Except this isn't a random input. These are self selected personal identifiers. If you were challenged to guess the gender of names like Fizzup, Derekloffin, Harlequin80, & ljw1004 I would expect you to get a fairly high % of the genders correct because most are either unknowable (ljw) or self obvious (Derek).
But irrespective of that to try and extrapolate any fine grain conclusions when 33% of your dataset cannot be identified and an additional 9% are incorrectly identified is just crazy. Their data noise level approaches 50%, you cannot draw conclusions at 3% with any confidence level with that level of noise.
But they only apply the logic to a sample of 100 and then extrapolate to their known gender data, there is nothing that implies they excluded the unguessables from the main data set. Unless you had someone sit down and manually exclude the unguessables or incorrects your results are skewed a mile.
From TFA "To see whether gender was apparent in an eBay auction, they challenged 400 people to guess the gender of 100 randomly chosen sellers. Just using clues like the names of the sellers and what other items they tended to sell, participants correctly guessed the gender of 56%, declared 35% unguessable, and got less than 9% wrong"
Hardly what I would have called particularly thorough for a study that is going to make gender based arguments.
And for something so pivotal it isn't particularly thorough.
From TFA "To see whether gender was apparent in an eBay auction, they challenged 400 people to guess the gender of 100 randomly chosen sellers. Just using clues like the names of the sellers and what other items they tended to sell, participants correctly guessed the gender of 56%, declared 35% unguessable, and got less than 9% wrong"
From that they get 44% incorrect. That is a fucking MASSIVE error level to apply to your stats. Particularly once you correct for other things and are claiming only a 3% differential.
The summary doesn't really reflect the articles findings. Yes female identified sellers received less money that sellers identified as male but once it was corrected for various things the difference was down to 97c per dollar. So the difference is no where near the level the summary, or even the opening part of the article claims.
There are also huge windows of error in their statistical sampling. They analysed 1.1m records but only had 400 sellers names assessed for expected gender. From those 35% couldn't be guessed an 9% got it wrong. So to apply that level of inaccuracy and then claim 3% as a significant difference is a bit of a stretch.
Bahahahaha. And MC Hammer pants really really really look good. That's why people wore them, not because they were having 5 minutes of fame. Nope definitely because they looked good.
How have you never heard of this guy? I could never pick one of his songs out of a medley but this guy is such an attention whore train wreck that even I, stuck in my basement, have heard of him. This was the guy who rushed on stage at an awards ceremony (VMA) to abuse the judges for picking Taylor Swift over Beyonce when he has absolutely nothing to do with Beyonce. And he married that Kardashian woman.
I'm kinda proud of you that you don't know who he is. Proud and really really jealous.
Uh huh. Close all gas storage facilities. Right... Do these people not even consider what the outcome of that would be? Next it will be get rid of all dams and water storage facilities, it's just as logical.
Yeah I mentioned Ducatis in another post. Reading back it wasn't obvious but I meant manually closing the valves in conjunction with an electric system. I've worked on ducati engines before and trying to get the shims and loaded vs unloaded gap right sucks.
Not necessarily at all. If you access the patient records via a citrix system for example there would be no reason to believe the patient records were compromised just because the host machine was. This can also be extended to applications that communicate with a database server. There is nothing that prevents that being encrypted every step of the way.
The only way would be via screenshots and your data rate would be terrible.
Short sighted from an industry view, probably not from the hospitals view. You would hope they have air gapped their network from the internet at this stage while they reappraise their security and plug holes. From my understanding the ransomware attackers don't normally attack the same target twice as you are less likely to pay up if you think it will happen again. So this should protect them from the current infection.
It also wouldn't surprise me if patient records were untouched. Those are probably behind higher levels of security than the rest of the network. What I suspect happened is they lost a way of accessing them because all their other systems went down.
By an absolute mile. At $17,000 you would just pay it straight away. They would have lost far more as a result of the systems being offline, and assuming the ransomware had got itself all through they systems it would have been orders of magnitude more to clean the system if it was even possible.
Just a silly thought. But if you put this into a current PC as it's main storage medium how long would it last before it was full assuming it could never delete anything written, so every single write is to a new block. You definitely wouldn't want it to swap to it but I'm kinda thinking along the lines of building up stacks of notebooks. Never accidentally delete anything ever again. Of course all your porn that got cached to disk would be there as well.
