Nuclear warfare may be a variable with two possible outcomes but the sun burning out will happen, therefore it's always on the agenda. Maybe on the long, long term agenda, but it's definitely on there.
Besides, the world's hardly awash with well known celebrity scientists. It's all very well that this view is strongly held in scientific circles, but the more recognised faces that come out and say this, the more it has a chance (however slim) of making it into the mainstream.
It's not so much getting off this rock as spreading the risk - nobody's saying it would be easy or fun to live on another planet (I guess plenty of people would still jump at the chance though), but at least then when we're all eaten by the elder gods or whatever we'll have banked some of our species elsewhere to carry on the line.
The chances of us evolving naturally even once are astronomical - to think you could reliably package up a "build your own human kit" into some kind of terrorforming probe (it would have to make some adjustments to the planet it hit unless you're happy with a huge failure rate) is hard to believe, and the only alternative (that humans are somehow the natural end-state for all evolution) is even harder to believe. Alternatively, if you assume they weren't aiming for "humans" but merely seeding life on other planets and humans were just an incidental event, I guess that's much more do-able... relatively!
What this will all come down to is profit. If a big enough corporation (or group thereof) thinks it can make a profit in space, it'll do so. Unlike humans, corporations don't have to worry about lifespans, they can take the bigger picture, but the payoff would need to be incredible for them to make the investments required (which is why it's still vital that governments support research spending on space projects, the governments might never shell out the money for a space mining facility, but they might just uncover the motivation for someone else to do so).
That's quite an odd assertion - I haven't the first clue how my car works but I drive safely, have never been the cause of an accident in 12 years of driving, am always courteous to other road users, etc. I don't need to understand how a car works to not want to kill myself or others or wreck my primary means of transportation - not being in an accident is all the motivation I need thankyouverymuch. And within my immediate circle of family and friends, I don't know anyone else who either knows much beyond the basics about the car or who would happily die in a screaming fireball on the motorway by driving like a maniac (and even if they did, I don't see how the two are in any way related - most of the people I've known in the past who were really into the mechanics side of things were also into the speed side of things, and speed and safety hardly go hand in glove either, but I still wouldn't claim all mechanics are bad drivers).
More recently, the early adopters for mobile phones seemed much more interested in how they worked and in what hacks and workarounds they could employ, now most people just take the devices for granted - they use them all the time and they're an essential part of life, but they don't particular care about the inner workings. It's likely more to do with the fact that early adopters are generally more techy people who do have a deeper interest in how things work, then those things reach a tipping point where they're cheap, easy and convenient enough for the masses and the noise from the average user drowns out the techy user. It doesn't mean there aren't just as many techy people in this new generation getting involved in the inner workings, they're just a hell of a lot harder to spot in the wild.
The argument that a tool is just a tool only works if you know nothing about the person wielding it. A gun is just a gun becomes a different argument if said gun is being held by someone we know is a convicted killer. Now Google's modus operandi is to use sat and street-level imagery to build up services like Google maps, so we can reasonably expect they are investigating the merit of using drones to add to that map data. Even without IR and night vision, that's enough to cause some people privacy concerns and that alone is worthy of further discussion. Of course the article makes the possibilities sound worse because that's how you sell clicks on the internet these days, but that doesn't mean the most likely purpose of the tool isn't also worth at least talking about.
Not sure about the minimal chance of hitting anyone part given that they're likely to be flying over the most populated areas - but maybe some kind of parachute that auto-deploys when the engine dies would minimise danger?
Even if every car had these systems it wouldn't guarantee safety - there are circumstances where a computer could probably do a reasonable or even a better job than a human brain, but there are other times when human insight is the better tool for the job. Imagine a 40MPH road with a bunch of kids on a school outing walking down the path, liable to run out at any time, or animals, horses, etc - a human would know instinctively that the safe thing would be to slow down (not all of them would of course, that's a different matter). The computer would happily plow on at 40MPH and that had better be some pretty damn impressive collision detection (both on the lead car and any following it) - and let's hope there's no ice or oil on the road, either...
Roadsign detection and speed governing would mean max speed = roadway speed limit. Smacking a properly-designed modern car into an immovable object at any legal roadway speed is generally not fatal.
