There is a very clear divide between progressives and establishment democrats at the municipal level. I live in Seattle, which isn't one of the cities you list but has similar problems of officially being controlled by the "liberal" party but the municipal policy effectively greatly favoring current land-owners over renters (according to this site, 46% of the population, but likely non-citizens are overrepresented as Seattle has a lot of immigrants), homeless, and future residents.
Because Washington state has top-two primaries (instead of Democrat and Republican party primaries), this divide is very visible in Seattle politics, especially in our mayoral race last week where the primary had the eventual winner establishment candidate Jenny Durkan with 28% of the vote and the two leading progressive candidates each with 17% of the vote (and another with 12% of the vote; if only we had ranked choice primaries...). One of the main issues was that Durkan wanted to zone for less new housing and slower. And she won in part because home owners think that increases their property values. But "increased property values" is bad for anyone who wants to live in the area who does not presently own a home.
If you want to see progressive housing policy, look to Seattle Transit Blog calling for upzoning near any major transit route. Multiple people in the comments put forth arguments for eliminating zoning limitation on residential construction entirely. These policies are not even within the Overton window of political discourse at the level of campaigns for Seattle city positions.
Moravec's paradox is the discovery by artificial intelligence and robotics researchers that, contrary to traditional assumptions, high-level reasoning requires very little computation, but low-level sensorimotor skills require enormous computational resources.
This paper gives an interesting summary of different assumptions about how detailed a brain simulation needs to be and what they mean for when simulating a brain would be feasible (assuming Moore's Law continues indefinitely, which is obviously not guaranteed). The classical estimates go as late as 2201 depending on what assumptions you accept. See the tables on pages 79-81 for the summary. The quantum estimate is just a question mark; they didn't even bother computing the cost of using classical computers to simulate an entire human brain as a quantum system.
A genuinely interesting paper would have specific ideas for architecture capable of solving problems beyond the scope of current CPUs and GPUs.
A couple cool projects I've seen on making good use of dark silicon are GreenDroid and Chlorophyll, both of which are recent research projects on compiling for weird architectures that are specially designed to be energy efficient. If it's specialized for different applications that you want, then Anton is the closest I've seen; it's specialized for running physical simulations so it can do things like protein folding.
I agree that statistics require careful interpretation, and any claims made based on statistics require critical thought to determine if they are using statistics properly. There is, after all, another study someone else referenced in this thread that concluded the exact opposite by studying the proportion (not absolute number) of harassing tweets sent to a set of 65 celebrity Twitter accounts. I am unsure what your discussion of domestic violence statistics adds to this discussion other than giving an example of another emotional charged area where statistics are complicated.
The link I reference had two parts: (1) the actual study that showed the fake users created by the researchers which were identical except for gendered names resulted in 25 times as much harassment directed at the bots with females names and (2) the references to previous studies showing that reports of harassment are much more common from women. The assertion made by the article is that (1) supports (2). Note that in (1) the same researchers are the ones coding messages as harassing or not, so there is no separate subjectivity of men vs. women on what messages constitute harassment. That's why I cited it: it answers the question of "Do women get harassed equal to men and are just more vocal about it?" In fact, the data suggests women are less vocal about it.
Lastly, as I stated above, this whole argument is a tangent to the real issue which is "Trolling" or "Abuse". While you go pull more made up numbers to back your tangent, nothing gets done to resolve the real issue.
No disagreement there. The article is about abuse/harassment, not trolling, so the title is misleading.
I really wish we could just drop the sexism part of this right now. Both genders get attacked by these people.
Both do, but it *is* sexist. It is far more widespread and vicious towards women. Ignoring that is not helping.
Really now? Proof by assertion is not proof, it's an informal fallacy. Yes, even if you claim that not believing your assertion is "not helping". Prove that women get trolled more, and prove that the trolling is more vicious as you claim. I await your great study of everyone trolled on the Internet with eagerness.
According to a University of Maryland study, online users who appear female are 25 times more likely to receive threats and sexually explicit messages than online users with male names.
and
The disproportionate targeting of women accords with statistics compiled by the organization Working to Halt Online Abuse (WHOA). In 2007, 61 percent of the individuals reporting online abuse to WHOA were female while 21 percent were male. 2006 followed a similar pattern: 70 percent of those reporting online harassment identified themselves as women. Overall, in the years covering 2000 to 2007, 72.5 percent of the 2,285 individuals reporting cyber harassment were female and 22 percent were male.
