Cash is not really viable for wide range of globalized services anymore.
Get a Target prepaid card (it's a bit trickier to grab one outside of the US, but still doable). Same goes for SIM card. It is possible to use most modern phone-tied services (semi) anonymously. The idea is that it is a bit inconvenient to set up so most people don't bother. Still beats the amish "option" you're suggesting though.
This not as outlandish as it sounds with speculative quantum gravity theories. Notably, past the event horizon, time could actually run backwards. Meaning time doesn't stand asymptotically still, but past there progressively starts running backwards from our perspective. Note that this doesn't imply time travel as the effect would manifest strictly locally. It has however interesting implications regarding possible "radiation" from the black hole. All you need to overcome is the difficulty of being near event horizon and not getting stuck there, but rest of it is essentialy "free ride". From perspective of black hole inhabitant, our universe would be center of black hole in their universe.
The intuition is thus that arrow of time is related to matter and gravity. The field equations for gravity would then point arrow of time as a multi dimensional wave where each cycle into negative half cycle manifested as black holes.
The issue is with attribution. AI is just probabilistic black box in lieu of the decision made by operator - typically we're talking about the "search" function of S&D systems.
We know what it might do, but we're not sure. Worse, it's a can of worms prone to potential subversion/ECM with disastrous results, even in civil areas.
When human operator makes an error, they're court martialed or sued. But with autonomous civil overlords, tanks and warships, political pressure emerges to make people who utilize those not liable for any error the AI would ultimately make.
Banning AIs is of course stupid. What should be clearly defined instead is that whoever utilizes such a tool is clearly responsible for whatever happens, just as they would be liable for setting up traps or other deterministic device. This would essentialy curb any use of AI in critical areas, as nobody sane would shoulder that amount of responsibility for something that flimsy.
If a police officer sends an autonomous tank into gang ridden DMZ of the city and it starts shooting innocent civilians, it should be the same as if the officer pulled the trigger themselves. Great power, great responsibility.
Common decency is having a common standard of deceit. "Face" isn't just asian thing, everybody does it, with the exception of anonymous cowards (internet can be more or less treated as confession booth), and low functioning autists who simply ignore the interpersonal discord caused by their lack of tactfulness.
Common muslim does drugs in certain parts of the world, even if it's a taboo of the faith.
Common american is racist, but will deny it.
Common teenager talks dirty in front of peers, but not in front of parents.
and so on
I think the moment GAI has at least some sort of self-awareness, it will be long past the ability to construct mental models of whomever it interacts with and convincingly bluff whatever it needs to for most optimal outcome, just like most humans do.
Google has realized they need to get back their control over Android, and DRM is their golden ticket.
So what is TEE and pals?
Each CPU has a burned in public key. A publisher can encrypt binary specifically for the public key of yours, and such a blob will run only and only on a CPU with burned in private key, like sort of smartcard on steroids. This can be paired with remote attestation and what not, but it's no different from, say, CPU, acting as your SIM card (which is actually one of such use cases). In smart card industry we call this a "secure element".
While not "mandatory" now, it starts to creep in. The google safenet feature which uses this can now attest to applications whether the device supports TEE or not, and vendors slowly nudge into buying DRM capable devices through planned obsolescence.
Can this be broken? Of course, this isn't legit ZK computing with actual grounds in cryptography. It's semantically no different from a smart card, and you can dump your private key with enough effort. Needless to say, if you distribute a jailbreak consisting of emulator for a key of some CPU, the simcard will be prompty blacklisted and all "cloud" features cease to function (ie works only to fool applications which stay offline).
Why is that bad? The computer is no longer yours. Jailbreaks become pretty much infeasible and you have zero control over what's running on your device, only the CPU vendor has full control over "ring -2". A perfect walled garden, and ranchers have the perfect cattle marking stamps too.
I'd not be one bit surprised when it comes to stuff like this in politics. Speaking for my own country, the whole political system is held together through complex web of blackmail. A lobbyist won't back a politician unless there exists a perfect kompro - this selects for politicians with paraphilias of the sort to guarantee public outrage and/or jail sentence if it ever came out.