Collision avoidance can decrease the risk but it'll never account for every possibility including all weather conditions, road conditions, loose debris, animals, people not in safety cars (cyclists, motorbikes, pedestrians). And I'm not sure that hitting an immovable object (concrete bridge support for instance) at 70MPH (the legal maximum here in the UK) is likely to be non-fatal - and even if by some miracle the driver survives, you've suddenly dumped a lot of dangerous debris all over the motorway that everyone else is now hitting at 70MPH. That's not accounting for the fact that migration to the new system would take a long time so you'll have a mix of safety and normal cars on the road (and even when the vast majority have migrated there will be some exceptions - vintage cars for instance are unlikely to get these features added). If we could magically switch all cars to these "fatality free" versions simultaneously tomorrow, I'd still feel confident giving odds of seeing the first fatality within a week.
Can you imagine a tiered internet that made those sites so painfully slow to use that people would be willing to give up their privacy to get around it?
There are enough high profile miscarriages of justice that if someone won't accept that a person with nothing to hide has nothing to fear then they're either burying their head in the sand or they really mean "but it's so much more convenient" - people seem very happy to give up privacy for convenience right up to the point where they realise they've given away too much. You can't ever fight that, it's human nature - sure certain defining moments in time may shock people into action, but they'll always return to the former position because it's too easy not to.
Number 6: Welcome friend, I'm number 6.
Number 15: I'm number 15. What number are you?
Homer: I am not a number! I am a man! And don't you ever... oh wait, I'm number 5. Ha, ha! In your face, number 6!
Exactly, I'm shocked by the standard of geekery I'm seeing here. I guess gone are the days where, if a/.er was identified from a photo taken with "friends", Google would assume he was Klingon.
I think the point is more that, once a computer can simulate the best of human thinking, it will be more productive to have the computer think about these issues. The computer can process the data faster and doesn't need food or sleep, and it's far easier to duplicate a computer process multiple times than to create more genius level humans (not to mention far quicker, none of that messy growing up and puberty stuff to deal with, just load the disk image and go).
That depends - you say you're assuming there's an equal chance of spotting a known and an unknown word. It's more likely that his software is better at recognising the word that has been generated by an algorithm rather than some randomly obscured word (that was then run through the algorithm), so while the success rate for two known words might not be 30% there's a good chance that it would be better than 1%, since that figure is factoring in the ability to recognise a word that even Google's OCR couldn't realiably recognise.
You can make a reasonable guess as a human - the unknown word is usually unknown for a reason (archaic type rendering two letters together, an ink smudge on one letter, etc) which a human can spot with reaonsable accuracy once they know the trick. If they introduced some of these tricks on the known words they might make it more difficult (and of course, just because a human can spot the difference reliably, doesn't mean it's easy to do in software, as supported by the results in the video).
It's a little confusing but the presentation seems to suggest if you get one of the words wrong too many times in succession - 32 it claims - it will log your IP and force you to get both right (I'm not sure how that works, maybe it uses a word with a large number of identical responses, or maybe it just generates two words it knows). It doesn't say if this is reset after you have successfully got both right, if it was you could still brute force it (the counter would reset after an average 100 entries if there's a 1% chance of matching both, in which case for every 132 entries you'd still get about a quarter of them through which is not a bad hit rate), if not the suggestion is "dynamic IP" - not particularly useful if you plan to spam thousands of these things.
Yes, centralised captcha would seem eventually destined to failure. The better approach would be that each instance be reasonably unique (even if it was only so far as each users uploading their own images), that way a determined spammer might break one site, but he can't use what he did there to attack others, each time he'll have to start from scratch. The obvious downside is the massive cost - instead of one person spending a week linking images in the database, you'd need one person per company. Still, for a reasonably large company for whom spam is a big issue it might eventually still become the best option (and to periodically add to and remove from the database of existing images).
Recaptcha allows you to make some mistakes as standard than most othert captcha solutions I think (and TFA's findings seem to support this also - it suggests you can get one word wrong and one letter from the other word, although when I tested that it was too much, but I have successfully tested one right and one wrong word and still passed the captcha). Really you're being served one captcha word and one word Google's book scanning project couldn't recognise, you're solving the captcha word but the other word (usually the harder word to read) you're only adding to the statistical weighting of what the word probably is, in other words you can afford to be a little wrong and still get the captcha right.
Slashdot Mirror
Nuclear warfare may be a variable with two possible outcomes but the sun burning out will happen, therefore it's always on the agenda. Maybe on the long, long term agenda, but it's definitely on there.
