 2.5 MB at the outside, or as little as 80 KB. If this were implemented, I'd expect a gig or two of flash memory to be included with any hardware based system, which would handle somewhere between 2 million and 62.5 million keys (depending on size of key and size of included memory). Or they maintain a separate file or partition on a hard drive, which has it's own protected key (on the hardware device), thereby eliminating the need for special purpose storage, and removing the cap on the number of files.
I suspect this is as much about resetting DRM to a real standard as it is about DPP. Since DPP would require a DRM-like system, if DPP were accepted, everyone would have a DRM capable system based on community developed standards. This doesn't make it a good idea, but it's not quite as half-assed as you think.
And anyone with a "link" to the key can assume ownership. So if you, or any of your friends' computers are compromised, they can "steal" your DPP protected stuff. And you can never get it back.
Of course, there is little reason to steal; people who want the files in question would simply get DPP-free versions. Only malicious sorts and vandals would bother, since there'd be no real gain from the act. But if you have a falling out with your friend, it doesn't look like you can "change the locks" so to speak. If I give a house key to a friend, and for some reason stop trusting him, I can change the locks on my house. This doesn't seem to support a similar mechanism. Also, unless you store the playkey online (which has its own problems), a hardware failure in the playkey storage device will cost you your files. Returning to the house analogy, it would be like your house burning down (okay, becoming inaccessible forever) because you lost the key to the front door.
According to this the standard Segway models can carry 260 pounds (that's passenger + cargo weight). The e series comes with side cargo bags for the purpose. So in my case (170 lbs), I could theoretically carry 90 lbs of groceries.
Of course, this is all a bit of a moot point for me personally. I live in Manhattan, and my grocery store is two blocks away, right next to where I get off the subway on my way home from work. So I do a couple small shopping runs a week, and only buy what I can easily carry.
Slashdot Mirror
I'm curious if insect death scent freaks out mammals. If it doesn't, this might be a good repellant for homes with pets (which rules out a number of poisons that the pets might ingest).
That does require you to inhale them. If they're creating sufficiently long tubes with sufficient durability and weaving them together, it's not as much of a problem. After all, you rarely inhale your shirt. :-) That said, I have no doubt it will be less durable than promised and this will be an issue, but they won't bother with safety tests up front because there are no regulation requiring it yet.
And to be fair, in some cases the new materials have taken over. Most passenger jet designs are switching to carbon fiber bodies; the cost is high, but the lighter material means that the you need far less fuel on every trip, eventually paying for itself. (Yes, the 787 is having problems in production, but I suspect that's more a matter of poor coordination than any intrinsic weakness in the material.) And the GP ignored plastics, which relatively recently displaced all sorts of time tested materials in the construction of all manner of products. Who's to say that we won't find a way to produce carbon nanotubes cheaply in the next few years?
Of course, if technologically enabled warrantless snooping is okay, they could track which subway entrances were used to determine where you've been. In the linked case, tracking confirmed an alibi. But it could just as easily be used for fishing expeditions if not confined to the scope of a warrant based on probable cause.
Exactly. The Massachusetts decision makes sense: If you can show probable cause, you can intrude upon a person's privacy, but *only* if you show probable cause. Wisconsin decided that privacy is subordinate to police effectiveness. Problem is, you follow that track too far and you end up with a police state and no rights to speak of. The police don't *intend* to violate your rights, they simply do whatever is allowable to uphold their mandate (keeping the peace). If you don't restrict the range of allowable activities, and they can use technology to supplement their numbers, upholding their mandate most effectively requires them to scan every phone call, track every car, open all mail, etc.
Technology allows quantitative differences to become qualitative differences: Police can already tail anyone on a public street. But limited numbers mean they are only able to do so for a small number of people, so they tend to have good reasons when they do tail. But if you can track every car effortlessly and keep a database of movements, you can go on fishing expeditions. Someone dumped a body on the side of a highway? Quick, pull up the logs and find every person who passed that stretch of highway recently. Then demand DNA and fingerprint samples from all of them (assuming you haven't already collected them). It's effective, at the cost of invading everyone's privacy.
For the same reasons all OSes end up using more memory over time:
The last two are particularly nasty sources of memory bloat. And of course, there are illegitimate reasons for bloat; poor design can allocate excessive memory or leak memory, reimplement the same functionality despite the existence of a sharable code (meaning different parts of the same program might load effectively identical code at different places because one developer reinvented the wheel unnecessarily), etc.
I meant to include a note at the top of that post that it was "just in case" someone was stupid enough to try it. Slashdot dropped the first paragraph though. Weird.