Of course occasionaly the blackmail ring is exposed accidentaly, in which case hasty cover up follows - the members did "nothing wrong", after all.
Not living or ever been to the US, I get the impression this is simply all swept under the rug with euphemism of "teen pregnancy".
In DC, 10% of all pregnancies are teenage.
A super politically incorrect fact is that fathers are almost twice as likely to be no longer in their teens. There's a marked bias for black fathers as well, while girls ethnicity is more or less uniformly distributed. What social dynamic is exactly at play there is anyone's guess.
Modern NiMH batteries keep 85% capacity over 3 years
That's capacity, but not actual single charge.
NiMh lose 50% of charge after a year. At best, there is a special type called LSD (Sanyo), which can retain about 80% C after a year.
Meanwhile, modern LSD alkaline batteries retain 80% of charge for 5 years, and some are even 60% C after 10 years. I have a kitchen LCD digital clock still running with single GP AAA manufactured in 2009.
Lithium thionyl chloride ("button cells") have even lower self-discharge, as well as other benefits (higher energy density).
However, for anything with routine high-peak-current usage (toddler, remote), where self discharge hardly is an issue, just standard cheapo NiMh truly is the way to go.
Basically, Zn-Mno2/Li-SoCl2 makes sense only for two settings:
1. Continuous, but very low current usage (eg digital clock) which is on par with self-discharge rate itself.
2. Safety kits. NiMh bottoms any useful C after about 3 years, while other chemistries can stay well above half C.
I didn't read the bill, only TFA summary, and the way I grok it is that sure, they can continue capture data all they want, but it won't longer be useful for arbitrary purpose as it is now. Meaning if the bill passes, and somebody puts scrapped data to some commercial use beyond the scope of the original service, they could be facing class action lawsuit should this come out to light.
Indeed this looks like 180 pivot into google direction, just more evil. Before the pivot, Facebook threw small scraps via it's API to attract the likes of Tinder, which boosted FB as a platform. Such apps are now facing the bait and switch which was coming all along - now tinder has to obtusely present opt-in forms to the user, or scram. While Facebook can now safely launch their own Tinder clone, with no annoying opt-ins. Gotta hand it to Zuck, the dude is the king of fucking everyone over, while making it look like he has suddenly seen the light, and cares about your privacy.
Presumably the bill doesn't cover data already farmed without consent, only further farming from now on.
It could be argued that FB has farmed as much data as possible already (since its popularity is more or less shrinking now). Zuck's move is "I got mine, now let's make sure nobody else gets hands on it".
Reminder that this discussion isn't about privacy, but straight competition between data brokers. Massive, and accurate human behavior corpuses, of which FB is one of the largest repository will be monetized in machine learning models soon enough.
I also wonder if google search will become pay service now, or what?
This. Driving criminal elements underground by banning the easy conduits might not be productive - it's basically an act of polishing a turd. Cops don't bother picking up peons slinging on the street, they use those to walk up the supply chain and nip the source in the bud, not just shallow appearance of it.
A sensible law would not seek to punish sex ad sites for being that. Instead, it would give more power to law enforcement to monitor these sites (request user info & IM dumps without court order, whatever). Carrot and stick.
My favorite anecdote about this is a subreddit dedicated to illegal bitcoin markets. A few years ago, a user in there casually posted that Silk Road server exposes phpinfo() - revealing real IP address and hostname.
A year later, Ross Ulbright was sentenced for lifetime for that single mistake.
This wouldn't be possible with moralizing laws which simply bans all public visibility of such endeavors.
Yes, "qubits" is a bit akin to "MHz race" when it comes quantum computing.
You can tell because d-wave now has 2000 "quibits" machines. In particular, quantum supremacy means you can run an actual algorithm with super-positioned program states (quantum logic, tiffoli gates), not just a fixed equation with superpositioned quibit registers.
