Hacker News Reader

Comment by Natsu | Hacker News Reader

For a story to appear on HN's front page, you need perhaps a few dozen upvotes. For a patent to be approved, years are spent examining it.

If a bad story hits HN, it's gone within the day. If a bad patent is issued, it can be used to extort millions of dollars out of people who are doing useful work.

The situations aren't glibly comparable. Every bad patent can cost millions of dollars. This isn't a system where we as a society should permit a high false positive rate.

In this situation, a high enough false positive rate, as evidenced by bad patents like these, can make it such that the cost of the system outweighs its value, such that a rational society would abolish it if there were no workable means of correction.

That aside, I note that the patent bar tries to spin the patent lawsuit revenue as being the value of patents, rather than the cost, so I suppose entrenched interests will continue with that line.

View on HN

> years are spent examining it

It's more like a day or two are spent actually examining any case in general. It takes so long because a) it has to (there is an 18 month period between submission and publication) and b) there are a lot of patents and the queue is long.

Also you've asserted there is a high false positive rate, but I would put money on the fact that the rate is extremely low. Like people having a fear of flying because they see news reports about all the accidents, the realite is the chances of a patent getting through or you dying on in a plane crash are miniscule, but successful flights and genuine inventions are rarely worth writing about.

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.

> 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.

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).

Reply on news.ycombinator.com