Mixing bleach and ammonia creates chlorine gas. In all likelihood, if you clean with one after the other, so only a few drops of one remain to react with the other, you'll only do a small amount of damage (though repeated exposure will cause progressively worsening effects). If you actually mix them in any serious quantities, particularly in a poorly ventilated space, you will gas yourself and you will die. So please don't mix bleach and ammonia, for cleaning purposes or anything else.
That said, the reactive properties of bleach and ammonia make bleaching one of the few ways of effectively removing the smell of cat urine from clothing, towels, etc. I've got a few towels at home that could not be used if I didn't bleach them to remove the cat urine odor (repeated non-bleach washings did little to nothing). Of course, they were originally blue, and now have large patches of pink due to the bleach, but I'm not picky about the appearance of my towels. :-)
Many modern installers cache the whole image (or at least, parts which are read more than once) from DVD to HD for precisely this reason.
It's not funny, it's true. I've worked at MS, and while I personally tested a whole bunch of install scenarios (for a specific bundled app), upgrade always got short shrift and had the most problems. Yes, the most egregious errors were addressed, but most of the intensive testing happens on clean installs. Back up your files and install clean, unless you're really interested in finding all the corner cases.
Interplay had little or nothing to do with the "legacy". Black Isle designed and wrote the both of the games you mentioned, Interplay just published them. Yes, Interplay owned the IP rights, but it doesn't mean they had some special attachment to them that no one could replace.
If you read the couple dozen earlier posts here, it's been pointed out repeatedly that it doesn't actually render them radioactively inert, just chemically inert and insoluble. If radioactive elements don't dissolve their barrels, and aren't soluble in water, then storage becomes a much easier problem.
You do realize that no one involved in developing Fallout (or the sequels) actually thought radiation could do the things it does in the game, right? The idea was to create a world based on what scientifically ignorant people in the late 40s, 50s and early 60s thought a post-nuclear apocalypse would be like. Attacking them for failure to adhere to real world physics/biology is missing the point. Virtually everyone with even a high school science education is aware that excessive radiation exposure kills, but for fictional purposes games and comics change the rules. Lighten up a bit.
Adding to the above point: These bacteria can't magically create compounds from random elements. While the article and summary discuss it in general terms ("heavy metals"), the only metal currently processed seems to be uranium. Presumably they would want to engineer bacteria to do the same thing to other soluble heavy metals, but the bacteria aren't magic. They can't tell the difference between a nonradioactive element and its radioactive isotopes, so selectively absorbing C14 isn't going to happen.
Beyond that, all bacteria everywhere make compounds out of carbon (carbon rings are the basis of most biologically active compounds). That doesn't mean they magically absorb the carbon from your body. If, for some reason, you had uranium as part of a chemical in your body, these bacteria still wouldn't do anything with it; they convert elemental uranium to a uranium compound, they don't strip it out of other compounds.
Pfffttt... Snail based networking is the way to go... But yeah, the latency characteristics are even worse.
Actually, it is. If they nest in one office, then returning to that office (from any starting location) is innate. Training helps with speed of transit, but they do it of their own accord. Of course, to make the trip again, they need to be boxed up and shipped back to the starting office.
Well, keep in mind, full screen on the OLPC isn't as big as on other machines. And if their new hardware has a GPU of some sort, some of the decoding and stretching can be offloaded.
Controlled through fear? If I may ask, what are we being controlled to do? Watch more TV news, mostly, as far as I can tell.
Which is a problem in itself. TV news isn't actually informative to a large extent. From the weeks of TV news coverage given to swine flu, the actual relevant information could easily be condensed down to half a dozen bullet points (I'll append them to the end of this post). No, in the case of the flu the motives may be benign, so it's not "controlled" in the puppet sense, but the hysteria over swine flu clogs emergency rooms (costing money, time and just for good measure, increasing the spread as the few people with swine flu give it to the hypochondriacs and patients with other ailments), causes people to horde Tamiflu (creating shortages and/or inflating the price) and leads to innumerable overreactions by public and private entities due to the hysteria of a few (e.g. closing schools for weeks).
Beyond the direct results of this sort of hysteria, there are indirect costs, namely a failure to properly evaluate risk. The money and time spent worrying about swine flu is vastly disproportionate to its danger. Inflate your tires, gets your brakes checked and don't drive drunk. You'll reduce your risk of death substantially more than wearing a face mask and stockpiling Tamiflu.
If something like SARS or influenza really did mutate and kill millions, it would not be a surprise, historically speaking, or biologically speaking. Seems to be worth letting people know about.