Only certain linear matrix algebra benefits from fast annealing (of note, gradient descent in neural networks).
But other than that, vast majority of number theoretic problems can't be translated to a single formula for which we can solve the global minima on d-wave and call it a day.
> the encryption on the chips themselves have been proven easy to crack for many years
Bullshit.
However you can't just "crack" the signature on the card if the reader actually does verify the signature. This is because there is no private key for PA on the card, thus classic key extraction attacks are useless. You can still clone the card and use somebody's else identity, but the encryption as such is fine (as long RSA-1024 is fine, which it barely is).
Yes, simple GPU NN filters are routinely used for upscaling and deinterlacing (fe. nnedi3 in madvr) these are coarse, but work in real time.
Filling missing parts of image is vastly more difficult problem though, and is very unlikely to work in real time - so it is useless for deinterlacing or upscaling.
Indeed, not an objectively verifiable truth, given that google doesn't share details of it's antispam measures.
Personally, I believe poisoning autocomplete is on par to poisoning the index itself in difficulty. A circumstancial evidence can be found by consulting blackhat seo forums - they mostly hire actual people to do the searches, it's really not something to be easily manipulated. Google antispam is thought to work by making a "bot or not" profile for each user profile it tracks.
As far I can tell, it's not possible to inject phrases at all via anonymous search.
Instances of google bombing work by simply crowdsourcing such credentials (think typical 4chan raids). The issue is that there is way too many of "outrageous" search phrases in the index, and only tiny fraction is traceable back to actual public troll operation.
Now, 'hillary...' phrases and such are entirely different story, and it's entirely plausible there was a secret campaign to poison the index. But for generally "outrageous" stuff, such an attractor is missing.
So, Google has no obligation to prevent its services from being leveraged by those who understand how it works from exploiting
Google bombing is exception, not a rule. It simply doesn't happen on this scale.
There are no secret neonazi meetings for this.
The truth is that people simply do search for these things - "are jews nice?" and "are jews evil?", the latter ends up being more frequent as a result of people encountering nazi propaganda elsewhere and following it up by questioning it via google.
Interestingly, whenever a moral outrage happens over some tardy phrase like that (last time, it was "how to have sex with kids"), the media shitstorm actually amplifies the phrase, in ultra-refined Streissand fashion and only then we can talk about the phrases being artificially rated up.
IMO, the main practical advantage is that checking for MP is fast, so they're easier to search for. Working with MP number fields is fast too (owing to close-to-pow2 property). so they find use whenever any large prime would do, it's just a nice to have MP in there.
A typical practical usage is PRNGs of girgantual periods (the period is typically the MP itself, or multiple of thereof) for HPC number crunching. The perfect number property indeed is a nice bonus in there, as it often leads to better k-distribution of the permutation.
State of the art machine intelligence is "machine intuition". Bots copy and remix the corpus matched to context, but we're still miles away from an observable thread of thought. It will just mimic and more or less randomly remix the corpus - basically replay thoughts of somebody else. In a crude sense, it's just very sophisticated markov model, to the point it will reasonably past turing test on a youtube comment level
This is not limited to text, a fashion or food instagram is even more trivial with current tools. Good example of this is Spiderman Elsa (which I suspect is made with good old honest-to-god sweatshop labor, not a bot), but the model of social spam has shown an immense profitability potential already in a format far more sophisticated than appealing to lowest sexual urges.
The good thing about this is that this will spell an early end to shallow internet memetics once advertising world discovers chatbots and context-aware media remix bots. No more need to bribe lowkey ecelebs to astroturf your product, when you can just unleash fake users in number. Even if the quality on average will be sub-par, statistically some will always get a traction if you spawn population large enough.
It's a post-scarcity scenario for internet drivel in a cost model where people engaged in drivel for social bond and validation, both points being moot when it's a machine on the other end.
This can lead to two possible outcomes:
1. The cancer spreads, remember the south park episode about living ads? This is it. People will literally lose grasp on reality and will feel about adverts as if they were people..
