Gates on spam

Bill Gates has a column in the Washington Post today in which he calls for both legislation and technological innovation by the software industry to help fight email spam.

The technology industry is committed to helping solve the spam problem, which hurts business productivity and is an unnecessary burden on everyone who uses e-mail. Relief is on the way. With strong legislation, continued vigorous enforcement and the technological advances we and other industry leaders are developing, the promise of a nearly spam-free future is coming closer to reality.

I don't see how more laws are going to stop bits of information from traveling from one person's computer to another's if copyright laws cannot stop online file-sharing. Technological innovation is more promising, and one thing Gates did not mention was using market incentives to find solutions. A previous TCS article by David Friedman spoke of using digital stamps to put a price - a very small price for the average user - on your emailbox.

Your digital stamps differ from the stamps provided by the post office in one important respect-yours are reusable. From the standpoint of the ordinary email user, digital stamps cost, on average, nothing, since he receives about as many as he sends. They may even cost less than nothing; if he sets a low enough price on his mailbox, some spam will make it through, giving him a positive balance of stamps received over stamps sent. Your online stamp machine will buy as well as sell, allowing such users to convert that balance into a small income. Spam is no longer a problem for those who do not want to receive it, and it is now a positive, if small, benefit for those who do.

A more thorough exploration of this idea is found in a draft version of Friedman's book Future Imperfect.

Share this

i liked Friedman's idea but

i liked Friedman's idea but you don't even need stamps, other solutions are already popping up. One scheme bounces all mail sent to you until a live human clicks on a link or enters a non-machine readable code. At that point the sender goes on your "white list" and can email you without futher interference. Anyone but a bot can reach you without paying.

who regulates and maintains

who regulates and maintains such a system? What about mailing lists? Would it cost a professor money to send assignments to students, as needed?

One way to stop spam is to

One way to stop spam is to offer bounties on spammers. I discussed this on my own blog last week, and Lawrence Lessig has discussed it recently as well. This solution leaves new legislation out of the picture completely, doesn't rule out technological innovations in the least, and allows those so motivated to do the rest of us a favor. With all the computer knowledge disseminated across the internet, there must be plenty of people with the know-how to catch spammers sitting around without any incentive to do the work.

Did you read the article,

Did you read the article, matt?

The user regulates his own mailbox. He decides who can send mail to him without a "stamp"--that's the whitelist. It would be pretty stupid for a student not to whitelist his professor, if he wants to succeed in the class.

ahhhh... thanks Qiwi. I

ahhhh... thanks Qiwi. I didn't get a chance to read the article, no, hence the rather ignorant questions.

In addition to the digital

In addition to the digital stamp scheme and the bounce/turing code scheme mentioned by JTK, there's also bayesian filtering, which can be quite good.

the biggest problem I see

the biggest problem I see with a white list is that if I were a spammer I would just spoof your email address. For example, my email address is: akschodh@yahoo.com and the spammers would just send me an email as if it were coming from akschodh@yahoo.com. Chances are that I am going to allow emails from myself to come through b/c I CC myself on a lot of group emails, etc.

The above problem is something the technological advances will hopefully fix. I don't like the bounty-hunter idea because I think the rules of engagement are not defined. Preferrably, legislation or the internet community would go after the companies that pay spammers for their "service".

I use a lot of Bayesian like filtering at work and it does pretty well but its not a 100% which is acceptable in many cases but I don't think in email.

Spoonie - is the only

Spoonie - is the only problem with bounties that the rules of engagement aren't currently defined? Because I don't think that defining those rules would take a lot of effort at all. I don't know what Microsoft's policy on engagement of virus writers is, for instance, but I imagine it's set up in a way that will work (otherwise why even offer the bounties?). And if Microsoft can figure out a way to do it, the civilized world can too.

As for legislation being aimed at the companies that pay spammers - well, legislating this sort of thing just isn't a solution if you really want to fix the problem. The DMCA didn't stop music or TV or movie pirating, did it? The methods Congress uses in drafting legislation almost guarantee that the final product most likely will not do what the preliminary version was intended to do (and most often won't do anything at all).

Jason, I agree that the

Jason, I agree that the rules of engagement would be simple to define, they just aren't done yet.

I believe the reason the DMCA/RIAA are not stopping piracy is that they are going for the individual pirates and not the source. Granted in their case it is much more difficult because the technology involved is readily available for legitimate reasons.

but from my understanding, there are only a handful of companies that pay spammers for their clicks (i.e. www.doubleclick.com) and if those guys are regulated (bad word, I know) that they do not pay for clicks to unwanted email addresses, in an ideal world, the spammers wouldn't bother emailing us anymore. You may need to put some negative incentive into the regulation as well.

Spoonie - Aren't the

Spoonie -

Aren't the individual pirates themselves the source of the piracy problem? Without individual pirates, all the file-sharing utilities in the world wouldn't result in piracy. If the members of the RIAA were intelligent about this whole thing, they'd realize the market is shifting beneath them and change their business model rather than clutching to outdated methods of distribution for the bulk of their income.

This is all beside the point, though - the point is that the DMCA, as a piece of government legislation, did nothing to reduce the problems it supposedly addressed. It has plenty of negative incentives (jail time, large fines) for violations, yet it has failed to influence those who pay no attention to the negative incentives. There is little reason to doubt that legislating against spam would have any different outcome than the DMCA has had on copyright violations. I think the catch in what you've said is the phrase "in an ideal world." Legislation of things that can't ever be completely eradicated is always based on the "ideal world" standard - and it always fails because of this standard as well.

Whoops. When I said "There

Whoops. When I said "There is little reason to doubt that legislating against spam would have any different outcome," what I really meant to say was, "There is little reason to believe that legislating against spam...."

I need to edit more carefully before I post.