As obnoxious as that is, you can disable that "feature" in the settings. That said, I mostly use Firefox because it does a much better job of respecting my privacy both by default and with more plugins for it.
As much as I like the idea of decentralized protocols, the problem with decentralizing chat is that most of the nodes we are talking about are mobile devices and decentralized protocols tend to require a lot more communication---and therefore battery power---than centralized protocols where you leave the organization to the servers. Any decentralized protocol would probably have to handle that by somehow offloading the extra communication and computation to devices that are currently plugged in.
There is the additional problem that authenticating users in a decentralized fashion means that the is no equivalent to password recovery, but users might be okay with an account tied to their physical phone.
Wikipedia's article on Mozilla Persona (which links to "How BrowserID differs from OpenID") clarifies that. While the site you are authenticating to gets the same information it would get via OpenID, the authentication provider doesn't know what sites you are using. Due to the indirection of storing the cryptographic credentials in the browser, the OpenID provider doesn't need to be contacted for every login and therefore doesn't know what sites you are logging into.
This is related to the design of Persona being browser-based instead of web-based, which also provides additional security (harder to fake a password entry box if it's normally generated by the browser).
The 47% figure is rather misleading because it only refers to the percentage of people paying federal income tax. It turns out there are multiple federal taxes on income, only one of which is called the federal income tax. Most of those 47% pay the payroll tax which is a regressive income tax. For detailed numbers see this chart which Google image search found on this CNN Money video. For those that don't want to click the link, the breakdown according to CNN is 53.6% pay income tax and the rest not paying income tax are split up as 28.6% pay payroll tax, 10.3% elderly with no income tax, 6.9% non-elderly with income under $20,000, and 1% other.
Once you eliminate people paying income/payroll tax and the retired elderly, that leaves at most 8% not retired but not making enough money to owe federal taxes. Some of those are unable to work. Some of those are unable to find a job. Some small proportion might really be lazy and leeching off the system like you are worried about... but that is almost certainly much less than 8% of the population and definitely a lot less than 47% of the population.
On top of that, remember this entire discussion is only about federal taxes. There are also state taxes, which are pretty universally regressive. Particularly, most states have a sales tax which hits the poor much harder as anyone earning so little they aren't paying income tax is probably buying necessities with all the money they do earn and therefore immediately paying sales tax on a large proportion of their income.
This article that I came across while searching for those figures tells a similar story with more exposition and citations.
I agree. Computer users should not need to know the inner details of how everything works on their computer in order to use it. Also, this goes back to my sig: having (effectively) a single video sharing website on the internet is bad because it can unilaterally do things like this.
I'm actually not clear on exactly how copyright works for the source code of closed-source products. My understanding was that the source is never published, so it is never copyrighted. Instead, I would expect it to be protected as a trade secret. That said, copyright of the binaries of a published product should be the same as the copyright on the source code, but presently there is no incentive for Microsoft to release the source code when their copyright expires (as that won't be for a very long time, they will have likely lost the source or simply no longer exist by the time that happens anyway).
The siblings have covered a lot of the issues with your suggestion. Wikipedia's page on philosophy of copyright might be informative as well. Other common arguments include the tragedy of the anticommons (having rightsholders for everything means doing anything new requires negotiating with too many different rightsholders) and the general fact that essentially all creative works build on prior creative works in some way, either direct retellings like many of Disney's movies or more indirectly like many fantasy books have elves that look a lot like those in Tolkien's Middle Earth.
There is also the complication that IP covers a lot of different things. Particularly I think there is a difference between artistic works like a novel and utilitarian works like Windows (and that there is not necessary a clear line between the two), but they are both covered by copyright under the exact same terms. Having copyright act differently for different works sounds messy and should probably be avoided in order to keep the law sensible, but both types of work have to be considered when arguing for how copyright should work.
For artistic works, the idea is that any published work is part of the collective culture and anyone should be able to build on it... with the exception that the author should have a limited monopoly on it in order to make money off of it. By having that time get too long, you get absurdities like the copyright status of the song "Happy Birthday to You" where the song has become part of American culture.
For utilitarian works, I think the argument might be closer to patents: the government wants to give some protection to new inventions in order to ensure a profit motive for developing them, but other companies should have access to old inventions in order to build on them. This doesn't quite work with software because there is no requirement tor release source code in order to get copyright on software. Of course, binaries alone can be useful and with effort can be modified to some extent if necessary.
The correct time-frame for both of those arguments is subjective and may be different, so the number that appears in copyright law should be a compromise between the two.
