The humans are of two types; One who control (type-1); the other who is controlled (type-2). All national boundaries or any other group formation is just to aid type-1. And yes, any information flow is to aid type-1. And yes all programming (of mind) is to aid type-1. Of course it aids type-1 to keep type-2 subservient and hence all system of information flow we have (schools, economy, even internet) exists to aid type-1 if you peel the onion.
Seems u hv developed dedicated hardware (neuron complex) to do fourier decomposition of a signal. That is given a function f, find a way to decompose/construct back f using fundamental frequencies. Essentially a way to decompose a complex object to its fundamental orthogonal building blocks; like say factor a number into its prime factors (fundamental theorem of arithmetic) or decompose a music piece into the notes/structure of notes; spoken language to basic phonemes [compressions like jpeg/mp3 work the same way.. keep the important n drop the fluff]. It's like finding the basis-set in linear algebra and figuring out their linear combination to get any given vector. I think you should be able to use this to do some cool things like very rapid learning of a new language/art (like singing or dance or any other art form) because all these involved finding the fundamental basis set and just doing linear combination of these.
what's wrong in ripping off of "gullible customers"? isn't the reason to be gullible is to be ripped off? how else will wealth move to the right place?
The evil doer is first getting power from the mother ship and destroying the mother ship with that power. That is it is not self powered or has external power source. So all you do to protect yourself is don't feed someone who can turn against you. The OS should first give a little power (few milli amps for few mins say) and then ask for clear authorization; a user should approve the device etc. Else just cut off power to that USB.
It's like you don't support (with money say) a family member or friend (say his business) who can become so powerful with your money that they can turn against you / your business.
that's north-north-east (not north-north-west)..seems they use a clock wise notation with north being 0, east 90, south 180, west 270. So 27 degree is close to north-north-east.
It did go wrong a few years ago. In South America, a plane crashed into the amazon because the coordinates entered were something like 270 degree rather than 27.0; so it flew near straight west instead of north-north-west. I think ti's because they introduced a 4 digit numbering 0270 was seen as 270 instead of 27.0 since the pilots were used to seeing only 3 digit code (like 270). There is an air crash investigation video about it. Not sure how this can be avoided -- may be every flight system should upload a read-only data to a central system to do a double check.
Seems it's a classic case of chicken n egg problem. The hardware [neural net/connections in spinal cord/brain] gets generated only if there is use/ a use-case. But you can't use if you don't have feedback. So both producer n consumer have stalled. You create a seed disturbance (using VR to coax the logical part to ask for input.. ).. n this gets into a virtuous positive feedback loop. Producer starts churning and the cycle goes. You came out of a dead-lock. In fact a determined human with enormous will-power/passion should be able to kick start the cycle without the VR aid; but it's rare for a normal human to do it without the external aid. It's like believing you can fly and taking a leap off the face of a cliff
Code should be rid of any fluff. Any description of what/how/why about the code should be in comments - well written and this is for human consumption; so it can be elaborate to avoid any misunderstanding. Code is written once but read/debugged many many times; Maintenance is the real cost. So you don't want to have noise in the code which takes a higher cognitive-load on your brain; your visual system too processes extra lines. These are unnecessary. Only after reading the code and thinking about it, you realize that code is harmless/unnecessary (eg while debugging or extending the software). The point is a line of code should have the least labor (cognitive/visual) from the human. A code which doesn't exist does this job the most perfectly. Hence write compact code; use standard libraries. Think at higher level of abstraction (like say containers instead of managing an array of objects); this needs spending lot more time in the design phase -- seeing the big picture of the various moving parts.
Except in embedded systems and few webscale systems, performance is really not an issue in most software. What is needed is how easy it is for a human to program and extend it. The issues of 90s n before is not there now for most applications. Programming is getting a human thought/idea..model it into a mathematical problem/model and from there the faster n easier the machine (language) can help, it's better. e.g you say you want a collection of data to be sorted, using some thing as key/ordering; and not care how the machine/language does it for you.