Besides, the world's hardly awash with well known celebrity scientists. It's all very well that this view is strongly held in scientific circles, but the more recognised faces that come out and say this, the more it has a chance (however slim) of making it into the mainstream.
It's not so much getting off this rock as spreading the risk - nobody's saying it would be easy or fun to live on another planet (I guess plenty of people would still jump at the chance though), but at least then when we're all eaten by the elder gods or whatever we'll have banked some of our species elsewhere to carry on the line.
The chances of us evolving naturally even once are astronomical - to think you could reliably package up a "build your own human kit" into some kind of terrorforming probe (it would have to make some adjustments to the planet it hit unless you're happy with a huge failure rate) is hard to believe, and the only alternative (that humans are somehow the natural end-state for all evolution) is even harder to believe. Alternatively, if you assume they weren't aiming for "humans" but merely seeding life on other planets and humans were just an incidental event, I guess that's much more do-able... relatively!
What this will all come down to is profit. If a big enough corporation (or group thereof) thinks it can make a profit in space, it'll do so. Unlike humans, corporations don't have to worry about lifespans, they can take the bigger picture, but the payoff would need to be incredible for them to make the investments required (which is why it's still vital that governments support research spending on space projects, the governments might never shell out the money for a space mining facility, but they might just uncover the motivation for someone else to do so).
That's quite an odd assertion - I haven't the first clue how my car works but I drive safely, have never been the cause of an accident in 12 years of driving, am always courteous to other road users, etc. I don't need to understand how a car works to not want to kill myself or others or wreck my primary means of transportation - not being in an accident is all the motivation I need thankyouverymuch. And within my immediate circle of family and friends, I don't know anyone else who either knows much beyond the basics about the car or who would happily die in a screaming fireball on the motorway by driving like a maniac (and even if they did, I don't see how the two are in any way related - most of the people I've known in the past who were really into the mechanics side of things were also into the speed side of things, and speed and safety hardly go hand in glove either, but I still wouldn't claim all mechanics are bad drivers).
For god's sake don't ever buy a dog.
More recently, the early adopters for mobile phones seemed much more interested in how they worked and in what hacks and workarounds they could employ, now most people just take the devices for granted - they use them all the time and they're an essential part of life, but they don't particular care about the inner workings. It's likely more to do with the fact that early adopters are generally more techy people who do have a deeper interest in how things work, then those things reach a tipping point where they're cheap, easy and convenient enough for the masses and the noise from the average user drowns out the techy user. It doesn't mean there aren't just as many techy people in this new generation getting involved in the inner workings, they're just a hell of a lot harder to spot in the wild.
And could the sharks have fricken lasers on their heads?
The argument that a tool is just a tool only works if you know nothing about the person wielding it. A gun is just a gun becomes a different argument if said gun is being held by someone we know is a convicted killer. Now Google's modus operandi is to use sat and street-level imagery to build up services like Google maps, so we can reasonably expect they are investigating the merit of using drones to add to that map data. Even without IR and night vision, that's enough to cause some people privacy concerns and that alone is worthy of further discussion. Of course the article makes the possibilities sound worse because that's how you sell clicks on the internet these days, but that doesn't mean the most likely purpose of the tool isn't also worth at least talking about.
Not sure about the minimal chance of hitting anyone part given that they're likely to be flying over the most populated areas - but maybe some kind of parachute that auto-deploys when the engine dies would minimise danger?
Even if every car had these systems it wouldn't guarantee safety - there are circumstances where a computer could probably do a reasonable or even a better job than a human brain, but there are other times when human insight is the better tool for the job. Imagine a 40MPH road with a bunch of kids on a school outing walking down the path, liable to run out at any time, or animals, horses, etc - a human would know instinctively that the safe thing would be to slow down (not all of them would of course, that's a different matter). The computer would happily plow on at 40MPH and that had better be some pretty damn impressive collision detection (both on the lead car and any following it) - and let's hope there's no ice or oil on the road, either...
Roadsign detection and speed governing would mean max speed = roadway speed limit. Smacking a properly-designed modern car into an immovable object at any legal roadway speed is generally not fatal.