I am not sure if this is exactly what you want, but you could get pretty close by setting your foe modifier to -6 (assuming you never browse at -1, of course) and marking those users as foes (and perhaps removing them from your foes list later if you think they might write something you would want to read in a later discussion). You can set modifiers here, which I got to by clicking on the score of a comment and clicking the edit link next to a modifier. Also, if you do want to block any comments from the sibling poster, you can set your anonymous modifier to -6.
Corn means maize in the US and some other places, but elsewhere it is a generic term for cereal crops. Perhaps the parent (or the author of the parent's source) is from one of those other countries.
I wish they had built a big RESET button into the US Government. I would be pushing the SHIT out of it right now.
There is one. It's called a convention to propose amendments to the United States Constitution. If two-thirds (34) of the state legislatures call for a constitutional convention, then Congress is obligated to arrange for a national convention during which any arbitrary amendments to the constitution can be proposed (the legality of the "arbitrary" part is not entirely clear due to the lack of precedence). They do not go into effect unless ratified by 3/4th of the states, though. See also: Second Constitution of the United States.
I am curious about your experiences with the Nokia N9 as I was planning on buying one in part due to liking the Nokia N810 and being able to use it as a pocket-sized Linux computer (I never got an N900). I had thought it was pretty easy to get access to a root terminal and do whatever on it just like the N810. Is this not the case?
O_o How did I manage to double post? I did get a "resource not valid" error the first time I tried to post, but I reloaded the thread and my post wasn't there...
This is certainly not a new idea. It is sometimes referred to as the "rapture of the nerds" version of a technological singularity. Ray Kurzweil is a big fan of the idea and one of the major proponents.
As to the actual feasibility, I ran across Whole Brain Emulation: A Roadmap a little while ago, which discusses the possibility given our current knowledge of how the brain works. It provides dates on how long Moore's Law would have to continue based on varyingly optimistic assumptions about how much work is necessary to actually emulate a brain.
Overall, I think there are two main problems with expecting immortality via brain uploading: (1) 40+ years is a very long time to assume Moore's Law for and (2) even if we can emulate a human brain, scanning an existing one and transferring it into a computer may not be possible.
This is certainly not a new idea. It is sometimes referred to as the "rapture of the nerds" version of a technological singularity. Ray Kurzweil is a big fan of the idea and one of the major proponents.
As to the actual feasibility, I ran across Whole Brain Emulation: A Roadmap a little while ago, which discusses the possibility given our current knowledge of how the brain works. It provides dates on how long Moore's Law would have to continue based on varyingly optimistic assumptions about how much work is necessary to actually emulate a brain.
Overall, I think there are two main problems with expecting immortality via brain uploading: (1) 40+ years is a very long time to assume Moore's Law for and (2) even if we can emulate a human brain, scanning an existing one and transferring it into a computer may not be possible.
The GP is referring to the Physical (or "strong") Church–Turing thesis which says that all physical processes (including, say, any computation done by the human brain) are Turing-computable. I do not know if Turing or Church actually suggested that version or if only later computer scientists came up with it. It cannot actually be proven without a much better understanding of physics, but it is generally believed to be true.
But what exactly does this get me over SSL Client Certificates?
Less importantly, e-mail verification: the third party is providing a federated e-mail verification service, which Mozilla hopes is a service which will be done by the e-mail provider but is also providing themselves (as well as allowing any other third-party to offer).
More importantly, by taking the [very common] assumption that control of the e-mail address for an account is equivalent to control of an account, this appears to essentially give the decision of which public keys are tied to an account to whoever controls the e-mail address. That means that having multiple devices with different keys is easy, and, more importantly, losing all of your private keys is not a problem as the public keys can be changed as long as you can still log into your e-mail. Of course, the downside to this is that, as far as I can tell, your e-mail provider can now log into any of your accounts without resetting the password. In fact, I am not seeing why this would not give Mozilla (or any other trusted third-party) the ability to log into any account supporting this. (Of course, to be fair, an OpenID provider has the same power and this has the additional advantage that the provider does not need to be told which websites the user is logging into.)
Using SSL Client Certificates, either each host you use would have to have the same certificate or each service you use would have to know about every public key you use. Or, I guess, you could give the service a public key used to sign the keys you do use, but then you would still have the problem of needing to use e-mail verification to recover if you lost your keys.