The force required to open a valve comes from the size of the spring that closes the valve. There isn't anything about engines that inherently requires that level of force but most valve systems are self closing with the cam only being responsible for opening. The force in the spring needs to be high enough to close the valve as fast as the cam allows and to be able to prevent the valve bouncing when it hits the seat.
In an ideal world we would be mechanically closing the valve as well as opening it. It requires significantly less force, and allows higher valve head speed and longer lift times. The disadvantages are you have to maintain the system a lot more and it is more complex. This is called a desmodromic valve system, Ducati motorcycles are the most common example.
While the desmo system still uses a cam if you could replace that system with an electrical system you could potentially see higher performance outcomes.
Firstly changing how soon you open a valve and how late you close it can be translated to duration, so you kinda contradict yourself there. But more importantly if you change the shape of your cam and you change the lift. The Honda VTEC system has multiple lobes which are engaged at various throttle and rpm levels. This has the effect of changing the duration and the lift. Specifically it changes the lift far more than the duration.
The biggest change that can be made to the valve system is to move away from springs and to a manual close system. This way your valves can physically move faster and hence have a longer period of time open without valve bounce or smashing into the piston.
If that is where they are. But I suspect they are probably somewhere more developed than that and somewhere that the US can exert a fair bit of pressure. It is unlikely to be a state sponsored attack so they won't be getting any support in hiding.
They picked the wrong target. If you hit a small business it's easier to pay. If you hit a large business you pay because you don't want people to find out. You hit a hospital though and people could die and it is very very public.
Right about now there will be a whole lot of resources targeted towards finding these people. They are fucked.
Would make fantastic replacement bones then. You could form a bone then make it malleable so it can be inserted via key hole or around something else you don't want to move and then have it move back to the original correct shape.
How you keep it in the malleable temp range during this process though may be the hardest part.
Slashdot Mirror
Except this isn't a random input. These are self selected personal identifiers. If you were challenged to guess the gender of names like Fizzup, Derekloffin, Harlequin80, & ljw1004 I would expect you to get a fairly high % of the genders correct because most are either unknowable (ljw) or self obvious (Derek).
But irrespective of that to try and extrapolate any fine grain conclusions when 33% of your dataset cannot be identified and an additional 9% are incorrectly identified is just crazy. Their data noise level approaches 50%, you cannot draw conclusions at 3% with any confidence level with that level of noise.
But they only apply the logic to a sample of 100 and then extrapolate to their known gender data, there is nothing that implies they excluded the unguessables from the main data set. Unless you had someone sit down and manually exclude the unguessables or incorrects your results are skewed a mile.
From TFA "To see whether gender was apparent in an eBay auction, they challenged 400 people to guess the gender of 100 randomly chosen sellers. Just using clues like the names of the sellers and what other items they tended to sell, participants correctly guessed the gender of 56%, declared 35% unguessable, and got less than 9% wrong"
Hardly what I would have called particularly thorough for a study that is going to make gender based arguments.
And for something so pivotal it isn't particularly thorough.
From TFA "To see whether gender was apparent in an eBay auction, they challenged 400 people to guess the gender of 100 randomly chosen sellers. Just using clues like the names of the sellers and what other items they tended to sell, participants correctly guessed the gender of 56%, declared 35% unguessable, and got less than 9% wrong"
From that they get 44% incorrect. That is a fucking MASSIVE error level to apply to your stats. Particularly once you correct for other things and are claiming only a 3% differential.
The summary doesn't really reflect the articles findings. Yes female identified sellers received less money that sellers identified as male but once it was corrected for various things the difference was down to 97c per dollar. So the difference is no where near the level the summary, or even the opening part of the article claims.
There are also huge windows of error in their statistical sampling. They analysed 1.1m records but only had 400 sellers names assessed for expected gender. From those 35% couldn't be guessed an 9% got it wrong. So to apply that level of inaccuracy and then claim 3% as a significant difference is a bit of a stretch.
Bahahahaha. And MC Hammer pants really really really look good. That's why people wore them, not because they were having 5 minutes of fame. Nope definitely because they looked good.
You're welcome. I can only assume it was the insightful reading of his character.
As for is that all there is to know? Pretty much. He's a media attention whore who thinks he is gods gift to the world.
