Email Is Broken
Fred Wilson touched on a topic I've been thinking about for a couple years now: e-mail as we currently know it is broken, but how can it be fixed?
There's no lack of e-mail alternatives: IM, SMS, web-forums, and blogs all replace at least some of the functionality of e-mail. But nothing seems well positioned to step up as the next generation of ubiquitous messaging service.
A big part of the problem is that the very feature which makes e-mail so useful--the ability to send and receive messages to and from anyone--is the same feature which makes it so vulnerable to spam.
It wouldn't be hard, for example, to set up a closed e-mail network where you will only get messages from your pre-approved "friends." As Fred correctly points out, many social networks already include this feature, and spam is generally not an issue on those networks.
Simply requiring a sender to ask to be a "friend" really just pushes the problem to a different level, though. A spammer could just as easily send millions of "be my friend" requests and be as much of a headache as they are with spam e-mail. Worse, it wouldn't do much about phishing schemes, since those could be crafted to look like legitimate requests from real companies.
The fundamental dilemma is as follows:
No replacement e-mail system will take off unless it's ubiquitous. The fact that anyone can communicate with anyone is the main reason the current e-mail system still dominates, rather than being replaced by some other (presumably closed) messaging system.
Companies and ISPs won't adopt a new messaging system unless they have the ability to administer their own accounts--a central Ministry of E-mail Management won't fly.
But as long as companies can create their own accounts, spammers can do the same to generate tons of spam (or possibly "be my friend" messages) from millions of unique addresses.
The only way to prevent the spammers is thus to break the anyone-to-anyone communications in e-mail.
So to fix e-mail, you either need centralized account management or you need to limit the ability for anyone to send a message to anyone else. The former is unacceptable to the businesses which actually run e-mail servers, and the latter limits the network effect which makes e-mail useful.
There are a couple of technical solutions which I could see as breaking the logjam:
A reliable, unobtrusive way to distinguish human- from machine-generated messages. The current "challenge-response" systems are very effective, but too obnoxious for widespread adoption. If you could figure this out, though, it would make it possible to apply different rules to messages depending on the sentience of the sender.
It may be possible to modestly limit the "anyone to anyone" messaging feature of e-mail in a way which blocks most unwanted messages without getting in the way of human senders. Whitelisting and Blacklisting are forms of this approach, but not powerful enough by themselves to solve the problem.