Slashdot Mirror
There is a very clear divide between progressives and establishment democrats at the municipal level. I live in Seattle, which isn't one of the cities you list but has similar problems of officially being controlled by the "liberal" party but the municipal policy effectively greatly favoring current land-owners over renters (according to this site, 46% of the population, but likely non-citizens are overrepresented as Seattle has a lot of immigrants), homeless, and future residents.
Because Washington state has top-two primaries (instead of Democrat and Republican party primaries), this divide is very visible in Seattle politics, especially in our mayoral race last week where the primary had the eventual winner establishment candidate Jenny Durkan with 28% of the vote and the two leading progressive candidates each with 17% of the vote (and another with 12% of the vote; if only we had ranked choice primaries...). One of the main issues was that Durkan wanted to zone for less new housing and slower. And she won in part because home owners think that increases their property values. But "increased property values" is bad for anyone who wants to live in the area who does not presently own a home.
If you want to see progressive housing policy, look to Seattle Transit Blog calling for upzoning near any major transit route. Multiple people in the comments put forth arguments for eliminating zoning limitation on residential construction entirely. These policies are not even within the Overton window of political discourse at the level of campaigns for Seattle city positions.
It's called Moravec's paradox:
Moravec's paradox is the discovery by artificial intelligence and robotics researchers that, contrary to traditional assumptions, high-level reasoning requires very little computation, but low-level sensorimotor skills require enormous computational resources.
This paper gives an interesting summary of different assumptions about how detailed a brain simulation needs to be and what they mean for when simulating a brain would be feasible (assuming Moore's Law continues indefinitely, which is obviously not guaranteed). The classical estimates go as late as 2201 depending on what assumptions you accept. See the tables on pages 79-81 for the summary. The quantum estimate is just a question mark; they didn't even bother computing the cost of using classical computers to simulate an entire human brain as a quantum system.
What about Drizzt?
A genuinely interesting paper would have specific ideas for architecture capable of solving problems beyond the scope of current CPUs and GPUs.
A couple cool projects I've seen on making good use of dark silicon are GreenDroid and Chlorophyll, both of which are recent research projects on compiling for weird architectures that are specially designed to be energy efficient. If it's specialized for different applications that you want, then Anton is the closest I've seen; it's specialized for running physical simulations so it can do things like protein folding.
I agree that statistics require careful interpretation, and any claims made based on statistics require critical thought to determine if they are using statistics properly. There is, after all, another study someone else referenced in this thread that concluded the exact opposite by studying the proportion (not absolute number) of harassing tweets sent to a set of 65 celebrity Twitter accounts. I am unsure what your discussion of domestic violence statistics adds to this discussion other than giving an example of another emotional charged area where statistics are complicated.
The link I reference had two parts: (1) the actual study that showed the fake users created by the researchers which were identical except for gendered names resulted in 25 times as much harassment directed at the bots with females names and (2) the references to previous studies showing that reports of harassment are much more common from women. The assertion made by the article is that (1) supports (2). Note that in (1) the same researchers are the ones coding messages as harassing or not, so there is no separate subjectivity of men vs. women on what messages constitute harassment. That's why I cited it: it answers the question of "Do women get harassed equal to men and are just more vocal about it?" In fact, the data suggests women are less vocal about it.
Lastly, as I stated above, this whole argument is a tangent to the real issue which is "Trolling" or "Abuse". While you go pull more made up numbers to back your tangent, nothing gets done to resolve the real issue.
No disagreement there. The article is about abuse/harassment, not trolling, so the title is misleading.
I really wish we could just drop the sexism part of this right now. Both genders get attacked by these people.
Both do, but it *is* sexist. It is far more widespread and vicious towards women. Ignoring that is not helping.
Really now? Proof by assertion is not proof, it's an informal fallacy. Yes, even if you claim that not believing your assertion is "not helping". Prove that women get trolled more, and prove that the trolling is more vicious as you claim. I await your great study of everyone trolled on the Internet with eagerness.
While a complete census of internet trolling has not been conducted, it turns out there are statistics and they do support the GP: here's a study showing women get harassed at a much higher rate than men.
According to a University of Maryland study, online users who appear female are 25 times more likely to receive threats and sexually explicit messages than online users with male names.
and
The disproportionate targeting of women accords with statistics compiled by the organization Working to Halt Online Abuse (WHOA). In 2007, 61 percent of the individuals reporting online abuse to WHOA were female while 21 percent were male. 2006 followed a similar pattern: 70 percent of those reporting online harassment identified themselves as women. Overall, in the years covering 2000 to 2007, 72.5 percent of the 2,285 individuals reporting cyber harassment were female and 22 percent were male.