C may still be good for those embedded/performance critical; for rest I see language like python does the human-idea/algorithm to final-result much faster. It enables you to think in high level data structures/data manipulation and not worry about machine level issues. The higher you can abstract things, the faster and more easily you can extend the software. And s/w is an iterative task; you never know how the system is going to grow; so the easier it is to extend it's better.
This must be made a specific instance of a general object recognition. With AI and image recognition seen great progress in the last 4 or 5 years, sending a picture to the backend should give all details of it. [eg send a pic of a flower, it should tell me what it is]
Tesla says two things: 1) I did not see the trailer (due to white color/sky what not) 2) I assumed it's a bridge or something above. So which is right? Either you saw or you didn't see. Assuming it is 2, it is very sloppy to not calculate how high the bridge is [using mobileeye or what not camera based system, it must be trivial to compute the clearance].
I believe the s/w assumed anything above say 3 feet is a bridge - in this case the trailer's clearance might have been say 4 feet. Not sure how tall a tesla is, may be 5 feet. It's really bad software design not to ensure that bridge or whatever you assumed has the 5 feet clearance you need to pass under. I don't see any reason why the fault is totally in the s/w in this case. Also for this problem, I don't think google's LIDAR or other highly complex/heavy-weight solution is necessary. The existing camera based thing should have worked if they had taken into the needed clearance.
That statistics comparison is misleading; because the population that drives a tesla and a normal car is way different. A tesla owner is likely to be someone well educated, well off and usually of higher IQ than the general population. To do apples to apples comparison, the whole population should be given a tesla to drive around and compare the improvement with non-tesla vehicle.
This statistics is like asking high school algebra problems to Phd students of math department in an university and comparing the results with Phd students of say music department.
The s/w in tesla is purely lacking here. It must have sensed a moving obstacle (bridges, billboards dont' move -- I bet tesla's "AI" is not smart enough to distinguish a stationary object from a potential moving collision object (say a herd of bison/elephants in a single file crossing the road). Once it senses, it could atleast alert the human.
In this case, it saw the the object's outline is above its path and assumed it has a safe "tunnel" to drive thru'. The crucial thing is the way tesla sees and a human sees is different. We see a lot more peripheral things; not just a small "tunnel" vision of clearance to give go ahead to motion. We can anticipate - a bison following another is very likely going to come in front of me, even though right now I see clearance ahead of me.
Still tesla might have tried to swerve in the last split second; just that there wasn't enough time to get the vehicle out -- this is strange given that electric motors can pump in huge torque almost instantly, I guess the limits were not that of physics/inertia. May be one day, it will swerve in such sudden appearance of obstacles and avoid a collision (like how say a honey bee flies and can do sudden change of direction/speed).
And it's also possible the real reason for the fatality could be the final hitting of the electric pole; may be if the pole weren't there, the driver may be alive given the safety features of the car.
The truck driver is not expecting the other guy to stop; he only expects him to be a bit easy on his gas; or even expect the other guy to "see" him. In this case, Tesla did not see (or rather saw and assumed it's a bridge). Normally we see the left turn is possible assuming the other driver sees you and gently takes his foot from the gas a bit. In this case, tesla was going at the same speed. True the truck driver could have waited..but it's very natural people see that the other guy is only going to slow down a bit (ie not expecting him to touch his brake..but just ease a bit on the gas) and that split second is enough to complete the left-turn. Humans drive and humans have some built in assumptions; we dont' go by the 'laws of right of way'. In fact at times, you need to do what is expected of you; if you stick to laws and slam the brakes, very well someone can rear end you.
