Original photo courtesy of Rich Legg. Used with permission.
It’s painful to watch the many approximate pattern-based spam-fighting attempts that come up from time to time that we all know will eventually be made obsolete. Ultimately such tricks will only end up leading to more time spent weeding out false positives while the spammers stay ahead of the curve (it is their business, after all).
So not long ago, I started dumping an external catch-all account into Gmail. Since I use a new email address with every account and new beta that I sign up for (in order to catch offenders who leak my data — GoDaddy being the worst as domain registration records are public unless you pay), I started getting blasted with spam sent to randomly generated addresses.
Initially Gmail did an incredible job catching the spam; since I’ve been using this technique for the past two months, Gmail has easily caught over 250,000 spam messages.
Now, that’s not to say it’s perfect. In fact, especially lately, far from it. Though Gmail is in the unique position to harvest email from across its entire user-base and adapt its algorithm instantly the moment one of its accounts gets hit, it still can’t hit everything 100%. So, even as this is one of the biggest advantages of using a hosted email service like Gmail, it still lets more spam through than I’d like.
As far as I know, Google does not exchange spam data with other email providers (though maybe it does, I’m not sure). Whatever the case, I’m always interested in diverse tactics to dealing with spam. And given the success I’ve found with Automattic’s spam-squashing Akismet plugin on my blog, I wonder if this technology couldn’t be adapted for email?
In particular, I think that early adopters suffer from a different kind of spam abuse than most. That’s only a hunch, but I think that we make ourselves more vulnerable, especially in case of using catch-all accounts (a cardinal sin of spam management, from what I hear).
Perhaps the application of Akismet to the early-adopter spam problem could act as an additional networked preventative measure, leveraging spam trends across all email platforms, just as Akismet is starting to do for blogging platforms.
I dunno, I’m not an expert in this domain, but Akismet is one of the most promising instances of spam fighting and prevention that I’ve found and I’d love to have the same piece of mind in my email that it affords me on my blog. Could we give an Akismet-bot POP3 access to Gmail and let it loose? Better yet, could we run Akismet client-side as a Greasemonkey or Firefox extension? Again, the details probably aren’t as important as the results.
So, Matt, what’d it take to sik Akismet on my email?
Yes, Akismet should be everywhere! It’s amazing at weeding out spam!
But have you tried Spamgourmet for your signups? I find it very handy, and it saves you the trouble of setting up a new account all the time.
Very cool pic, btw! 🙂
The problems I see are:
1. In my experience, blog comment spam just isn’t that sophisticated, compared to e-mail spam. Almost every comment spam Akismet has caught for me has the same exact format. E-mail spammers seem to do a better job of looking exactly like a legitimate message from a friend or a company you do business with.
2. False positives are less devastating in blog comments. If someone wanted to say “nice job” or offer an insight, nothing too bad will happen when that comment gets eaten by the filter. But in e-mail, the thing that gets eaten could have been an important receipt, a license code, or substantial correspondence from a loved one.
What’d it take? Like 50 more servers. 😉
But seriously, email spam is pretty hairy and tackling it would be a pretty significant departure from what Akismet does. I don’t forsee us moving in that direction anytime soon, though I’d be lying if I said I haven’t considered it when dealing with the influx of email spam that’s been getting through lately.
Well, it’s not that I expect you to do it, Matt… but I see a business opportunity given the success of Akismet.
The question is: could Akismet be used on email? And could something like Amazon’s Elastic Cloud help in this respect?