8 ms·

I guess it depends on what one's of "high" is with respect to false positives. I think he's arguing that the costs associated with these false positives are su

by sinwave 12y ago

I guess it depends on what one's of "high" is with respect to false positives. I think he's arguing that the costs associated with these false positives are such that even an "extremely low" absolute number (like one in a thousand) is unacceptable. Consider, for example, if airplanes carried tens of thousands, rather than hundreds, of passengers. Then our acceptable false positive rate may change by an order of magnitude.

bostik 12y ago
> even an "extremely low" absolute number (like one in a thousand) is unacceptable. Like herd immunity in epidemology? The really large corporations are mostly immune, because they have the resources to fight the disease from the carriers. Everyone else is subject to ill effects. With luck the carriers die off before they can infect the non-immune population. When you get more carriers, the spread of the disease increases exponentially. In case of patents we see two things playing out at the same time: the race to get more bad patents issued, and more actors trying to abuse the existing ones. The herd immunity is long gone, and everyone suffers.
Natsu 12y ago
Exactly right. If 1% of 100 patents are bad, there's 1 bad patent out there. If 0.001% of 100,000,000 patents are bad, you have 1,000 bad patents and the problem is 1,000 times worse due to scale, because you only got a thousand times better and you needed to be a million times better to keep pace with the problem. And no, you can't sensibly distribute the costs of that across the population in a case like this. Patents can be asserted against everyone independently, so a larger population makes things worse, no better, as it gives trolls more victims, rather than distributing the cost. It's not as if patent litigation is over for a patent once they've sued one person (though it is interesting to imagine a model where a patent might be 'exhausted' once the holder collected 'enough' money through litigation).

by sinwave 12y ago

bostik 12y ago

> even an "extremely low" absolute number (like one in a thousand) is unacceptable. Like herd immunity in epidemology? The really large corporations are mostly immune, because they have the resources to fight the disease from the carriers. Everyone else is subject to ill effects. With luck the carriers die off before they can infect the non-immune population. When you get more carriers, the spread of the disease increases exponentially. In case of patents we see two things playing out at the same time: the race to get more bad patents issued, and more actors trying to abuse the existing ones. The herd immunity is long gone, and everyone suffers.

Natsu 12y ago

Exactly right. If 1% of 100 patents are bad, there's 1 bad patent out there. If 0.001% of 100,000,000 patents are bad, you have 1,000 bad patents and the problem is 1,000 times worse due to scale, because you only got a thousand times better and you needed to be a million times better to keep pace with the problem. And no, you can't sensibly distribute the costs of that across the population in a case like this. Patents can be asserted against everyone independently, so a larger population makes things worse, no better, as it gives trolls more victims, rather than distributing the cost. It's not as if patent litigation is over for a patent once they've sued one person (though it is interesting to imagine a model where a patent might be 'exhausted' once the holder collected 'enough' money through litigation).