Of course the s/w must be better. First a general system should look for sudden intrusions (like left turns, intersections). And not be able to distinguish from a trailer and a bridge is clearly not acceptable. Yeah saying that I assumed it's like a hole, I can pass thru' -- then your system should detect how the hole is narrowing and do you have the time to clear it (in this case it would've passed under the trailer without collision (like in a james bond movie))
I guess economies of scale might still prefer doing mass manufacture of highly intricate - less volume components in one or few global centers. e.g like making iphones or CPUs. Even for an automobile, the engine could be mass produced in one or two places while the body-building can be done locally. The thing is anything that is too intricate and involves lot of skilled labor is better done in one or two places to reduce cost. Not sure how doing this is say a dozen places around the world will be cost effective. Since the component is small, it should be relatively less expensive to ship it.
It's kind of an answer to this question in the quoted text "what are we going to do with all the unemployable people? ".
The point is society can't take care of an individual by telling him/her how he should spend the waking hours. We do assume that tech takes care of basic needs [thur' say a society UBI/universal basic income/ food stamps/housing etc]. Even today if you apply for a loan (govt sponsored) or go for govt assistance, you will get it. If you are sober, you can follow the steps I had listed.
Basically humans have reached that point in evolution where they don't need to work for survival; they need to find higher purpose (don't wanna bring in any religious mumbo-jumbo). This higher purpose I alluded thru' the use of the word 'creative/creation'. Just like we used to take for granted free-oxygen, we need to take for granted free food/shelter/clothing/basic-entertainment. The greatest challenge then is what to do with all the free time.
[btw health care is not really an issue, if you stop doing all the non-sense capitalism has made a man to do -- all unnatural sleep/work/eating patterns]
Since socialism will never take hold in the US until things are at the French Revolution level, what are we going to do with all the unemployable people? It's not nice to say, but there are a group of people who are absolutely incapable of doing anything beyond warehouse work or factory work. Heck, there are corporate employees who are incapable of doing anything outside a narrow processing-type job description. For these people, I do kind of wish for a return to the pre-automation days when you had 10,000+ people working in a steel mill, or another 10,000+ just churning out paperwork at a corporate job. Those people earned a decent middle class salary, and had a good life. I doubt anyone growing up now is going to have it so good.
As technology takes care of your basic needs, you need to find ways to fill your time. Build healthy hobbies; be creative, meditate, become spiritual. You need to be a creator even if no one cares to see your creation (let alone pay money to buy your stuff or watch you perform). In the short term, most will end up drug addicts. A few that don't get killed will wake up, take up meditation and find inner peace.
The very reason each person is given the right to carry weapon is to fight against rogue power (the reason as per the constitution). Imagine a state where most (99%) of folks are brainwashed to believe the current power is just/fair; only a few handful sense the truth and want to fight back. Don't you want them to be enabled in that situation? So any way to mitigate their ability is against the spirit of the constitution.
Exceptions should be allowed only for the military and the police.
Why? the 2nd amendment wants you to be able to fight against precisely these when a rogue force takes over power. You don't want to be shooting with a liability when the other side has more powerful weapon.
Finding NULL return usage from malloc/calloc is something static-analysis tools (like beefed up lint tools) easily spot. Not sure why they didn't run the source thru' static analysis or marked the flagging as noise. This case is finding the input arg to calloc could be zero and hence can get a NULL return (they say implementation dependent; most cases it's NULL when you ask for zero bytes/items to a calloc library)
assert originated when the extra if check was costly, usually in embedded systems. You don't want to do a check when you know 100% sure it is never going to happen. Of course, cpu cycles are not that critical these days in 99.99% or more applications that this style of assert is no longer needed; not only not-needed they are dangerous as seen in this exploit. [of course if you an embedded system like say switches/network processors, where each cycle counts, you may still dont' want to put in unnecessary checks (like if (not is-earth-round()) do...)
Moreover if someone is really after the last cycle, they should take the source and build a custom one; instead of making the library vulnerable like in this case.
No, the residents in the address are sorted using some method (likely age) and if the recipient is in the top 80%, the mail gets delivered; else the next day.
