Hacker News Reader

I upvoted, not because I agree, but because I believe the people downvoting you are downvoting because they disagree and not because your comment shouldn't be heard.

Imagine he had said: I believe a free and open market for weapons would be a good thing, because it would reduce the number of defenseless people, would result in a power imbalance that puts generally-okay actors at an advantage (say what you will, but the mob doesn't have 1% of the resources the US government does), and would therefore reduce crime.

I do not personally find that argument compelling (and it is of identical structure to the above), but me disagreeing with it does not mean it isn't of sufficient quality for Hacker News.

Thanks. I was disappointed to see downvotes rather than engagement in the argument.

You prompt a fascinating hypothetical: suppose all the world's atomic weapons were put up in a free, transparent auction? (i.e. bidders and bids are disclosed)

I'd argue that most would end up in the hands of good actors (e.g. the world bands behind the Dalai Lama to buy and destroy them all), but who knows? Good debate topic for happy hour, though.

Unfortunately, the size of the world's nuclear arsenal being what it is, it doesn't much matter if "most" of the nuclear weapons end up in the hands of good actors.

I think this is also the case for vulnerabilities as well unfortunately.

That was my point.

> I do not personally find that argument compelling (and it is of identical structure to the above)

Replacing "vulnerabilities" with "firearms" while retaining the same argument structure does not mean that both arguments are logically identical (a false equivalency). The implications of a free market for information security and a free market for firearms are totally different.

Analogies rarely, if ever, imply a total equivalence between the things being compared. That's what makes them analogies, not formal deductive proofs.

I've noticed this becoming a trend. The votes are not for the quality of the comment but rather the agreeability. It's too bad...

With some round of the vote tweaks they did earlier this year it seemed like there was a huge amount of extra volatility injected in the early up/downvotes. Almost any comment on how a comment has been voted down seems to wind up wrong within a half hour or so. And this happens quite a lot.

I have a running hypothesis that those particular "wow, this is receiving a lot of downvotes" comments cause upvotes to the post they're attached to. (Unless they're by the original author of the downvoted post; then they just cause snotty arguments.)

If you cared a lot about Internet Points, I'm betting keeping a sockpuppet acccount just to post "why was [master]'s post downvoted?" comments would be one of the highest-ROI things you could do to get more of them.

Yeah, I had considered this, but I can't imagine the effect continuing beyond the un-greying of the parent post. I can see it rallying to that point, though.

I should also note that I've seen a lot of early-greyed posts come back from the dead after 30 minutes or so even without the "omgwtfisgoingon" posts underneath them as well. Sometimes even becoming clear winners on their strata of the post's comments.

I don't think the weapons analogy is a good one. For the developer of the software the purchase of the "weapon" is also essentially makes all other instances of that weapon useless forever (assuming customers patch promptly).

Perhaps restricting the analogy to nuclear weapons would make sense. Only nation states (software companies) and terrorists (malicious hackers... and perhaps intelligence agencies) would be interested in purchasing such weapons.

I think that is one of many salient differences between the scenarios.

My thought is: if one argument is good and another is bad, clearly the reason for the difference does not exist within the properties they share. It must exist where the arguments are different, or more likely in this case (as what the poster said and what I said are so similar), in the outside world of facts and understanding.

But if the shape of the argument allows for things that may successfully convince you, then clearly that information is not enough to dismiss it outright. It should be the start of a dialog with the parent comment.

Now, for something completely different:

If, for some reason, you want to actually engage _me_ on this particular policy issue, angle your thought at this: my understanding is that it's 2014, and if you want to regulate the information people share or sell over the internet you've accepted a challenge I do not envy.

I would be particularly skeptical you could do this without building a comprehensive computer spying system and perhaps outlawing crypto (it's been tried), as you're trying to read communications of highly security-conscious people.

As for if Mitnick is a good person (or similar things under discussion here), I have not met him.

If and when you do meet him, just take all he says with a grain of salt. Very full of himself :)

I'm not sure you can, in practice, have a completely free and open market for vulnerabilities, since any information about the vulnerability released broadly is likely to shorten the path to others discovering the same vulnerability [1], devaluing it implicitly through partial disclosure. So there's an interest in the market being small and secret. For the seller to keep the price high and for the buyer to maintain exclusivity on the exploit.

[1] As a case in point, I showed the headline of the email for the bash vulnerability to a coworker today on the commute and he instantly described in accurate detail how it probably works. Not that this was a particularly difficult case, but I think the principle holds.

> This is fairly axiomatic. An open market increases the price of vulnerabilities which in turn increases the number of vulnerabilities found (unless you want to argue the ability to find vulnerabilities is inelastic for some reason)

Wait a second...won't increasing the number of vulnerabilities found push prices down? If I'm looking to penetrate a system I only need to buy one vulnerability, so in effect different vulnerabilities are somewhat fungible and so should compete on price. Hence, if more vulnerabilities are being found and coming to market, prices should be going down.

On the other hand, with a free and open market for vulnerabilities there would likely be people who would NOT have bought vulnerabilities on the black market buying vulnerabilities on the safer, easier to use free and open market, so demand could go up, raising prices.

>An open market increases the price of vulnerabilities

That's impossible to tell. You could just as easily say that the price will crash when you take away all the costs and risk of running a black market and give buyers a place to compare multiple "products." The demand side could just as easily be inelastic (or at least saturated) as the supply.

> the good actors should always win the bid

This works if you're talking about Microsoft, but not if you're talking about smaller companies or open source products. Maybe a Google or a Facebook would step up and pay off the market for things that they use, but "the rich people will take care of us" is not a setup that I'm comfortable with.

But you aren't choosing between a perfect world, where bugs get reported only to the vendors and the vendors promptly fix them, and an open market. Right now, governments and shady corporations are already buying. If it costs some company $250k or $1M to buy their vulnerability, maybe they'll work harder in the future. Or hire their own pen testers. And they'll have the money, because almost by definition attractive vulnerabilities are those in widely used software.

In the complaints about auctioning off vulnerabilities it's hard to avoid hearing companies bitching that they may have to pay security researchers, and it will be harder to intimidate them with law enforcement.

I mostly disagree with arguments to rationalize the sale of exploits, they create a massive power balance towards bad actors, but we have to be honest with ourselves, and like drugs, 0days are not going away.

Our only proper response is secure software development practices, employment of security reseachers, and adoption of security-centric practices in critical systems... such as the Linux kernel. Which is embarassingly not the case at the moment. For ex: http://unix.stackexchange.com/questions/59020/why-are-the-gr...

On your second point.... Governments should be assumed to be bad actors and they certainly have some of the deepest pockets. If an open market increases the price, it would seem to make it a better market for bad actors.

Especially if we think of small software companies or open-source projects (like OpenSSL) who cant afford to pay hundreds of thousands of dollars to secure their own exploit.

On your overall point... I think this issue of selling 0days is more a debate of ethics, and I don't think economics can solve a problem of ethics.

An open market increases the price of vulnerabilities

What's your logic behind this? I believe this to be false. To my knowledge the black market commands artificially high prices on illicit goods as a rule, except when the good is available on the open market. See:

1) The goods are stolen and need to be unloaded quickly.

2) Open market prices are artificially high thanks to things like taxes (example: alcohol, cigarettes)

The market concept is nice for the reasons you listed, but doesn't it add an incentive to introduce obfuscated exploits into code then sell an exploit to them later?

I don't agree with you (on all points) but none the less applaud your ideas and the way you formulate your arguments.