2. It's a chemo which will bring us back to 1993. Folks will recognize low effort posts lost all of its shreds of utility for validation, pushing the bar for social network posts a lot higher (low effort posts being implicitly assumed a bot when it becomes a common case).
In either case, there will be constant market pressure for "better ads" as users adapt, there will be this arms race for ever better "living ad" until the bots start having so much grasp of context we'll enter a very weak GAI era.
> Rounding error compared to what?
Google 'bunker oil' and 'container ship'.
> Yes, as a fraction of total pollution these are small, but we do better by reducing pollution from all sources in general
No, we reduce Co2 by going after the few and worst offenders, like coal powerplants and car emissions, not 1% in the long trail. CO2 producers follow a Pareto pattern, so better use sorting algorithm optimized for it.
>One of the reasons that SpaceX avoided using solid rocket boosters is because they are terrible from a pollution perspective
Proton-M flies on Lox+kerosene, too. Russians didn't switch from hydrazine in 2012 because they give a crap about environment, but simply because current technology favors that combo in performance/cost/complexity, especially if you add reusability/serviceability as your business objective.
Truly "green rocket" would be lox+hydrogen, which is something phenomenally stupid to do for first stage (remember space shuttle?).
Documented Daly City earthquakes of 1906 (~7.8mag) and 1957 (~5.7mag) are comparable to what at least one japanese metropolis gets every few years, yes.
2011 tsunami was indeed different - the epicenter was 90 miles from the shore, of "megathrust" geologic class, equivalent of which in bay area would be https://en.wikipedia.org/wiki/.... When this happens, the biggest problem is obviously the tsunami, the earthquake as such is fine (as long you build for it). The trouble of when this got twisted and politicized is that SA alone is weak and manageable by modern standards and ironically, the old buildings is what are most dangerous and would get trashed first and foremost. While if Cascadia triggers, I suppose one would have to adopt retarded building codes all way up to oregon to "stop" the tsunami wave.
If anything, the fault line argument makes old buildings a hazard, not modern skyscrapers. Modern office buildings are technically subject to much bigger forces from the lateral wind (150Mph+). Earthquakes (even 9.0) are comparably puny to that (although buildings are specially tailored in highly seismic areas). Think places like Japan.
Not enough bandwidth in the overcrowded free spectrum (tragedy of the commons emerges in urban areas with healthy WISP market). Wifi great to cover rural areas tho (both long haul and last mile).
Also stop calling it mesh, its a buzzword at this point. Just WISP.
The problem US wired broadband is facing now is pains most of civilized world went through 10-15 years ago. Worst part is that FCC stance is just a spin on correct market shape - the desired endgame result. But they skimp over how you get there, which is always a disaster. The skewed market wont fix itself just by suddenly deregulating (the arguably already shitty regulation).
US needs to do what rest of the world did - deregulate slowly, and only as a "reward" for improving competetiveness of market - first, force ISPs to open the door, on federal level, demand they sell local loops and coax docsis channels at fixed prices (simply by copying market prices from europe), allow competition to slowly build their last mile infra, and don't listen to retarded arguments like "theres no need for parallel last mile". There is shitton of need - incubent telco copper is ancient and neigh useless and no, they wont install more if there's no competitor. The more last mile (glass) bandwidth, the merrier. Less aggregation and oversubscription for the end user.
The issue with zerotier is that its code is "preconfigured" to use their servers (ie its hardcoded everywhere), and they advertise (spam?) very aggresively (last year on HN). So instead of repeating it, here goes what others said:
Naive ideas like zerotier depend on central "tracker" nodes, not the torrent kind, but more like DNS. Sure, you can run DNS alt roots, but nobody will use those, because DNS isn't federated, DNS authority is a hiearchy.
People should know better than DNS these days. Networks like cjdns and tinc can achieve same effect like zerotier, with far less "need" for central ownership of the network.
Slashdot Mirror
Sorry, but I'm a cash strapped millennial, not a boomer flush with pension money - in my area taxis are significantly more expensive option.