Even more sad then.... I don't live in America :(
How have you never heard of this guy? I could never pick one of his songs out of a medley but this guy is such an attention whore train wreck that even I, stuck in my basement, have heard of him. This was the guy who rushed on stage at an awards ceremony (VMA) to abuse the judges for picking Taylor Swift over Beyonce when he has absolutely nothing to do with Beyonce. And he married that Kardashian woman.
I'm kinda proud of you that you don't know who he is. Proud and really really jealous.
No he kept it on Tidal and removed it from his personal website.
And his music is terrible.
Uh huh. Close all gas storage facilities. Right... Do these people not even consider what the outcome of that would be? Next it will be get rid of all dams and water storage facilities, it's just as logical.
Yeah I mentioned Ducatis in another post. Reading back it wasn't obvious but I meant manually closing the valves in conjunction with an electric system. I've worked on ducati engines before and trying to get the shims and loaded vs unloaded gap right sucks.
Not necessarily at all. If you access the patient records via a citrix system for example there would be no reason to believe the patient records were compromised just because the host machine was. This can also be extended to applications that communicate with a database server. There is nothing that prevents that being encrypted every step of the way.
The only way would be via screenshots and your data rate would be terrible.
There is one behaviour which he demonstrates which I don't think I have ever seen anyone else do. Is there a way I could msg it to you?
Short sighted from an industry view, probably not from the hospitals view. You would hope they have air gapped their network from the internet at this stage while they reappraise their security and plug holes. From my understanding the ransomware attackers don't normally attack the same target twice as you are less likely to pay up if you think it will happen again. So this should protect them from the current infection.
It also wouldn't surprise me if patient records were untouched. Those are probably behind higher levels of security than the rest of the network. What I suspect happened is they lost a way of accessing them because all their other systems went down.
Oh no you've attracted APK. On the plus side through perhaps this will mean Coren22 gets a little less stalked.
By an absolute mile. At $17,000 you would just pay it straight away. They would have lost far more as a result of the systems being offline, and assuming the ransomware had got itself all through they systems it would have been orders of magnitude more to clean the system if it was even possible.
Just a silly thought. But if you put this into a current PC as it's main storage medium how long would it last before it was full assuming it could never delete anything written, so every single write is to a new block. You definitely wouldn't want it to swap to it but I'm kinda thinking along the lines of building up stacks of notebooks. Never accidentally delete anything ever again. Of course all your porn that got cached to disk would be there as well.
Well I did say close as fast as the cam allows but yes it might not have been clear.
The force required to open a valve comes from the size of the spring that closes the valve. There isn't anything about engines that inherently requires that level of force but most valve systems are self closing with the cam only being responsible for opening. The force in the spring needs to be high enough to close the valve as fast as the cam allows and to be able to prevent the valve bouncing when it hits the seat.
In an ideal world we would be mechanically closing the valve as well as opening it. It requires significantly less force, and allows higher valve head speed and longer lift times. The disadvantages are you have to maintain the system a lot more and it is more complex. This is called a desmodromic valve system, Ducati motorcycles are the most common example.
While the desmo system still uses a cam if you could replace that system with an electrical system you could potentially see higher performance outcomes.
Firstly changing how soon you open a valve and how late you close it can be translated to duration, so you kinda contradict yourself there. But more importantly if you change the shape of your cam and you change the lift. The Honda VTEC system has multiple lobes which are engaged at various throttle and rpm levels. This has the effect of changing the duration and the lift. Specifically it changes the lift far more than the duration.
The biggest change that can be made to the valve system is to move away from springs and to a manual close system. This way your valves can physically move faster and hence have a longer period of time open without valve bounce or smashing into the piston.
If that is where they are. But I suspect they are probably somewhere more developed than that and somewhere that the US can exert a fair bit of pressure. It is unlikely to be a state sponsored attack so they won't be getting any support in hiding.
They picked the wrong target. If you hit a small business it's easier to pay. If you hit a large business you pay because you don't want people to find out. You hit a hospital though and people could die and it is very very public.
Right about now there will be a whole lot of resources targeted towards finding these people. They are fucked.
Would make fantastic replacement bones then. You could form a bone then make it malleable so it can be inserted via key hole or around something else you don't want to move and then have it move back to the original correct shape.
How you keep it in the malleable temp range during this process though may be the hardest part.