Well you're aware of it, right? I find it hard to believe that anyone (or at least, non-hermits in the U.S.) is *not* aware of the return of the swine flu. Until it actually does show signs of increasing mortality, why bother announcing it? Do we make a special announcement every time someone at a convention comes down with the seasonal flu? Again, disproportionate weight is being assigned to a negligible risk.
Swine flu bullet points:
Good lord, you can't even quote your own source correctly. It says "People who die after getting swine flu are 100 times more likely, compared to seasonal flu, to have been killed by the virus itself rather than secondary causes." Which is a rather odd way of putting it to start with (by "dying from the virus itself" they really mean "dying due to overzealous immune reaction"), and not helpful from assessing mortality rates in general. If 1000 people die of swine flu, and another 1000 die of seasonal flu with complications, it's not really that important from a lethality standpoint. If the seasonal group's secondary infections would not have occurred absent the seasonal flu, then the seasonal flu was still responsible for the death.
Secondly, the article itself notes that "Both seasonal influenza and the new A(H1N1) virus that has swept the globe since May appear to have roughly the same mortality rate of one-to-five per 1,000 infections," though it notes that the swine flu figures are sketchy.
Finally, your source is suspect to start with. It's a single scientist, and the results haven't actually been published, just posted to a research sharing website (no real vetting has occurred). In case you've forgotten, science reporting goes for sensation over accuracy.
Being small costs a lot. Netbooks are cheap because they use relatively cheap components, usually underpowered and ordered in bulk to keep costs low. And they're still large enough that they can use some commodity components; hard drives, memory, adapters, etc., can all be borrowed from larger laptops. The problem is that miniaturization costs more as you get smaller and smaller. Once you drop below the netbook size range, into the sub-7" screens, it's hard to fit everything into the package without:
One way or another, you'll have to compromise on the design; a general purpose phone that performs well without overheating will require expensive components. The phone can be much cheaper if it's less versatile and/or underpowered (which is how regular cell phones are kept in the $100 range). Similarly, a short battery life can keep costs down. Few phones compromise on this, but a lot of laptops and older MP3 players keep costs down this way. They require more frequent charging, allowing the use of either cheaper, more wasteful CPUs and storage or smaller/cheaper batteries.
Unlike the iPhone though, that's the whole cost. The iPhone's "real" price is buried in the mandatory contract. Outside of the U.S., you don't have a phone bundled with the plan, but the plan itself is cheaper. So yes, it costs $650 (or equivalent local currency), but the plan only runs $20-40/month, not $60-70/month. Over two years, you will have paid quite a bit more for the iPhone. And if you choose not to upgrade after two years, the savings over the iPhone accrue even faster.
It seems to indicate that playkeys would be per file. And the cost to store a key maxes out at about half a KB (for an RSA prime number based system); substantially less if it uses either a private key style encryption system or an elliptic curve based public key system. So for your files, that would be around 2.5 MB at the outside, or as little as 80 KB. If this were implemented, I'd expect a gig or two of flash memory to be included with any hardware based system, which would handle somewhere between 2 million and 62.5 million keys (depending on size of key and size of included memory). Or they maintain a separate file or partition on a hard drive, which has it's own protected key (on the hardware device), thereby eliminating the need for special purpose storage, and removing the cap on the number of files.
I suspect this is as much about resetting DRM to a real standard as it is about DPP. Since DPP would require a DRM-like system, if DPP were accepted, everyone would have a DRM capable system based on community developed standards. This doesn't make it a good idea, but it's not quite as half-assed as you think.
And anyone with a "link" to the key can assume ownership. So if you, or any of your friends' computers are compromised, they can "steal" your DPP protected stuff. And you can never get it back.
Of course, there is little reason to steal; people who want the files in question would simply get DPP-free versions. Only malicious sorts and vandals would bother, since there'd be no real gain from the act. But if you have a falling out with your friend, it doesn't look like you can "change the locks" so to speak. If I give a house key to a friend, and for some reason stop trusting him, I can change the locks on my house. This doesn't seem to support a similar mechanism. Also, unless you store the playkey online (which has its own problems), a hardware failure in the playkey storage device will cost you your files. Returning to the house analogy, it would be like your house burning down (okay, becoming inaccessible forever) because you lost the key to the front door.
Presumably, if you actually share files out, it will unblock the relevant port. If you aren't sharing anything, it *should* be blocked.
According to this the standard Segway models can carry 260 pounds (that's passenger + cargo weight). The e series comes with side cargo bags for the purpose. So in my case (170 lbs), I could theoretically carry 90 lbs of groceries.
Of course, this is all a bit of a moot point for me personally. I live in Manhattan, and my grocery store is two blocks away, right next to where I get off the subway on my way home from work. So I do a couple small shopping runs a week, and only buy what I can easily carry.