Cash is not really viable for wide range of globalized services anymore.
Get a Target prepaid card (it's a bit trickier to grab one outside of the US, but still doable). Same goes for SIM card. It is possible to use most modern phone-tied services (semi) anonymously. The idea is that it is a bit inconvenient to set up so most people don't bother. Still beats the amish "option" you're suggesting though.
This not as outlandish as it sounds with speculative quantum gravity theories. Notably, past the event horizon, time could actually run backwards. Meaning time doesn't stand asymptotically still, but past there progressively starts running backwards from our perspective. Note that this doesn't imply time travel as the effect would manifest strictly locally. It has however interesting implications regarding possible "radiation" from the black hole. All you need to overcome is the difficulty of being near event horizon and not getting stuck there, but rest of it is essentialy "free ride". From perspective of black hole inhabitant, our universe would be center of black hole in their universe.
The intuition is thus that arrow of time is related to matter and gravity. The field equations for gravity would then point arrow of time as a multi dimensional wave where each cycle into negative half cycle manifested as black holes.
The issue is with attribution. AI is just probabilistic black box in lieu of the decision made by operator - typically we're talking about the "search" function of S&D systems.
We know what it might do, but we're not sure. Worse, it's a can of worms prone to potential subversion/ECM with disastrous results, even in civil areas.
When human operator makes an error, they're court martialed or sued. But with autonomous civil overlords, tanks and warships, political pressure emerges to make people who utilize those not liable for any error the AI would ultimately make.
Banning AIs is of course stupid. What should be clearly defined instead is that whoever utilizes such a tool is clearly responsible for whatever happens, just as they would be liable for setting up traps or other deterministic device. This would essentialy curb any use of AI in critical areas, as nobody sane would shoulder that amount of responsibility for something that flimsy.
If a police officer sends an autonomous tank into gang ridden DMZ of the city and it starts shooting innocent civilians, it should be the same as if the officer pulled the trigger themselves. Great power, great responsibility.
Common decency is having a common standard of deceit. "Face" isn't just asian thing, everybody does it, with the exception of anonymous cowards (internet can be more or less treated as confession booth), and low functioning autists who simply ignore the interpersonal discord caused by their lack of tactfulness.
Common muslim does drugs in certain parts of the world, even if it's a taboo of the faith.
Common american is racist, but will deny it.
Common teenager talks dirty in front of peers, but not in front of parents.
and so on
I think the moment GAI has at least some sort of self-awareness, it will be long past the ability to construct mental models of whomever it interacts with and convincingly bluff whatever it needs to for most optimal outcome, just like most humans do.
Google has realized they need to get back their control over Android, and DRM is their golden ticket. So what is TEE and pals?
Each CPU has a burned in public key. A publisher can encrypt binary specifically for the public key of yours, and such a blob will run only and only on a CPU with burned in private key, like sort of smartcard on steroids. This can be paired with remote attestation and what not, but it's no different from, say, CPU, acting as your SIM card (which is actually one of such use cases). In smart card industry we call this a "secure element".
While not "mandatory" now, it starts to creep in. The google safenet feature which uses this can now attest to applications whether the device supports TEE or not, and vendors slowly nudge into buying DRM capable devices through planned obsolescence.
Can this be broken? Of course, this isn't legit ZK computing with actual grounds in cryptography. It's semantically no different from a smart card, and you can dump your private key with enough effort. Needless to say, if you distribute a jailbreak consisting of emulator for a key of some CPU, the simcard will be prompty blacklisted and all "cloud" features cease to function (ie works only to fool applications which stay offline).
Why is that bad? The computer is no longer yours. Jailbreaks become pretty much infeasible and you have zero control over what's running on your device, only the CPU vendor has full control over "ring -2". A perfect walled garden, and ranchers have the perfect cattle marking stamps too.
I'd not be one bit surprised when it comes to stuff like this in politics. Speaking for my own country, the whole political system is held together through complex web of blackmail. A lobbyist won't back a politician unless there exists a perfect kompro - this selects for politicians with paraphilias of the sort to guarantee public outrage and/or jail sentence if it ever came out.