Slashdot Mirror
The humans are of two types; One who control (type-1); the other who is controlled (type-2). All national boundaries or any other group formation is just to aid type-1. And yes, any information flow is to aid type-1. And yes all programming (of mind) is to aid type-1. Of course it aids type-1 to keep type-2 subservient and hence all system of information flow we have (schools, economy, even internet) exists to aid type-1 if you peel the onion.
Seems u hv developed dedicated hardware (neuron complex) to do fourier decomposition of a signal. That is given a function f, find a way to decompose/construct back f using fundamental frequencies. Essentially a way to decompose a complex object to its fundamental orthogonal building blocks; like say factor a number into its prime factors (fundamental theorem of arithmetic) or decompose a music piece into the notes/structure of notes; spoken language to basic phonemes [compressions like jpeg/mp3 work the same way.. keep the important n drop the fluff]. It's like finding the basis-set in linear algebra and figuring out their linear combination to get any given vector. I think you should be able to use this to do some cool things like very rapid learning of a new language/art (like singing or dance or any other art form) because all these involved finding the fundamental basis set and just doing linear combination of these.
why not :x? single char n no caps
Is 'machine learning' ok?
what's wrong in ripping off of "gullible customers"? isn't the reason to be gullible is to be ripped off? how else will wealth move to the right place?
The evil doer is first getting power from the mother ship and destroying the mother ship with that power. That is it is not self powered or has external power source. So all you do to protect yourself is don't feed someone who can turn against you. The OS should first give a little power (few milli amps for few mins say) and then ask for clear authorization; a user should approve the device etc. Else just cut off power to that USB.
It's like you don't support (with money say) a family member or friend (say his business) who can become so powerful with your money that they can turn against you / your business.
that's north-north-east (not north-north-west)..seems they use a clock wise notation with north being 0, east 90, south 180, west 270. So 27 degree is close to north-north-east.
It did go wrong a few years ago. In South America, a plane crashed into the amazon because the coordinates entered were something like 270 degree rather than 27.0; so it flew near straight west instead of north-north-west. I think ti's because they introduced a 4 digit numbering 0270 was seen as 270 instead of 27.0 since the pilots were used to seeing only 3 digit code (like 270). There is an air crash investigation video about it. Not sure how this can be avoided -- may be every flight system should upload a read-only data to a central system to do a double check.
We've had our eye on you for sometime, Mr Anderson.
Seems it's a classic case of chicken n egg problem. The hardware [neural net/connections in spinal cord/brain] gets generated only if there is use/ a use-case. But you can't use if you don't have feedback. So both producer n consumer have stalled. You create a seed disturbance (using VR to coax the logical part to ask for input.. ).. n this gets into a virtuous positive feedback loop. Producer starts churning and the cycle goes. You came out of a dead-lock. In fact a determined human with enormous will-power/passion should be able to kick start the cycle without the VR aid; but it's rare for a normal human to do it without the external aid. It's like believing you can fly and taking a leap off the face of a cliff
Code should be rid of any fluff. Any description of what/how/why about the code should be in comments - well written and this is for human consumption; so it can be elaborate to avoid any misunderstanding. Code is written once but read/debugged many many times; Maintenance is the real cost. So you don't want to have noise in the code which takes a higher cognitive-load on your brain; your visual system too processes extra lines. These are unnecessary. Only after reading the code and thinking about it, you realize that code is harmless/unnecessary (eg while debugging or extending the software). The point is a line of code should have the least labor (cognitive/visual) from the human. A code which doesn't exist does this job the most perfectly. Hence write compact code; use standard libraries. Think at higher level of abstraction (like say containers instead of managing an array of objects); this needs spending lot more time in the design phase -- seeing the big picture of the various moving parts.
Except in embedded systems and few webscale systems, performance is really not an issue in most software. What is needed is how easy it is for a human to program and extend it. The issues of 90s n before is not there now for most applications. Programming is getting a human thought/idea ..model it into a mathematical problem/model and from there the faster n easier the machine (language) can help, it's better. e.g you say you want a collection of data to be sorted, using some thing as key/ordering; and not care how the machine/language does it for you.