As obnoxious as that is, you can disable that "feature" in the settings. That said, I mostly use Firefox because it does a much better job of respecting my privacy both by default and with more plugins for it.
Gah, replying to undo accidental negative moderation.
As much as I like the idea of decentralized protocols, the problem with decentralizing chat is that most of the nodes we are talking about are mobile devices and decentralized protocols tend to require a lot more communication---and therefore battery power---than centralized protocols where you leave the organization to the servers. Any decentralized protocol would probably have to handle that by somehow offloading the extra communication and computation to devices that are currently plugged in.
There is the additional problem that authenticating users in a decentralized fashion means that the is no equivalent to password recovery, but users might be okay with an account tied to their physical phone.
Wikipedia's article on Mozilla Persona (which links to "How BrowserID differs from OpenID") clarifies that. While the site you are authenticating to gets the same information it would get via OpenID, the authentication provider doesn't know what sites you are using. Due to the indirection of storing the cryptographic credentials in the browser, the OpenID provider doesn't need to be contacted for every login and therefore doesn't know what sites you are logging into.
This is related to the design of Persona being browser-based instead of web-based, which also provides additional security (harder to fake a password entry box if it's normally generated by the browser).
The 47% figure is rather misleading because it only refers to the percentage of people paying federal income tax. It turns out there are multiple federal taxes on income, only one of which is called the federal income tax. Most of those 47% pay the payroll tax which is a regressive income tax. For detailed numbers see this chart which Google image search found on this CNN Money video. For those that don't want to click the link, the breakdown according to CNN is 53.6% pay income tax and the rest not paying income tax are split up as 28.6% pay payroll tax, 10.3% elderly with no income tax, 6.9% non-elderly with income under $20,000, and 1% other.
Once you eliminate people paying income/payroll tax and the retired elderly, that leaves at most 8% not retired but not making enough money to owe federal taxes. Some of those are unable to work. Some of those are unable to find a job. Some small proportion might really be lazy and leeching off the system like you are worried about... but that is almost certainly much less than 8% of the population and definitely a lot less than 47% of the population.
On top of that, remember this entire discussion is only about federal taxes. There are also state taxes, which are pretty universally regressive. Particularly, most states have a sales tax which hits the poor much harder as anyone earning so little they aren't paying income tax is probably buying necessities with all the money they do earn and therefore immediately paying sales tax on a large proportion of their income.
This article that I came across while searching for those figures tells a similar story with more exposition and citations.
Posting to remove accidental negative moderation.
I agree. Computer users should not need to know the inner details of how everything works on their computer in order to use it. Also, this goes back to my sig: having (effectively) a single video sharing website on the internet is bad because it can unilaterally do things like this.
I'm actually not clear on exactly how copyright works for the source code of closed-source products. My understanding was that the source is never published, so it is never copyrighted. Instead, I would expect it to be protected as a trade secret. That said, copyright of the binaries of a published product should be the same as the copyright on the source code, but presently there is no incentive for Microsoft to release the source code when their copyright expires (as that won't be for a very long time, they will have likely lost the source or simply no longer exist by the time that happens anyway).
The siblings have covered a lot of the issues with your suggestion. Wikipedia's page on philosophy of copyright might be informative as well. Other common arguments include the tragedy of the anticommons (having rightsholders for everything means doing anything new requires negotiating with too many different rightsholders) and the general fact that essentially all creative works build on prior creative works in some way, either direct retellings like many of Disney's movies or more indirectly like many fantasy books have elves that look a lot like those in Tolkien's Middle Earth.
There is also the complication that IP covers a lot of different things. Particularly I think there is a difference between artistic works like a novel and utilitarian works like Windows (and that there is not necessary a clear line between the two), but they are both covered by copyright under the exact same terms. Having copyright act differently for different works sounds messy and should probably be avoided in order to keep the law sensible, but both types of work have to be considered when arguing for how copyright should work.
For artistic works, the idea is that any published work is part of the collective culture and anyone should be able to build on it... with the exception that the author should have a limited monopoly on it in order to make money off of it. By having that time get too long, you get absurdities like the copyright status of the song "Happy Birthday to You" where the song has become part of American culture.
For utilitarian works, I think the argument might be closer to patents: the government wants to give some protection to new inventions in order to ensure a profit motive for developing them, but other companies should have access to old inventions in order to build on them. This doesn't quite work with software because there is no requirement tor release source code in order to get copyright on software. Of course, binaries alone can be useful and with effort can be modified to some extent if necessary.