Of course occasionaly the blackmail ring is exposed accidentaly, in which case hasty cover up follows - the members did "nothing wrong", after all.
> If Rotterham had happened in the US
Not living or ever been to the US, I get the impression this is simply all swept under the rug with euphemism of "teen pregnancy".
In DC, 10% of all pregnancies are teenage.
A super politically incorrect fact is that fathers are almost twice as likely to be no longer in their teens. There's a marked bias for black fathers as well, while girls ethnicity is more or less uniformly distributed. What social dynamic is exactly at play there is anyone's guess.
Modern NiMH batteries keep 85% capacity over 3 years
That's capacity, but not actual single charge.
NiMh lose 50% of charge after a year. At best, there is a special type called LSD (Sanyo), which can retain about 80% C after a year.
Meanwhile, modern LSD alkaline batteries retain 80% of charge for 5 years, and some are even 60% C after 10 years. I have a kitchen LCD digital clock still running with single GP AAA manufactured in 2009.
Lithium thionyl chloride ("button cells") have even lower self-discharge, as well as other benefits (higher energy density).
However, for anything with routine high-peak-current usage (toddler, remote), where self discharge hardly is an issue, just standard cheapo NiMh truly is the way to go.
Basically, Zn-Mno2/Li-SoCl2 makes sense only for two settings:
1. Continuous, but very low current usage (eg digital clock) which is on par with self-discharge rate itself.
2. Safety kits. NiMh bottoms any useful C after about 3 years, while other chemistries can stay well above half C.
> Collection of data will continue same as ever.
I didn't read the bill, only TFA summary, and the way I grok it is that sure, they can continue capture data all they want, but it won't longer be useful for arbitrary purpose as it is now. Meaning if the bill passes, and somebody puts scrapped data to some commercial use beyond the scope of the original service, they could be facing class action lawsuit should this come out to light.
Indeed this looks like 180 pivot into google direction, just more evil. Before the pivot, Facebook threw small scraps via it's API to attract the likes of Tinder, which boosted FB as a platform. Such apps are now facing the bait and switch which was coming all along - now tinder has to obtusely present opt-in forms to the user, or scram. While Facebook can now safely launch their own Tinder clone, with no annoying opt-ins. Gotta hand it to Zuck, the dude is the king of fucking everyone over, while making it look like he has suddenly seen the light, and cares about your privacy.
Presumably the bill doesn't cover data already farmed without consent, only further farming from now on.
It could be argued that FB has farmed as much data as possible already (since its popularity is more or less shrinking now). Zuck's move is "I got mine, now let's make sure nobody else gets hands on it".
Reminder that this discussion isn't about privacy, but straight competition between data brokers. Massive, and accurate human behavior corpuses, of which FB is one of the largest repository will be monetized in machine learning models soon enough.
I also wonder if google search will become pay service now, or what?
This. Driving criminal elements underground by banning the easy conduits might not be productive - it's basically an act of polishing a turd. Cops don't bother picking up peons slinging on the street, they use those to walk up the supply chain and nip the source in the bud, not just shallow appearance of it.
A sensible law would not seek to punish sex ad sites for being that. Instead, it would give more power to law enforcement to monitor these sites (request user info & IM dumps without court order, whatever). Carrot and stick.
My favorite anecdote about this is a subreddit dedicated to illegal bitcoin markets. A few years ago, a user in there casually posted that Silk Road server exposes phpinfo() - revealing real IP address and hostname.
A year later, Ross Ulbright was sentenced for lifetime for that single mistake.
This wouldn't be possible with moralizing laws which simply bans all public visibility of such endeavors.
Yes, "qubits" is a bit akin to "MHz race" when it comes quantum computing.
You can tell because d-wave now has 2000 "quibits" machines. In particular, quantum supremacy means you can run an actual algorithm with super-positioned program states (quantum logic, tiffoli gates), not just a fixed equation with superpositioned quibit registers.