C may still be good for those embedded/performance critical; for rest I see language like python does the human-idea/algorithm to final-result much faster. It enables you to think in high level data structures/data manipulation and not worry about machine level issues. The higher you can abstract things, the faster and more easily you can extend the software. And s/w is an iterative task; you never know how the system is going to grow; so the easier it is to extend it's better.
This must be made a specific instance of a general object recognition. With AI and image recognition seen great progress in the last 4 or 5 years, sending a picture to the backend should give all details of it. [eg send a pic of a flower, it should tell me what it is]
Tesla says two things: 1) I did not see the trailer (due to white color/sky what not) 2) I assumed it's a bridge or something above. So which is right? Either you saw or you didn't see. Assuming it is 2, it is very sloppy to not calculate how high the bridge is [using mobileeye or what not camera based system, it must be trivial to compute the clearance].
I believe the s/w assumed anything above say 3 feet is a bridge - in this case the trailer's clearance might have been say 4 feet. Not sure how tall a tesla is, may be 5 feet. It's really bad software design not to ensure that bridge or whatever you assumed has the 5 feet clearance you need to pass under. I don't see any reason why the fault is totally in the s/w in this case. Also for this problem, I don't think google's LIDAR or other highly complex/heavy-weight solution is necessary. The existing camera based thing should have worked if they had taken into the needed clearance.
That statistics comparison is misleading; because the population that drives a tesla and a normal car is way different. A tesla owner is likely to be someone well educated, well off and usually of higher IQ than the general population. To do apples to apples comparison, the whole population should be given a tesla to drive around and compare the improvement with non-tesla vehicle.
This statistics is like asking high school algebra problems to Phd students of math department in an university and comparing the results with Phd students of say music department.
The s/w in tesla is purely lacking here. It must have sensed a moving obstacle (bridges, billboards dont' move -- I bet tesla's "AI" is not smart enough to distinguish a stationary object from a potential moving collision object (say a herd of bison/elephants in a single file crossing the road). Once it senses, it could atleast alert the human.
In this case, it saw the the object's outline is above its path and assumed it has a safe "tunnel" to drive thru'. The crucial thing is the way tesla sees and a human sees is different. We see a lot more peripheral things; not just a small "tunnel" vision of clearance to give go ahead to motion. We can anticipate - a bison following another is very likely going to come in front of me, even though right now I see clearance ahead of me.
Still tesla might have tried to swerve in the last split second; just that there wasn't enough time to get the vehicle out -- this is strange given that electric motors can pump in huge torque almost instantly, I guess the limits were not that of physics/inertia. May be one day, it will swerve in such sudden appearance of obstacles and avoid a collision (like how say a honey bee flies and can do sudden change of direction/speed).
And it's also possible the real reason for the fatality could be the final hitting of the electric pole; may be if the pole weren't there, the driver may be alive given the safety features of the car.
The truck driver is not expecting the other guy to stop; he only expects him to be a bit easy on his gas; or even expect the other guy to "see" him. In this case, Tesla did not see (or rather saw and assumed it's a bridge). Normally we see the left turn is possible assuming the other driver sees you and gently takes his foot from the gas a bit. In this case, tesla was going at the same speed. True the truck driver could have waited ..but it's very natural people see that the other guy is only going to slow down a bit (ie not expecting him to touch his brake..but just ease a bit on the gas) and that split second is enough to complete the left-turn. Humans drive and humans have some built in assumptions; we dont' go by the 'laws of right of way'. In fact at times, you need to do what is expected of you; if you stick to laws and slam the brakes, very well someone can rear end you.