Collision avoidance can decrease the risk but it'll never account for every possibility including all weather conditions, road conditions, loose debris, animals, people not in safety cars (cyclists, motorbikes, pedestrians). And I'm not sure that hitting an immovable object (concrete bridge support for instance) at 70MPH (the legal maximum here in the UK) is likely to be non-fatal - and even if by some miracle the driver survives, you've suddenly dumped a lot of dangerous debris all over the motorway that everyone else is now hitting at 70MPH. That's not accounting for the fact that migration to the new system would take a long time so you'll have a mix of safety and normal cars on the road (and even when the vast majority have migrated there will be some exceptions - vintage cars for instance are unlikely to get these features added). If we could magically switch all cars to these "fatality free" versions simultaneously tomorrow, I'd still feel confident giving odds of seeing the first fatality within a week.
Can you imagine a tiered internet that made those sites so painfully slow to use that people would be willing to give up their privacy to get around it?
There are enough high profile miscarriages of justice that if someone won't accept that a person with nothing to hide has nothing to fear then they're either burying their head in the sand or they really mean "but it's so much more convenient" - people seem very happy to give up privacy for convenience right up to the point where they realise they've given away too much. You can't ever fight that, it's human nature - sure certain defining moments in time may shock people into action, but they'll always return to the former position because it's too easy not to.
Exactly - them defaulting to thinking you have something to hide only matters if they can find you.
There was an internet before Geocities, that was just one of the earlier attempts to mainstream it so users could be exploited.
Number 6: Welcome friend, I'm number 6.
Number 15: I'm number 15. What number are you?
Homer: I am not a number! I am a man! And don't you ever... oh wait, I'm number 5. Ha, ha! In your face, number 6!
Exactly, I'm shocked by the standard of geekery I'm seeing here. I guess gone are the days where, if a /.er was identified from a photo taken with "friends", Google would assume he was Klingon.
I think the point is more that, once a computer can simulate the best of human thinking, it will be more productive to have the computer think about these issues. The computer can process the data faster and doesn't need food or sleep, and it's far easier to duplicate a computer process multiple times than to create more genius level humans (not to mention far quicker, none of that messy growing up and puberty stuff to deal with, just load the disk image and go).
That depends - you say you're assuming there's an equal chance of spotting a known and an unknown word. It's more likely that his software is better at recognising the word that has been generated by an algorithm rather than some randomly obscured word (that was then run through the algorithm), so while the success rate for two known words might not be 30% there's a good chance that it would be better than 1%, since that figure is factoring in the ability to recognise a word that even Google's OCR couldn't realiably recognise.
You can make a reasonable guess as a human - the unknown word is usually unknown for a reason (archaic type rendering two letters together, an ink smudge on one letter, etc) which a human can spot with reaonsable accuracy once they know the trick. If they introduced some of these tricks on the known words they might make it more difficult (and of course, just because a human can spot the difference reliably, doesn't mean it's easy to do in software, as supported by the results in the video).
It's a little confusing but the presentation seems to suggest if you get one of the words wrong too many times in succession - 32 it claims - it will log your IP and force you to get both right (I'm not sure how that works, maybe it uses a word with a large number of identical responses, or maybe it just generates two words it knows). It doesn't say if this is reset after you have successfully got both right, if it was you could still brute force it (the counter would reset after an average 100 entries if there's a 1% chance of matching both, in which case for every 132 entries you'd still get about a quarter of them through which is not a bad hit rate), if not the suggestion is "dynamic IP" - not particularly useful if you plan to spam thousands of these things.
Yes, centralised captcha would seem eventually destined to failure. The better approach would be that each instance be reasonably unique (even if it was only so far as each users uploading their own images), that way a determined spammer might break one site, but he can't use what he did there to attack others, each time he'll have to start from scratch. The obvious downside is the massive cost - instead of one person spending a week linking images in the database, you'd need one person per company. Still, for a reasonably large company for whom spam is a big issue it might eventually still become the best option (and to periodically add to and remove from the database of existing images).
Recaptcha allows you to make some mistakes as standard than most othert captcha solutions I think (and TFA's findings seem to support this also - it suggests you can get one word wrong and one letter from the other word, although when I tested that it was too much, but I have successfully tested one right and one wrong word and still passed the captcha). Really you're being served one captcha word and one word Google's book scanning project couldn't recognise, you're solving the captcha word but the other word (usually the harder word to read) you're only adding to the statistical weighting of what the word probably is, in other words you can afford to be a little wrong and still get the captcha right.