What d-wave does is quantum annealing - it has one "hardcoded", specific algorithm it can run.
Only certain linear matrix algebra benefits from fast annealing (of note, gradient descent in neural networks).
But other than that, vast majority of number theoretic problems can't be translated to a single formula for which we can solve the global minima on d-wave and call it a day.
> the encryption on the chips themselves have been proven easy to crack for many years
Bullshit.
However you can't just "crack" the signature on the card if the reader actually does verify the signature. This is because there is no private key for PA on the card, thus classic key extraction attacks are useless. You can still clone the card and use somebody's else identity, but the encryption as such is fine (as long RSA-1024 is fine, which it barely is).
Yes, simple GPU NN filters are routinely used for upscaling and deinterlacing (fe. nnedi3 in madvr) these are coarse, but work in real time.
Filling missing parts of image is vastly more difficult problem though, and is very unlikely to work in real time - so it is useless for deinterlacing or upscaling.
Indeed, not an objectively verifiable truth, given that google doesn't share details of it's antispam measures.
Personally, I believe poisoning autocomplete is on par to poisoning the index itself in difficulty. A circumstancial evidence can be found by consulting blackhat seo forums - they mostly hire actual people to do the searches, it's really not something to be easily manipulated. Google antispam is thought to work by making a "bot or not" profile for each user profile it tracks.
As far I can tell, it's not possible to inject phrases at all via anonymous search.
Instances of google bombing work by simply crowdsourcing such credentials (think typical 4chan raids). The issue is that there is way too many of "outrageous" search phrases in the index, and only tiny fraction is traceable back to actual public troll operation.
Now, 'hillary...' phrases and such are entirely different story, and it's entirely plausible there was a secret campaign to poison the index. But for generally "outrageous" stuff, such an attractor is missing.
Google bombing is exception, not a rule. It simply doesn't happen on this scale.
There are no secret neonazi meetings for this.
The truth is that people simply do search for these things - "are jews nice?" and "are jews evil?", the latter ends up being more frequent as a result of people encountering nazi propaganda elsewhere and following it up by questioning it via google.
Interestingly, whenever a moral outrage happens over some tardy phrase like that (last time, it was "how to have sex with kids"), the media shitstorm actually amplifies the phrase, in ultra-refined Streissand fashion and only then we can talk about the phrases being artificially rated up.
IMO, the main practical advantage is that checking for MP is fast, so they're easier to search for. Working with MP number fields is fast too (owing to close-to-pow2 property). so they find use whenever any large prime would do, it's just a nice to have MP in there.
A typical practical usage is PRNGs of girgantual periods (the period is typically the MP itself, or multiple of thereof) for HPC number crunching. The perfect number property indeed is a nice bonus in there, as it often leads to better k-distribution of the permutation.
http://maths-people.anu.edu.au...
State of the art machine intelligence is "machine intuition". Bots copy and remix the corpus matched to context, but we're still miles away from an observable thread of thought. It will just mimic and more or less randomly remix the corpus - basically replay thoughts of somebody else. In a crude sense, it's just very sophisticated markov model, to the point it will reasonably past turing test on a youtube comment level
This is not limited to text, a fashion or food instagram is even more trivial with current tools. Good example of this is Spiderman Elsa (which I suspect is made with good old honest-to-god sweatshop labor, not a bot), but the model of social spam has shown an immense profitability potential already in a format far more sophisticated than appealing to lowest sexual urges.
The good thing about this is that this will spell an early end to shallow internet memetics once advertising world discovers chatbots and context-aware media remix bots. No more need to bribe lowkey ecelebs to astroturf your product, when you can just unleash fake users in number. Even if the quality on average will be sub-par, statistically some will always get a traction if you spawn population large enough.
It's a post-scarcity scenario for internet drivel in a cost model where people engaged in drivel for social bond and validation, both points being moot when it's a machine on the other end.
This can lead to two possible outcomes:
1. The cancer spreads, remember the south park episode about living ads? This is it. People will literally lose grasp on reality and will feel about adverts as if they were people..