Of course the s/w must be better. First a general system should look for sudden intrusions (like left turns, intersections). And not be able to distinguish from a trailer and a bridge is clearly not acceptable. Yeah saying that I assumed it's like a hole, I can pass thru' -- then your system should detect how the hole is narrowing and do you have the time to clear it (in this case it would've passed under the trailer without collision (like in a james bond movie))
I guess economies of scale might still prefer doing mass manufacture of highly intricate - less volume components in one or few global centers. e.g like making iphones or CPUs. Even for an automobile, the engine could be mass produced in one or two places while the body-building can be done locally. The thing is anything that is too intricate and involves lot of skilled labor is better done in one or two places to reduce cost. Not sure how doing this is say a dozen places around the world will be cost effective. Since the component is small, it should be relatively less expensive to ship it.
It's kind of an answer to this question in the quoted text "what are we going to do with all the unemployable people? ". The point is society can't take care of an individual by telling him/her how he should spend the waking hours. We do assume that tech takes care of basic needs [thur' say a society UBI/universal basic income/ food stamps/housing etc]. Even today if you apply for a loan (govt sponsored) or go for govt assistance, you will get it. If you are sober, you can follow the steps I had listed.
Basically humans have reached that point in evolution where they don't need to work for survival; they need to find higher purpose (don't wanna bring in any religious mumbo-jumbo). This higher purpose I alluded thru' the use of the word 'creative/creation'. Just like we used to take for granted free-oxygen, we need to take for granted free food/shelter/clothing/basic-entertainment. The greatest challenge then is what to do with all the free time. [btw health care is not really an issue, if you stop doing all the non-sense capitalism has made a man to do -- all unnatural sleep/work/eating patterns]
Since socialism will never take hold in the US until things are at the French Revolution level, what are we going to do with all the unemployable people? It's not nice to say, but there are a group of people who are absolutely incapable of doing anything beyond warehouse work or factory work. Heck, there are corporate employees who are incapable of doing anything outside a narrow processing-type job description. For these people, I do kind of wish for a return to the pre-automation days when you had 10,000+ people working in a steel mill, or another 10,000+ just churning out paperwork at a corporate job. Those people earned a decent middle class salary, and had a good life. I doubt anyone growing up now is going to have it so good.
As technology takes care of your basic needs, you need to find ways to fill your time. Build healthy hobbies; be creative, meditate, become spiritual. You need to be a creator even if no one cares to see your creation (let alone pay money to buy your stuff or watch you perform). In the short term, most will end up drug addicts. A few that don't get killed will wake up, take up meditation and find inner peace.
The very reason each person is given the right to carry weapon is to fight against rogue power (the reason as per the constitution). Imagine a state where most (99%) of folks are brainwashed to believe the current power is just/fair; only a few handful sense the truth and want to fight back. Don't you want them to be enabled in that situation? So any way to mitigate their ability is against the spirit of the constitution.
Exceptions should be allowed only for the military and the police.
Why? the 2nd amendment wants you to be able to fight against precisely these when a rogue force takes over power. You don't want to be shooting with a liability when the other side has more powerful weapon.
Finding NULL return usage from malloc/calloc is something static-analysis tools (like beefed up lint tools) easily spot. Not sure why they didn't run the source thru' static analysis or marked the flagging as noise. This case is finding the input arg to calloc could be zero and hence can get a NULL return (they say implementation dependent; most cases it's NULL when you ask for zero bytes/items to a calloc library)
assert originated when the extra if check was costly, usually in embedded systems. You don't want to do a check when you know 100% sure it is never going to happen. Of course, cpu cycles are not that critical these days in 99.99% or more applications that this style of assert is no longer needed; not only not-needed they are dangerous as seen in this exploit. [of course if you an embedded system like say switches/network processors, where each cycle counts, you may still dont' want to put in unnecessary checks (like if (not is-earth-round()) do...)
Moreover if someone is really after the last cycle, they should take the source and build a custom one; instead of making the library vulnerable like in this case.
No, the residents in the address are sorted using some method (likely age) and if the recipient is in the top 80%, the mail gets delivered; else the next day.