The correct time-frame for both of those arguments is subjective and may be different, so the number that appears in copyright law should be a compromise between the two.
I am not sure if this is exactly what you want, but you could get pretty close by setting your foe modifier to -6 (assuming you never browse at -1, of course) and marking those users as foes (and perhaps removing them from your foes list later if you think they might write something you would want to read in a later discussion). You can set modifiers here, which I got to by clicking on the score of a comment and clicking the edit link next to a modifier. Also, if you do want to block any comments from the sibling poster, you can set your anonymous modifier to -6.
Corn means maize in the US and some other places, but elsewhere it is a generic term for cereal crops. Perhaps the parent (or the author of the parent's source) is from one of those other countries.
I wish they had built a big RESET button into the US Government. I would be pushing the SHIT out of it right now.
There is one. It's called a convention to propose amendments to the United States Constitution. If two-thirds (34) of the state legislatures call for a constitutional convention, then Congress is obligated to arrange for a national convention during which any arbitrary amendments to the constitution can be proposed (the legality of the "arbitrary" part is not entirely clear due to the lack of precedence). They do not go into effect unless ratified by 3/4th of the states, though. See also: Second Constitution of the United States.
I am curious about your experiences with the Nokia N9 as I was planning on buying one in part due to liking the Nokia N810 and being able to use it as a pocket-sized Linux computer (I never got an N900). I had thought it was pretty easy to get access to a root terminal and do whatever on it just like the N810. Is this not the case?
O_o How did I manage to double post? I did get a "resource not valid" error the first time I tried to post, but I reloaded the thread and my post wasn't there...
This is certainly not a new idea. It is sometimes referred to as the "rapture of the nerds" version of a technological singularity. Ray Kurzweil is a big fan of the idea and one of the major proponents.
As to the actual feasibility, I ran across Whole Brain Emulation: A Roadmap a little while ago, which discusses the possibility given our current knowledge of how the brain works. It provides dates on how long Moore's Law would have to continue based on varyingly optimistic assumptions about how much work is necessary to actually emulate a brain.
Overall, I think there are two main problems with expecting immortality via brain uploading: (1) 40+ years is a very long time to assume Moore's Law for and (2) even if we can emulate a human brain, scanning an existing one and transferring it into a computer may not be possible.
This is certainly not a new idea. It is sometimes referred to as the "rapture of the nerds" version of a technological singularity. Ray Kurzweil is a big fan of the idea and one of the major proponents.
As to the actual feasibility, I ran across Whole Brain Emulation: A Roadmap a little while ago, which discusses the possibility given our current knowledge of how the brain works. It provides dates on how long Moore's Law would have to continue based on varyingly optimistic assumptions about how much work is necessary to actually emulate a brain.
Overall, I think there are two main problems with expecting immortality via brain uploading: (1) 40+ years is a very long time to assume Moore's Law for and (2) even if we can emulate a human brain, scanning an existing one and transferring it into a computer may not be possible.
The GP is referring to the Physical (or "strong") Church–Turing thesis which says that all physical processes (including, say, any computation done by the human brain) are Turing-computable. I do not know if Turing or Church actually suggested that version or if only later computer scientists came up with it. It cannot actually be proven without a much better understanding of physics, but it is generally believed to be true.
But what exactly does this get me over SSL Client Certificates?
Less importantly, e-mail verification: the third party is providing a federated e-mail verification service, which Mozilla hopes is a service which will be done by the e-mail provider but is also providing themselves (as well as allowing any other third-party to offer).
More importantly, by taking the [very common] assumption that control of the e-mail address for an account is equivalent to control of an account, this appears to essentially give the decision of which public keys are tied to an account to whoever controls the e-mail address. That means that having multiple devices with different keys is easy, and, more importantly, losing all of your private keys is not a problem as the public keys can be changed as long as you can still log into your e-mail. Of course, the downside to this is that, as far as I can tell, your e-mail provider can now log into any of your accounts without resetting the password. In fact, I am not seeing why this would not give Mozilla (or any other trusted third-party) the ability to log into any account supporting this. (Of course, to be fair, an OpenID provider has the same power and this has the additional advantage that the provider does not need to be told which websites the user is logging into.)
Using SSL Client Certificates, either each host you use would have to have the same certificate or each service you use would have to know about every public key you use. Or, I guess, you could give the service a public key used to sign the keys you do use, but then you would still have the problem of needing to use e-mail verification to recover if you lost your keys.
You are thinking of the story Manna by Marshall Brain.