2. It's a chemo which will bring us back to 1993. Folks will recognize low effort posts lost all of its shreds of utility for validation, pushing the bar for social network posts a lot higher (low effort posts being implicitly assumed a bot when it becomes a common case).
In either case, there will be constant market pressure for "better ads" as users adapt, there will be this arms race for ever better "living ad" until the bots start having so much grasp of context we'll enter a very weak GAI era.
> Rounding error compared to what?
Google 'bunker oil' and 'container ship'.
> Yes, as a fraction of total pollution these are small, but we do better by reducing pollution from all sources in general
No, we reduce Co2 by going after the few and worst offenders, like coal powerplants and car emissions, not 1% in the long trail. CO2 producers follow a Pareto pattern, so better use sorting algorithm optimized for it.
>One of the reasons that SpaceX avoided using solid rocket boosters is because they are terrible from a pollution perspective
Proton-M flies on Lox+kerosene, too. Russians didn't switch from hydrazine in 2012 because they give a crap about environment, but simply because current technology favors that combo in performance/cost/complexity, especially if you add reusability/serviceability as your business objective.
Truly "green rocket" would be lox+hydrogen, which is something phenomenally stupid to do for first stage (remember space shuttle?).
Documented Daly City earthquakes of 1906 (~7.8mag) and 1957 (~5.7mag) are comparable to what at least one japanese metropolis gets every few years, yes.
2011 tsunami was indeed different - the epicenter was 90 miles from the shore, of "megathrust" geologic class, equivalent of which in bay area would be https://en.wikipedia.org/wiki/.... When this happens, the biggest problem is obviously the tsunami, the earthquake as such is fine (as long you build for it). The trouble of when this got twisted and politicized is that SA alone is weak and manageable by modern standards and ironically, the old buildings is what are most dangerous and would get trashed first and foremost. While if Cascadia triggers, I suppose one would have to adopt retarded building codes all way up to oregon to "stop" the tsunami wave.
If anything, the fault line argument makes old buildings a hazard, not modern skyscrapers. Modern office buildings are technically subject to much bigger forces from the lateral wind (150Mph+). Earthquakes (even 9.0) are comparably puny to that (although buildings are specially tailored in highly seismic areas). Think places like Japan.
Not even merit. Equality of opportunity.
Not enough bandwidth in the overcrowded free spectrum (tragedy of the commons emerges in urban areas with healthy WISP market). Wifi great to cover rural areas tho (both long haul and last mile).
Also stop calling it mesh, its a buzzword at this point. Just WISP.
The problem US wired broadband is facing now is pains most of civilized world went through 10-15 years ago. Worst part is that FCC stance is just a spin on correct market shape - the desired endgame result. But they skimp over how you get there, which is always a disaster. The skewed market wont fix itself just by suddenly deregulating (the arguably already shitty regulation).
US needs to do what rest of the world did - deregulate slowly, and only as a "reward" for improving competetiveness of market - first, force ISPs to open the door, on federal level, demand they sell local loops and coax docsis channels at fixed prices (simply by copying market prices from europe), allow competition to slowly build their last mile infra, and don't listen to retarded arguments like "theres no need for parallel last mile". There is shitton of need - incubent telco copper is ancient and neigh useless and no, they wont install more if there's no competitor. The more last mile (glass) bandwidth, the merrier. Less aggregation and oversubscription for the end user.
The issue with zerotier is that its code is "preconfigured" to use their servers (ie its hardcoded everywhere), and they advertise (spam?) very aggresively (last year on HN). So instead of repeating it, here goes what others said:
https://news.ycombinator.com/i...
Naive ideas like zerotier depend on central "tracker" nodes, not the torrent kind, but more like DNS. Sure, you can run DNS alt roots, but nobody will use those, because DNS isn't federated, DNS authority is a hiearchy.
People should know better than DNS these days. Networks like cjdns and tinc can achieve same effect like zerotier, with far less "need" for central ownership of the network.